Binary Option Payoff Formula Example

What's new in macOS 11, Big Sur!

It's that time of year again, and we've got a new version of macOS on our hands! This year we've finally jumped off the 10.xx naming scheme and now going to 11! And with that, a lot has changed under the hood in macOS.
As with previous years, we'll be going over what's changed in macOS and what you should be aware of as a macOS and Hackintosh enthusiast.

Has Nvidia Support finally arrived?

Sadly every year I have to answer the obligatory question, no there is no new Nvidia support. Currently Nvidia's Kepler line is the only natively supported gen.
However macOS 11 makes some interesting changes to the boot process, specifically moving GPU drivers into stage 2 of booting. Why this is relevant is due to Apple's initial reason for killing off Web Drivers: Secure boot. What I mean is that secure boot cannot work with Nvidia's Web Drivers due to how early Nvidia's drivers have to initialize at, and thus Apple refused to sign the binaries. With Big Sur, there could be 3rd party GPUs however the chances are still super slim but slightly higher than with 10.14 and 10.15.

What has changed on the surface

A whole new iOS-like UI

Love it or hate it, we've got a new UI more reminiscent of iOS 14 with hints of skeuomorphism(A somewhat subtle call back to previous mac UIs which have neat details in the icons)
You can check out Apple's site to get a better idea:

macOS Snapshotting

A feature initially baked into APFS back in 2017 with the release of macOS 10.13, High Sierra, now macOS's main System volume has become both read-only and snapshotted. What this means is:
However there are a few things to note with this new enforcement of snapshotting:

What has changed under the hood

Quite a few things actually! Both in good and bad ways unfortunately.

New Kernel Cache system: KernelCollections!

So for the past 15 years, macOS has been using the Prelinked Kernel as a form of Kernel and Kext caching. And with macOS Big Sur's new Read-only, snapshot based system volume, a new version of caching has be developed: KernelCollections!
How this differs to previous OSes:

Secure Boot Changes

With regards to Secure Boot, now all officially supported Macs will also now support some form of Secure Boot even if there's no T2 present. This is now done in 2 stages:
While technically these security features are optional and can be disabled after installation, many features including OS updates will no longer work reliably once disabled. This is due to the heavy reliance of snapshots for OS updates, as mentioned above and so we highly encourage all users to ensure at minimum SecureBootModel is set to Default or higher.

No more symbols required

This point is the most important part, as this is what we use for kext injection in OpenCore. Currently Apple has left symbols in place seemingly for debugging purposes however this is a bit worrying as Apple could outright remove symbols in later versions of macOS. But for Big Sur's cycle, we'll be good on that end however we'll be keeping an eye on future releases of macOS.

New Kernel Requirements

With this update, the AvoidRuntimeDefrag Booter quirk in OpenCore broke. Because of this, the macOS kernel will fall flat when trying to boot. Reason for this is due to cpu_count_enabled_logical_processors requiring the MADT (APIC) table, and so OpenCore will now ensure this table is made accessible to the kernel. Users will however need a build of OpenCore 0.6.0 with commit bb12f5f or newer to resolve this issue.
Additionally, both Kernel Allocation requirements and Secure Boot have also broken with Big Sur due to the new caching system discussed above. Thankfully these have also been resolved in OpenCore 0.6.3.
To check your OpenCore version, run the following in terminal:
nvram 4D1FDA02-38C7-4A6A-9CC6-4BCCA8B30102:opencore-version
If you're not up-to-date and running OpenCore 0.6.3+, see here on how to upgrade OpenCore: Updating OpenCore, Kexts and macOS

Broken Kexts in Big Sur

Unfortunately with the aforementioned KernelCollections, some kexts have unfortunately broken or have been hindered in some way. The main kexts that currently have issues are anything relying on Lilu's userspace patching functionality:
Thankfully most important kexts rely on kernelspace patcher which is now in fact working again.

MSI Navi installer Bug Resolved

For those receiving boot failures in the installer due to having an MSI Navi GPU installed, macOS Big Sur has finally resolved this issue!

New AMD OS X Kernel Patches

For those running on AMD-Based CPUs, you'll want to also update your kernel patches as well since patches have been rewritten for macOS Big Sur support:

Other notable Hackintosh issues

Several SMBIOS have been dropped

Big Sur dropped a few Ivy Bridge and Haswell based SMBIOS from macOS, so see below that yours wasn't dropped:
If your SMBIOS was supported in Catalina and isn't included above, you're good to go! We also have a more in-depth page here: Choosing the right SMBIOS
For those wanting a simple translation for their Ivy and Haswell Machines:

Dropped hardware

Currently only certain hardware has been officially dropped:

Extra long install process

Due to the new snapshot-based OS, installation now takes some extra time with sealing. If you get stuck at Forcing CS_RUNTIME for entitlement, do not shutdown. This will corrupt your install and break the sealing process, so please be patient.

X79 and X99 Boot issues

With Big Sur, IOPCIFamily went through a decent rewriting causing many X79 and X99 boards to fail to boot as well as panic on IOPCIFamily. To resolve this issue, you'll need to disable the unused uncore bridge:
You can also find prebuilts here for those who do not wish to compile the file themselves:

New RTC requirements

With macOS Big Sur, AppleRTC has become much more picky on making sure your OEM correctly mapped the RTC regions in your ACPI tables. This is mainly relevant on Intel's HEDT series boards, I documented how to patch said RTC regions in OpenCorePkg:
For those having boot issues on X99 and X299, this section is super important; you'll likely get stuck at PCI Configuration Begin. You can also find prebuilts here for those who do not wish to compile the file themselves:

SATA Issues

For some reason, Apple removed the AppleIntelPchSeriesAHCI class from AppleAHCIPort.kext. Due to the outright removal of the class, trying to spoof to another ID (generally done by SATA-unsupported.kext) can fail for many and create instability for others. * A partial fix is to block Big Sur's AppleAHCIPort.kext and inject Catalina's version with any conflicting symbols being patched. You can find a sample kext here: Catalina's patched AppleAHCIPort.kext * This will work in both Catalina and Big Sur so you can remove SATA-unsupported if you want. However we recommend setting the MinKernel value to 20.0.0 to avoid any potential issues.

Legacy GPU Patches currently unavailable

Due to major changes in many frameworks around GPUs, those using ASentientBot's legacy GPU patches are currently out of luck. We either recommend users with these older GPUs stay on Catalina until further developments arise or buy an officially supported GPU

What’s new in the Hackintosh scene?

Dortania: a new organization has appeared

As many of you have probably noticed, a new organization focusing on documenting the hackintoshing process has appeared. Originally under my alias, Khronokernel, I started to transition my guides over to this new family as a way to concentrate the vast amount of information around Hackintoshes to both ease users and give a single trusted source for information.
We work quite closely with the community and developers to ensure information's correct, up-to-date and of the best standards. While not perfect in every way, we hope to be the go-to resource for reliable Hackintosh information.
And for the times our information is either outdated, missing context or generally needs improving, we have our bug tracker to allow the community to more easily bring attention to issues and speak directly with the authors:

Dortania's Build Repo

For those who either want to run the lastest builds of a kext or need an easy way to test old builds of something, Dortania's Build Repo is for you!
Kexts here are built right after commit, and currently supports most of Acidanthera's kexts and some 3rd party devs as well. If you'd like to add support for more kexts, feel free to PR: Build Repo source

True legacy macOS Support!

As of OpenCore's latest versioning, 0.6.2, you can now boot every version of x86-based builds of OS X/macOS! A huge achievement on @Goldfish64's part, we now support every major version of kernel cache both 32 and 64-bit wise. This means machines like Yonah and newer should work great with OpenCore and you can even relive the old days of OS X like OS X 10.4!
And Dortania guides have been updated accordingly to accommodate for builds of those eras, we hope you get as much enjoyment going back as we did working on this project!

Intel Wireless: More native than ever!

Another amazing step forward in the Hackintosh community, near-native Intel Wifi support! Thanks to the endless work on many contributors of the OpenIntelWireless project, we can now use Apple's built-in IO80211 framework to have near identical support to those of Broadcom wireless cards including features like network access in recovery and control center support.
For more info on the developments, please see the itlwm project on GitHub: itlwm

Clover's revival? A frankestien of a bootloader

As many in the community have seen, a new bootloader popped up back in April of 2019 called OpenCore. This bootloader was made by the same people behind projects such as Lilu, WhateverGreen, AppleALC and many other extremely important utilities for both the Mac and Hackintosh community. OpenCore's design had been properly thought out with security auditing and proper road mapping laid down, it was clear that this was to be the next stage of hackintoshing for the years we have left with x86.
And now lets bring this back to the old crowd favorite, Clover. Clover has been having a rough time of recent both with the community and stability wise, with many devs jumping ship to OpenCore and Clover's stability breaking more and more with C++ rewrites, it was clear Clover was on its last legs. Interestingly enough, the community didn't want Clover to die, similarly to how Chameleon lived on through Enoch. And thus, we now have the Clover OpenCore integration project(Now merged into Master with r5123+).
The goal is to combine OpenCore into Clover allowing the project to live a bit longer, as Clover's current state can no longer boot macOS Big Sur or older versions of OS X such as 10.6. As of writing, this project seems to be a bit confusing as there seems to be little reason to actually support Clover. Many of Clover's properties have feature-parity in OpenCore and trying to combine both C++ and C ruins many of the features and benefits either languages provide. The main feature OpenCore does not support is macOS-only ACPI injection, however the reasoning is covered here: Does OpenCore always inject SMBIOS and ACPI data into other OSes?

Death of x86 and the future of Hackintoshing

With macOS Big Sur, a big turning point is about to happen with Apple and their Macs. As we know it, Apple will be shifting to in-house designed Apple Silicon Macs(Really just ARM) and thus x86 machines will slowly be phased out of their lineup within 2 years.
What does this mean for both x86 based Macs and Hackintoshing in general? Well we can expect about 5 years of proper OS support for the iMac20,x series which released earlier this year with an extra 2 years of security updates. After this, Apple will most likely stop shipping x86 builds of macOS and hackintoshing as we know it will have passed away.
For those still in denial and hope something like ARM Hackintoshes will arrive, please consider the following:
So while we may be heart broken the journey is coming to a stop in the somewhat near future, hackintoshing will still be a time piece in Apple's history. So enjoy it now while we still can, and we here at Dortania will still continue supporting the community with our guides till the very end!

Getting ready for macOS 11, Big Sur

This will be your short run down if you skipped the above:
For the last 2, see here on how to update: Updating OpenCore, Kexts and macOS
In regards to downloading Big Sur, currently gibMacOS in macOS or Apple's own software updater are the most reliable methods for grabbing the installer. Windows and Linux support is still unknown so please stand by as we continue to look into this situation, macrecovery.py may be more reliable if you require the recovery package.
And as with every year, the first few weeks to months of a new OS release are painful in the community. We highly advise users to stay away from Big Sur for first time installers. The reason is that we cannot determine whether issues are Apple related or with your specific machine, so it's best to install and debug a machine on a known working OS before testing out the new and shiny.
For more in-depth troubleshooting with Big Sur, see here: OpenCore and macOS 11: Big Sur
submitted by dracoflar to hackintosh [link] [comments]

No gods, no kings, only NOPE - or divining the future with options flows. [Part 2: A Random Walk and Price Decoherence]

tl;dr -
1) Stock prices move continuously because different market participants end up having different ideas of the future value of a stock.
2) This difference in valuations is part of the reason we have volatility.
3) IV crush happens as a consequence of future possibilities being extinguished at a binary catalyst like earnings very rapidly, as opposed to the normal slow way.
I promise I'm getting to the good parts, but I'm also writing these as a guidebook which I can use later so people never have to talk to me again.
In this part I'm going to start veering a bit into the speculation territory (e.g. ideas I believe or have investigated, but aren't necessary well known) but I'm going to make sure those sections are properly marked as speculative (and you can feel free to ignore/dismiss them). Marked as [Lily's Speculation].
As some commenters have pointed out in prior posts, I do not have formal training in mathematical finance/finance (my background is computer science, discrete math, and biology), so often times I may use terms that I've invented which have analogous/existing terms (e.g. the law of surprise is actually the first law of asset pricing applied to derivatives under risk neutral measure, but I didn't know that until I read the papers later). If I mention something wrong, please do feel free to either PM me (not chat) or post a comment, and we can discuss/I can correct it! As always, buyer beware.
This is the first section also where you do need to be familiar with the topics I've previously discussed, which I'll add links to shortly (my previous posts:
1) https://www.reddit.com/thecorporation/comments/jck2q6/no_gods_no_kings_only_nope_or_divining_the_future/
2) https://www.reddit.com/thecorporation/comments/jbzzq4/why_options_trading_sucks_or_the_law_of_surprise/
---
A Random Walk Down Bankruptcy
A lot of us have probably seen the term random walk, maybe in the context of A Random Walk Down Wall Street, which seems like a great book I'll add to my list of things to read once I figure out how to control my ADD. It seems obvious, then, what a random walk means - when something is moving, it basically means that the next move is random. So if my stock price is $1 and I can move in $0.01 increments, if the stock price is truly randomly walking, there should be roughly a 50% chance it moves up in the next second (to $1.01) or down (to $0.99).
If you've traded for more than a hot minute, this concept should seem obvious, because especially on the intraday, it usually isn't clear why price moves the way it does (despite what chartists want to believe, and I'm sure a ton of people in the comments will tell me why fettucini lines and Batman doji tell them things). For a simple example, we can look at SPY's chart from Friday, Oct 16, 2020:

https://preview.redd.it/jgg3kup9dpt51.png?width=1368&format=png&auto=webp&s=bf8e08402ccef20832c96203126b60c23277ccc2
I'm sure again 7 different people can tell me 7 different things about why the chart shape looks the way it does, or how if I delve deeply enough into it I can find out which man I'm going to marry in 2024, but to a rationalist it isn't exactly apparent at why SPY's price declined from 349 to ~348.5 at around 12:30 PM, or why it picked up until about 3 PM and then went into precipitous decline (although I do have theories why it declined EOD, but that's for another post).
An extremely clever or bored reader from my previous posts could say, "Is this the price formation you mentioned in the law of surprise post?" and the answer is yes. If we relate it back to the individual buyer or seller, we can explain the concept of a stock price's random walk as such:
Most market participants have an idea of an asset's true value (an idealized concept of what an asset is actually worth), which they can derive using models or possibly enough brain damage. However, an asset's value at any given time is not worth one value (usually*), but a spectrum of possible values, usually representing what the asset should be worth in the future. A naive way we can represent this without delving into to much math (because let's face it, most of us fucking hate math) is:
Current value of an asset = sum over all (future possible value multiplied by the likelihood of that value)
In actuality, most models aren't that simple, but it does generalize to a ton of more complicated models which you need more than 7th grade math to understand (Black-Scholes, DCF, blah blah blah).
While in many cases the first term - future possible value - is well defined (Tesla is worth exactly $420.69 billion in 2021, and maybe we all can agree on that by looking at car sales and Musk tweets), where it gets more interesting is the second term - the likelihood of that value occurring. [In actuality, the price of a stock for instance is way more complicated, because a stock can be sold at any point in the future (versus in my example, just the value in 2021), and needs to account for all values of Tesla at any given point in the future.]
How do we estimate the second term - the likelihood of that value occurring? For this class, it actually doesn't matter, because the key concept is this idea: even with all market participants having the same information, we do anticipate that every participant will have a slightly different view of future likelihoods. Why is that? There's many reasons. Some participants may undervalue risk (aka WSB FD/yolos) and therefore weight probabilities of gaining lots of money much more heavily than going bankrupt. Some participants may have alternative data which improves their understanding of what the future values should be, therefore letting them see opportunity. Some participants might overvalue liquidity, and just want to GTFO and thereby accept a haircut on their asset's value to quickly unload it (especially in markets with low liquidity). Some participants may just be yoloing and not even know what Fastly does before putting their account all in weekly puts (god bless you).
In the end, it doesn't matter either the why, but the what: because of these diverging interpretations, over time, we can expect the price of an asset to drift from the current value even with no new information added. In most cases, the calculations that market participants use (which I will, as a Lily-ism, call the future expected payoff function, or FEPF) ends up being quite similar in aggregate, and this is why asset prices likely tend to move slightly up and down for no reason (or rather, this is one interpretation of why).
At this point, I expect the 20% of you who know what I'm talking about or have a finance background to say, "Oh but blah blah efficient market hypothesis contradicts random walk blah blah blah" and you're correct, but it also legitimately doesn't matter here. In the long run, stock prices are clearly not a random walk, because a stock's value is obviously tied to the company's fundamentals (knock on wood I don't regret saying this in the 2020s). However, intraday, in the absence of new, public information, it becomes a close enough approximation.
Also, some of you might wonder what happens when the future expected payoff function (FEPF) I mentioned before ends up wildly diverging for a stock between participants. This could happen because all of us try to short Nikola because it's quite obviously a joke (so our FEPF for Nikola could, let's say, be 0), while the 20 or so remaining bagholders at NikolaCorporation decide that their FEPF of Nikola is $10,000,000 a share). One of the interesting things which intuitively makes sense, is for nearly all stocks, the amount of divergence among market participants in their FEPF increases substantially as you get farther into the future.
This intuitively makes sense, even if you've already quit trying to understand what I'm saying. It's quite easy to say, if at 12:51 PM SPY is worth 350.21 that likely at 12:52 PM SPY will be worth 350.10 or 350.30 in all likelihood. Obviously there are cases this doesn't hold, but more likely than not, prices tend to follow each other, and don't gap up/down hard intraday. However, what if I asked you - given SPY is worth 350.21 at 12:51 PM today, what will it be worth in 2022?
Many people will then try to half ass some DD about interest rates and Trump fleeing to Ecuador to value SPY at 150, while others will assume bull markets will continue indefinitely and SPY will obviously be 7000 by then. The truth is -- no one actually knows, because if you did, you wouldn't be reading a reddit post on this at 2 AM in your jammies.
In fact, if you could somehow figure out the FEPF of all market participants at any given time, assuming no new information occurs, you should be able to roughly predict the true value of an asset infinitely far into the future (hint: this doesn't exactly hold, but again don't @ me).
Now if you do have a finance background, I expect gears will have clicked for some of you, and you may see strong analogies between the FEPF divergence I mentioned, and a concept we're all at least partially familiar with - volatility.
Volatility and Price Decoherence ("IV Crush")
Volatility, just like the Greeks, isn't exactly a real thing. Most of us have some familiarity with implied volatility on options, mostly when we get IV crushed the first time and realize we just lost $3000 on Tesla calls.
If we assume that the current price should represent the weighted likelihoods of all future prices (the random walk), volatility implies the following two things:
  1. Volatility reflects the uncertainty of the current price
  2. Volatility reflects the uncertainty of the future price for every point in the future where the asset has value (up to expiry for options)
[Ignore this section if you aren't pedantic] There's obviously more complex mathematics, because I'm sure some of you will argue in the comments that IV doesn't go up monotonically as option expiry date goes longer and longer into the future, and you're correct (this is because asset pricing reflects drift rate and other factors, as well as certain assets like the VIX end up having cost of carry).
Volatility in options is interesting as well, because in actuality, it isn't something that can be exactly computed -- it arises as a plug between the idealized value of an option (the modeled price) and the real, market value of an option (the spot price). Additionally, because the makeup of market participants in an asset's market changes over time, and new information also comes in (thereby increasing likelihood of some possibilities and reducing it for others), volatility does not remain constant over time, either.
Conceptually, volatility also is pretty easy to understand. But what about our friend, IV crush? I'm sure some of you have bought options to play events, the most common one being earnings reports, which happen quarterly for every company due to regulations. For the more savvy, you might know of expected move, which is a calculation that uses the volatility (and therefore price) increase of at-the-money options about a month out to calculate how much the options market forecasts the underlying stock price to move as a response to ER.
Binary Catalyst Events and Price Decoherence
Remember what I said about price formation being a gradual, continuous process? In the face of special circumstances, in particularly binary catalyst events - events where the outcome is one of two choices, good (1) or bad (0) - the gradual part gets thrown out the window. Earnings in particular is a common and notable case of a binary event, because the price will go down (assuming the company did not meet the market's expectations) or up (assuming the company exceeded the market's expectations) (it will rarely stay flat, so I'm not going to address that case).
Earnings especially is interesting, because unlike other catalytic events, they're pre-scheduled (so the whole market expects them at a certain date/time) and usually have publicly released pre-estimations (guidance, analyst predictions). This separates them from other binary catalysts (e.g. FSLY dipping 30% on guidance update) because the market has ample time to anticipate the event, and participants therefore have time to speculate and hedge on the event.
In most binary catalyst events, we see rapid fluctuations in price, usually called a gap up or gap down, which is caused by participants rapidly intaking new information and changing their FEPF accordingly. This is for the most part an anticipated adjustment to the FEPF based on the expectation that earnings is a Very Big Deal (TM), and is the reason why volatility and therefore option premiums increase so dramatically before earnings.
What makes earnings so interesting in particular is the dramatic effect it can have on all market participants FEPF, as opposed to let's say a Trump tweet, or more people dying of coronavirus. In lots of cases, especially the FEPF of the short term (3-6 months) rapidly changes in response to updated guidance about a company, causing large portions of the future possibility spectrum to rapidly and spectacularly go to zero. In an instant, your Tesla 10/30 800Cs go from "some value" to "not worth the electrons they're printed on".
[Lily's Speculation] This phenomena, I like to call price decoherence, mostly as an analogy to quantum mechanical processes which produce similar results (the collapse of a wavefunction on observation). Price decoherence occurs at a widespread but minor scale continuously, which we normally call price formation (and explains portions of the random walk derivation explained above), but hits a special limit in the face of binary catalyst events, as in an instant rapid portions of the future expected payoff function are extinguished, versus a more gradual process which occurs over time (as an option nears expiration).
Price decoherence, mathematically, ends up being a more generalizable case of the phenomenon we all love to hate - IV crush. Price decoherence during earnings collapses the future expected payoff function of a ticker, leading large portions of the option chain to be effectively worthless (IV crush). It has interesting implications, especially in the case of hedged option sellers, our dear Market Makers. This is because given the expectation that they maintain delta-gamma neutral, and now many of the options they have written are now worthless and have 0 delta, what do they now have to do?
They have to unwind.
[/Lily's Speculation]
- Lily
submitted by the_lilypad to thecorporation [link] [comments]

Wasteland 3 - Review Thread

Game Information

Game Title: Wasteland 3
Platforms:
Trailers:
Developer: inXile Entertainment
Publisher: Deep Silver
Review Aggregator:
OpenCritic - 84 average - 97% recommended - 39 reviews

Critic Reviews

33bits - Juanma F. Padilla - Spanish - 95 / 100
After the excellent Wasteland 2, we were excited to get our hands on the new installment, and we can say without fear that it has met expectations. Wasteland 3 is a sign of the love that InXile has for his work and Brian Fargo for the genre that has created a name for him. If you are a lover of the saga or the genre, do not hesitate to enjoy it.
ACG - Jeremy Penter - Buy

Video Review - Quote not available

Attack of the Fanboy - Diego Perez - 4.5 / 5 stars
Wasteland 3 is one of the best RPGs I've played in years, and it's one you absolutely should not skip.
CGMagazine - Lane Martin - 9 / 10
Wasteland 3 is a lovely return to the post nuclear apocalypse with fun gameplay and interesting choices at its forefront, though at times it can be a bit clumsy in its implementation.
COGconnected - Tony Bae - 90 / 100
Wasteland 3 doesn’t pull any punches with its subject matter in sexuality, violence, and language. But if you are fine with that, I would highly recommend you give Wasteland 3 a shot, especially if you were (or still are) a Fallout fan.
Cerealkillerz - Julian Bieder - German - 8.8 / 10
On Paper Wasteland 3 sounds like the perfect RPG-Dream but the execution leaves much to be desired. Bugs, Glitches and graphics that doesn't really represent a game that releases and the end of this console generation are a bit of a letdown. Everything else from the great story, entertaining NPCs, solid battle system, clever leveldesign over to the love for details is amazing, besides some flaws that should soon be fixed, as inXile and Brian Fargo promise. Everyone that wasn't happy with the latest Fallout Games will surely love Wasteland 3.
Chicas Gamers - Adrián de Francisco - Spanish - Unscored
Wasteland 3 is a old-school role-playing game, with a compelling story, a combat system that promises but is not groundbreaking and some funny moments and black mood, which always remind us that we are in a post apocalyptic world, but with a smile. Don't forget the powerful character editor, rhythm voices, and the beautiful scenery that puts you in that atmosphere of cold and snowy Colorado.
Cram-Gaming - Robert Cram - 8.5 / 10
Wasteland 3 can be a bit of slog if you're gunning for marathon gaming sessions with it at the helm. Combat, whilst exciting initially can fall into the traps of repetition. A little more variety could have negated some of the repeated player actions. That said, the story is compelling and the characters an interesting assortment of misfit survivors, although perhaps fitting post-apocalyptic stereotypes. It's a fun, easy to play game overall though that should well-please fans of the series and keep players entertained for quite some time with its high replay-value. However, aside from some bugs here and there, the impressive amount of voice-work on offer, the character building is the best part of the experience where you can really nurture your ranger squad in this snowy post-apocalyptic world.
Digital Trends - Tom Caswell - 4 / 5 stars
Wasteland 3 is a rewarding game that offers unprecedented choice and is a great jumping on point for new players.
DualShockers - Kris Cornelisse - 9 / 10
Improving on its predecessor in almost every way, Wasteland 3 is one of the best and most reactive RPGs I've played in a long time.
EGM - Mollie L Patterson - Unscored
At least in my time with it, Wasteland 3 has been a fascinating experience. I’ve come to appreciate its depth of gameplay, character, building, and exploration, even if some of its pieces and parts still feel very foreign to me.
Entertainium - Eduardo Rebouças - Unscored
I will be even happier with Wasteland 3 once it’s patched and most of the bugs that bit me end up getting squashed. Even in its current state I’m having a grand ol’ time bringing some justice to the cold depths where no Ranger has dared to before. But for as much of a blast as I’m having out northeast in the cold, I hope I can make it back to sunny Arizona in time to save my fellow lawmen!
Eurogamer - Wesley Yin-Poole - Recommended
inXile's old-school RPG is the Fallout game we've been craving.
Fextralife - Castielle - 8.3 / 10
Wasteland 3 is a throwback to the old School RPGs of yesteryear, while providing a new combat experience and a bigger world. Players that liked previous Fallout Games, or games like Wasteland 2 or Baldur's Gate will feel right at home with this title, and will have the opportunity to try X-Com like combat. For the amount of content provided, 60 USD is a very good price, and fans of the genre should get more than their money's worth.
GAMES.CH - Nedžad Hurabašić - German - 83 / 100
Wasteland 3 is absolutely worth the money - the RPG brings dozens of hours of fun gameplay to the table. A must-buy for roleplayers.
Game Revolution - Jason Faulkner - 4 / 5 stars
Wasteland 3 is a marvel of a game, especially from a small studio like inExile. It’s not without its flaws, but the excellent writing and enthralling world overshadow those.
GameSkinny - Daniel Hollis - 9 / 10 stars
Wasteland 3 invokes feelings of classic RPGs such as Fallout and manages to nail the feel and tone perfectly in a modernized setting.
GameWatcher - Marcello Perricone - 8.5 / 10
A fantastic RPG that superbly mixes player choice and great combat to something bigger than the sum of its parts.
GamesRadar+ - Andrew King - 4 / 5 stars
Wasteland 3 doesn't bring much new to the table, both as a CRPG and as a piece of post-apocalyptic fiction. But, it's a terrifically executed role-playing game that rewards player investment from beginning to end.
GamingBolt - Ravi Sinha - 9 / 10
Wasteland 3 is a heady crescendo of post-apocalyptic story-telling. Its combat is compelling and fun while its characters and overall plot are engrossing, even when it goes to some dark places. A must-play for tactical RPG fans.
Gert Lush Gaming - Jim Smale - 9 / 10
Wasteland 3 is the defacto strategy experience and one that every gamer owes themself the pleasure of playing.
God is a Geek - Mick Fraser - 9.5 / 10
Wasteland 3 is a huge undertaking, marrying deep, choice-driven role play with fast-paced tactical combat and vast areas to explore.
IGN Spain - Álex Pareja - Spanish - 8 / 10
Wasteland 3 knows how to open to new players keeping the old school essence. It's not a revolution on the genre or in the post apocaliptic proposal, but it won't matter to the franchise lovers.
Niche Gamer - Cwb - 3.5 / 10
We’ll update this review if the game is fixed, and the issues outlined are fixed or at least addressed; and then I’ll pick it back up. As it stands now, I’ll be playing something else that isn’t as apt to crash. Buyer beware.
PC Gamer - Jody Macgregor - 84 / 100
A wilfully strange setting explored through a predictable but enjoyable old school RPG thats been streamlined just enough.
PC Invasion - Jason Rodriguez - 8.5 / 10
There are a few misgivings related to Wasteland 3's technical aspects, mechanics, and overall challenge. However, its cast of characters (both old and new), the switch to a traditional turn-based combat system, and branching paths filled with decisions and dire consequences make for a superb journey with the Desert Rangers.
PCGamesN - Gina Lees - 9 / 10
Lurid characters, a deep RPG system, and captivating combat set in an unhinged apocalypse - inXile Entertainment's latest shouldn't be missed.
Player2.net.au - Matt Hewson - A or higher
With a focus on freedom of choice that is second-to-none, Wasteland 3 has set the benchmark for CRPG narratives, all the while being supported by wonderfully engaging gameplay and roleplaying mechanics.
PowerUp! - Leo Stevenson - 9.7 / 10
If you’re an RPG fan, a Fallout fan or even just a videogame fan, do yourself a favour and play one of this year’s very best games; Wasteland 3.
Saving Content - Scott Ellison II - 5 / 5 stars
It took me a while to realize how much these interactions, whether it be the interpersonal conversation or combat encounters themselves, stuck with me. Wasteland 3 has rules, but they only exist for you to bend them. With limitless character creation combinations, branching dialogue choices that affect what quests you do or don’t experience, and multiple endings, Wasteland 3 is an expanse of content and opportunity. The change in locale does wonders, no longer relying on a tired post-apocalyptic biome. Wasteland 3 has a wonderful backdrop in Colorado’s frozen wastes, making it the perfect place to spend a nuclear winter.
Screen Rant - Christopher Teuton - 4 / 5 stars
Wasteland 3 takes players to a new location and presents them with equally unfamiliar challenges, yet still perfectly demonstrates all of the reasons why this series has had die-hard fans for over three decades, and is absolutely worth playing for anyone looking for their next post-apocalyptic fix.
Shacknews - Josh Hawkins - 9 / 10
If you’re a big fan of the original Wasteland games, or just an RPG fan in general, then I highly recommend picking up Wasteland 3 and giving it a try.
Spaziogames - Paolo Sirio - Italian - 8.3 / 10
Wasteland 3 doesn't change its predecessor's successful formula but, outside of certain design limitations, it perfects and modernizes it. It's easily the best game in the franchise, in terms of pure technique, and one that clearly gives you an idea of what inXile is able to achieve.
The Games Machine - Danilo Dellafrana - Italian - 8.7 / 10
Wasteland 3 is a good role-playing game, technically passable but enriched by a dense network of intriguing subplots that will push the most dedicated to play it several times. Watch out for the ever-present release bugs, though – best to wait a couple patches if you want to avoid unnecessary hurdles.
TrustedReviews - Alastair Stevenson - 4 / 5 stars
Wasteland 3 is a solid tactical RPG that will keep fans of the genre entertained for hours upon hours. But it doesn't do enough to bring the genre forward to a mainstream audience.
WayTooManyGames - Thomas Medina - 9 / 10
All in all, this is the game I wanted so badly for Wasteland 2 to be. It doesn’t just repeat what came before, but expands upon it all. Not just mechanically, but story wise as well.
Wccftech - Francesco De Meo - 9 / 10
Wasteland 3 features everything only the best role-playing games do: an engaging story powered by excellent writing, compelling characters, tons of customization options, and a deep tactical combat system that feels fresh even after dozens of hours. But, most of all, it features a living world that reacts to what the player does, and changes depending on how the player decides to deal with the troubles ahead, providing a role-playing experience of the highest degree, one that very few games can boast of.
Windows Central - Jez Corden - 5 / 5 stars
Wasteland 3 is a testament to the power of the branching narrative, taking it far beyond binary choices and into a grand canopy of cause and effect. It gives the wintry climbs of Colorado a lifelike quality that must have been painstaking to build. The most impressive RPG in years, Wasteland 3 is a masterpiece.
XboxEra - Jesse Norris - 9.7 / 10
Wasteland 3 shines with clear dedication to crafting the best game its genre has ever seen. Excellent visuals are matched by top notch voice work and some of the best and most natural writing I have seen in a video game not made by Naughty Dog. The combat is a brutal dance where one wrong move can spell disaster, but victory is an exhilarating rush that never becomes old. Wasteland 3 cements inXile as one of the best in the business in the RPG genre and affirms that Xbox has something truly special on their hands.
submitted by diogenesl to Games [link] [comments]

Over-Optimizing for Performance

Recently on the csharp subreddit, the post C# 9.0 records: immutable classes linked to a surprisingly controversial article discussing how C# 9.0's records are, underneath it all, immutable classes. The comments are full of back-&-forth over whether one should use records for ease or structs for performance. The pro-struct argument revolved around the belief that performance should always be a developer's #1 priority, and anything less was the realm of the laggard.
Here is a real-world example that shows with stark clarity why that kind of thinking is wrong.
Consider the following scenario:

1

You're working on a game with dozens, maybe hundreds of people on the team; you don't know because when you were cross with facilities about them removing all the fluorescents, you got accused of being against the new energy saving initiative. Now you swim in a malevolent ocean of darkness that on some very late nights alone in the office, you swear is actively trying to consume you.
 

2

The team that preceded you inherited an engine that is older than OOP, when source repositories were stacks of 8-inch floppies, and it looked as if Jefferson Starship was going to take over the world. One year ago they bequeathed upon the company this nightmare of broken, undocumented GOTO spaghetti & anti-patterns. You're convinced this was their sadistic revenge for all getting fired post-acquisition.
 

3

Management denied your request to get headcount for an additional technical artist, but helpfully supplied you with an overly nervous intern. After several weeks working alongside them, you're beginning to suspect they're pursuing something other than a liberal arts degree.
 

4

Despite the many getting started guides you spent countless evenings writing, the endless brownbags nobody attended, and the daily dozen emails you forward to oppressively inquisitive artists comprised of a single passive-aggressive sentence suggesting they scroll down to the part that begins FW: FW: FW: FW: FW: FW: RE: WE BROKE TOOL NEED WORKAROUND ASAP ...
 
...yes, despite all of that, the engineering team still spent days tracking down why the game kept crashing with Error 107221: У вас ошибка after re-re-re-re-re-throwing an ex_exception when it couldn't (and should never even try to) load a 16K-textured floor mat.
 

5

Despite your many attempts to politely excuse yourself, one blissfully unaware artist exhausts 48 minutes of your lunch break explaining how the Pitchfork review for the latest "dope slab" of this TikTok-Instagram-naphouse artist you never heard of was just sooooo unfair.
 
And then in their hurry to finish up & catch the 2:30 PM bus home, they forget to toggle Compress To CXIFF (Custom Extended Interchange File Format), set the Compression slider 5/6ths of the way between -3 & -2, look to their left, look to their right, click Export As .MA 0.9.3alpha7, and make absolutely, positively, 100% SURE not to be working in prod. And THAT is how the game explodicated.
 

6

You know better than anyone the intermediate file format the main game loop passes to Game.dll, memory mapping it as a reverse top-middle Endian binary structure.
 
You know for 381 of the parameter fields what their 2-7 character names probably mean.
 
YOU know which 147 fields always have to be included, but with a null value, and that the field ah_xlut must ALWAYS be set to 0 unless it's Thursday, in which case that blackbox from hell requires its internal string equivalent: TRUE.
 
YOU know that the two tech artists & one rapidly aging intern that report to you would totally overhaul tooling so artists would never "happen" again, but there just aren't enough winters, springs, summers, falls, July 4ths, Christmas breaks, Presidents Days, and wedding anniversaries in a year to properly do so.
 

7

If you could just find the time between morning standups, after lunch standups, watersprint post-mortems, Milbert's daily wasting of an hour at your desk trying to convince you engineering should just rebuild the engine from the ground up in JavaScript & React, & HR's mandatory EKG Monitor job satisfaction surveys, you might be able to get at least some desperately-needed tooling done.
 
And so somehow you do. A blurry evening or two here. A 3:00 AM there. Sometimes just a solitary lunch hour.
 
Your dog no longer recognizes you.
 
You miss your wife calling to say she's finally cleaning out the hall closet and if you want to keep this box of old cards & something in plastic that says Underground Sea Beta 9.8 Grade, you better call her back immediately.
 
And your Aunt Midge, who doesn't understand how SMS works, bombards you one evening:
your father is...
no longer with us...
they found him...
1 week ago...
in an abandoned Piggly Wiggly...
by an old culvert...
split up...
he was then...
laid down to rest...
sent to St. Peter's...
and your father...
he's in a better place now...
don't worry...
it's totally okay...
we decided we will all go...
up to the mountain
 
You call your sister in a panic and, after a tidal wave of confusion & soul-rending anxiety, learn it was just Hoboken Wireless sending the messages out of order. This causes you to rapidly cycle.
 

8

On your bipolar's upswing, you find yourself more productive than you've ever been. Your mind is aglow with whirling, transient nodes of thought careening through a cosmic vapor of invention. It's like your brain is on 200mg of pure grade Adderall.
 
Your fingers ablaze with records, clean inheritance, beautiful pattern matching, bountiful expression syntax, aircraft carriers of green text that generate the most outstanding CHM for an internal tool the world has ever seen. Readable. PERFECTLY SOLID.
 
After much effort, you gaze upon the completed GUI of your magnum opus with the kind of pride you imagine one would feel if they hadn't missed the birth of their son. Clean, customer-grade WPF; tooltips for every control; sanity checks left & right; support for plugins & light scripting. It's even integrated with source control!
 
THOSE GODDAMNED ARTISTS CAN'T FAIL. YOUR PIPELINE TOOL WON'T LET THEM.
 
All they have to do is drag content into the application window, select an options template or use the one your tool suggests after content analysis, change a few options, click Export, and wait for 3-5 minutes to generate Game.dll-compatible binary.
 
Your optimism shines through the commit summary, your test plan giddy & carefree. With great anticipation, you await code review.
 

9

A week goes by. Then two. Then three. Nothing. The repeated pinging of engineers, unanswered.
 
Two months in you've begun to lose hope. Three months, the pangs of defeat. Four months, you write a blog post about how fatalism isn't an emotion or outlook, but the TRANSCENDENCE of their sum. Two years pass by. You are become apathy, destroyer of wills.
 

10

December 23rd, 2022: the annual Winter Holidays 2-hour work event. The bar is open, the Kokanee & Schmidt's flowing (max: 2 drink tickets). The mood a year-high ambivalent; the social distancing: acceptable. They even have Pabst Blue Ribbon, a beer so good it won an award once.
 
Standing beside you are your direct reports, Dave "Macroman" Thorgletop and wide-eyed The Intern, the 3 of you forming a triumvirate of who gives a shit. Dave is droning on & on about a recent family trip to Myrtle Beach. You pick up something something "can you believe that's when my daughter Beth scooped up a dead jellyfish? Ain't that something? A dead jellyfish," and "they even had a Ron Jons!"
 
You barely hear him, lost as you are in thought: "I wish I had 2 days of vacation." You stare down ruefully at your tallboy.
 
From the corner of your eye you spot Milbert, index finger pointed upward, face a look of pure excitement.
 
"Did I tell you about my OpenWinamp project? It's up on SourceForge", he says as he strides over. It's unsettling how fast this man is.
 
"JAVASCRIPT IS JUST A SUBSET OF JAVA!" you yell behind you, tossing the words at him like a German potato masher as you power walk away. It does its job, stopping Milbert dead in his tracks.
 
Dave snickers. The Intern keeps staring wide-eyed. You position yourself somewhat close to the studio's 3 young receptionists, hoping they serve as a kind of ritual circle of protection.
 
It works... kind of. Milbert is now standing uncomfortably close to The Intern, Dave nowhere to be seen.
 
From across the room you distinctly hear "Think about it, the 1st-person UI could be Lua-driven Electron."
 
The Intern clearly understands that words are being spoken to them, but does not comprehend their meaning.
 
You briefly feel sorry for the sacrificial lamb.
 

11

You slide across the wall, putting even more distance between you & boredom made man. That's when you spot him, arrogantly aloof in the corner: Glen Glengerry. Core engineering's most senior developer.
 
Working his way up from a 16-year old game tester making $4.35 an hour plus free Dr. Shasta, to pulling in a cool $120K just 27-years later, plus benefits & Topo Chicos. His coding style guides catechism, his Slack pronouncements ex cathedra; he might as well be CTO.
 
You feel lucky your team is embedded with the artists. You may have sat through their meetings wondering why the hell you should care about color theory, artistic consistency, & debates about whether HSL or CMYK was the superior color space (spoiler: it's HSL), you were independent and to them, a fucking code wizard, man.
 
And there he stands, this pseudo-legend, so close you could throw a stapler at him. Thinning grey-blonde tendrils hanging down from his CodeWarrior hat, white tee with This Guy VIMs on the back, tucked into light blue jeans. He's staring out into the lobby at everything and yet... nothing all at.
 

12

Maybe it's the 4.8% ABV. Maybe it's the years of crushing down anger into a singularity, waiting for it to undergo rapid fiery expansion, a Big Bang of righteous fury. Maybe it's those sandals with white socks. Maybe it's all three. But whatever it is, it's as if God himself compels you to march over & give him a piece of your mind, seniority be damned.
 
"Listen, you big dumb bastard..."
 
That... is maybe a little too aggressive. But Glen Glengerry barely reacts. Pulling a flask out of his back pocket, he doesn't look over as he passes it to you.
 
Ugh. Apple Pucker.
 

13

"I thought bringing in your own alcohol was against company policy", wiping sticky green sludge from your lips. He turns with a look of pure disdain & snorts.
 
"You think they're going to tell ME what I can & can't bring in?" He grabs the flask back, taking a big swig.
 
For what feels like an eternity, you both stand in silence. You swallow, speaking softly. "None of you even looked at my code. I worked very, very hard on that. My performance review for that year simply read 'recommend performance improvement plan." The words need no further context.
 
"I know", Glen² replies. "That was me."
 

14

Now you're not a weak man, and maybe in some other circumstance you would have punched him in the goddamn lip. But you feel nothing, just a hollowness inside. "Why?", you ask, wondering if the answer would even matter.
 
"Because you don't use Bulgarian notation. Because your method names aren't lower camel case. Because good code doesn't require comments. Because you use classes & records over more performant structs, pointlessly burdening the heapstack. BECAUSE. YOUR CODE. IS. SHIT."
 
You clinch your fists so tightly the knuckles whiten.
 

15

He looks away from you, taking another sip of green goo. "You're not a coder. You're an artist masquerading as one" he speaks, as if it were fact.
 
The only thing artistic about you is the ability to create user-friendly internal tooling using nothing but a UI framework, broken down garbage nobody wants to touch, & sheer willpower. If your son's life depended on you getting accepted into art instruction school, you couldn't even draw a turtle.
 
He doesn't pause. "I'll champion ruthless micro-optimization until the day I die. But buddy, I'm going to let you in on a little secret: you aren't here to improve workflow. You're here to LOOK like you're doing something NOBODY else can."
 
He goes on. "What do you think those artists are going to do when they have to stare at a progress bar for 4, 5 minutes? They're going to complain your tool is slow."
 
"Sure, it may take them 20, 30 minutes to do it the old way, there'll be an error, and either they'll stare at it for 30 minutes before adding that missing semi-colon or they'll come get you. And you'll fix it. And 1 week later, they won't remember how. And you'll stay employed. And every. Body. Wins."
 

16

A little bit of the pride, the caring, wells back up inside from somewhere long forgotten.
 
"You don't think we should care about rapid application development & KISS, quickly getting things out that help our team, instead devoting ourselves to shaving off ticks here & there? What do you think artists are going to do with those 4 minutes you talk about?
 
You don't stop. "I'll tell you what they'll do. They'll 9GAG for 20 minutes straight. They'll listen to podcasts about dialectical materialism vis-a-vis the neo-feudalism that is a natural extension of the modern world's capitalist prison. They'll Reddit."
 
His silence gives you the bravery to push the limits.
 
"Christ, man. Are you only in it for the $120K..."
 
He corrects you: "...$123K."
 
"...only in it for the $123K/year? The free snacks from the microkitchen? The adulation? Have you no sense of comraderie?? No desire to push us to something better?! No integrity?!!!"
 
His eyes sharply narrow, face creases in anger. You clearly have overstepped your bounds.
 

17

"You think I don't have integrity? No sense of teamwork? I'm only in it for the cold cash? You think I don't care about you all?", he roars.
 
A light volley of small green flecks land on your face.
 
"Why do you think they made a 16-year old tester the lead developer of a 1993 Doom clone?! Because my code was clean & painless to work with?! Because I made coding look easy?! No! IT WAS BECAUSE I WAS A GOD TO THEM.
 
And from a God, a PANTHEON. We built monuments to over-engineering! We crafted that of 7 weeks onboarding, that of immortal bugs, demonic hosts spawned by legion from the very loins of a fix. It took 2 years before a developer could BEGIN to feel confident they knew what they were doing. And by that time, they were one of US!
 
You think the team we laid off November '19 was fired because they were bad at their jobs? NO! It was because they worked themselves out of one. They didn't leave us a broken pipeline. They left an internal Wiki, a wealth of tools & example projects, and a completely transparent code base.
 
We couldn't have THAT, now could we? No, we couldn't. So we got rid of it. ALL OF IT. Poof. Gone. Just like that. Before anyone even knew a THING."
 
He leans forward, so close his psoriasis almost touches yours.  
With an intensity that borders on frightening, he whispers "You think they left us Game.dll? I fucking *MADE** Game.dll."*
 
The words hit hard like a freight train.
 

18

And without another word, he turns & leaves. You're left there, alone, coworkers milling about, with only one thought.
     
Were one to get a hobby, should it be cocaine?
 

In Conclusion

It's these kinds of situations that make me believe there are far more important considerations than a ruthless dedication to performance, even in the game industry as my real-world scenario so clearly demonstrates.
 
Like, records are cool & shit.
submitted by form_d_k to shittyprogramming [link] [comments]

The Meta Unicorn - Diana Invoke

The Meta Unicorn - Diana Invoke

(CEBASAYJC4RTQP2JJRKFMYACAECQCKACAEBQKAQCAMEQKSYBAEBQSXQ)

A Pre Note

I am not a top-100 player, as given the inevitable RNG of effects, matchups and draws in card games, and the lack of rewards for ranking up in LoR, I simply don't see the point in painstakingly grinding up to such a level. The highest I've climbed is low Diamond, but considering the above knowledge, I believe that at a certain skill level (perhaps at around Platinum), it's more about how much time one can put in than how skilled they are. HOWEVER, considering all this, I have the absolute conviction that this deck is a top-100 worthy deck.

Introduction

This is the only deck that I've played since Day 1 of Call of the Mountain, with various modifications, and I believe that it is a completely undiscovered meta unicorn. I've never faced a similar deck on ladder, and my deckbuilding experiments with any other archtypes have left me completely unsatisfied with the lack of interaction and agency, as well as the sheer counterability of the vast majority of tools currently out there.
A lot of people are frustrated with the current meta - a lot of points of which are covered by BruisedByGod in his recent video critique. To summarize his main points:
  1. Most answers are completely outclassed by threats
  2. Sheer lack of healing options locks out deckbuilding choices
  3. Most top-tier strategies prey on lack of interactivity (Pirate Burn, Lee Sin OTK, Star Spring)
This is a Control deck which, while originally devised to prey on the inevitably popular Aurelion Sol and Troll Chant and abuse the broken, flexible toolbox of Invoke on Day 1, also manages to both answer all 3 of these problems efficiently.

Card Choices

Early Tempo/Nightfall

Simply the best available early-game that an Invoke Targon deck could hope to muster - Diana functioning as both early game and late-game removal (we have just enough Nightfall Synergy) for practically no investment, Pale Cascade being legitimately one of the most broken cards currently in the game, and the ping cards also serving a modicum of uses at all stages of a match.
Spacey Sketcher has been severely underrated so far - providing critical tools for certain matchups and/or providing early game minions without needing to actually run them (a fundamental weakness of faster decks top-decking late). Its 'discard-replace' synergy with our late-game, as well as Duskpetal Dust and meta-call flex cards is just icing on the cake.
Finally, note how every early game card I've chosen scales well and still plays a role as the game goes later; as removal, Elusive blocking, tool-building, Burst-speed Nightfall, pings and cantrip Combat Tricks. This is an often overlooked but fundamental difference between Control early-drops, and aggro early-drops (such as Precious Pet).
~

Removal

These two cards, combined with any generated Obliterates, form the only proper removal this deck has - and were the catalyst for me creating this deck in the first place. All three of these removal types leave almost NO room for the opponent to interact with them, and I believe that is the sole condition for a high-cost removal spell to be playable in the current game state.
NOTE: Ruination is easily and always played around at a high-level of play - and leaves the opponent with ALL of the agency/choice to play around it/bait it exactly how they wish, instead of you (whose only options are to play the card too early and get out-tempo'd afterward, use more than 3 mana elsewhere to catch-up at which point it becomes unplayable, or lose the game to a sudden-attack completely at your opponent's discretion) - the ultimate NO-NO for this deck: I never even considered putting it in.
~

Meta Call Flex Spots

At times I feel as if this card could be cut to 1 copy, but right now 2 feels great against the current meta, and drawing into at least one is almost necessary in order to compete with Star Spring (Obliterate is conditional and too great a tempo loss early on). In other metas previously, I've experimented with 1 copy of Passage Unearned, as well as 2 extra copies of Lunari Shadestalker.
~

Literally Everything Else One Could Ever Hope to Need

I still believe that Invoke is one of the most broken mechanics currently in the game. This is one of the heaviest late-game decks I can possibly imaginable, yet the only cards above 5-mana we run are removal, and our mid-game minions and healing straight up provide whatever early OR late-game tools we might possibly need in any matchup - it's simply overly flexible (flexilibity in card games being a MUCH bigger deal than most people give it credit for) and not enough of a tempo/stat sacrifice IMO. I think that Invoke as a mechanic is even stronger when ran in bulk, and especially in a Control deck - as the game goes on slowly you generate a toolbox that can handle just about any dynamic situation that meta decks can throw your way.
The spell-mana nerf to Living Legends has balanced it out quite a bit, however the same-nerf to Cosmic Inspiration still hasn't convinced me that it isn't in the top 5 least healthy effects that a game based on carefully stat-balanced of minion trading could ever have (hit me up with your Cosmic Inspiration hate!) - a large proportion our games are won by this disgusting effect.
Solari Priestess and Starshaping need no introduction as some of the most popular, utilitarian Invoke cards, however Mountain Scryer and Moondreamer (not so much Lunari Priestess) really put in the work, and I've never seen anyone else play these cards. The former provides crazy mana-advantage as the game goes on given our huge focus on Celestials (it's a shame we can't afford to push its Invoke chances even higher), and the latter has juuussst the right stat distribution at 3/5 to blockade most midgame tempo plays out opponent might go for.
NOTE: Aurelion Sol is straight up unnecessary to compete late-game, is always burdensome and clunky draw, ruins our surprise factor (though that doesn't exist anymore with this post being made), and we often outvalue decks running him anyway (don't forget that the original premise of this deck was 'How can I best remove Aurelion?').
~

Matchups/Strategy (Order Based on Mobalytics Tier List)

Lee Sin (60/40)

A somewhat favored matchup - although more recent lists that have cut Bastion in favor of Nopify may be a bit more in their favor (a proper Ping Counter). Hard mulligan for Spacey Sketcher, Sunburst and our pings. Generating Silence (Equinox) for Mentor of the Stones/Zenith Blade is our main early game goal. Our Mid-to-Late game goal is removing all 3 Lee Sin's at the expense of practically everything else (the rest of their deck is pretty much completely irrelevant, but rushing them down is also pretty much impossible) - after which our win is basically guaranteed.
~

Swain/TF (80/20)

I believe that we are very, very heavily favored if played properly (although it's a VERY nuanced matchup to play right), and most of our losses come from bricking our early-game draws and/or not drawing/generating a single Starshaping/Golden Sister as their burn damage inevitably builds up. Hard mulligan for all 1/2 cost cards (only keep 1 Pale Cascade with a 1/2 cost minion).
~

Pirate Aggro (55/45)

We are much more prone to bricking on draws here than Swain/TF, as we need quite a specific hand to deal with their onslaught - This is probably our most draw-dependent, low-agency matchup by far - as face-deck matchups tend to be. In addition - Captain Farron is much more effective against our removal strategy than the likes of Leviathan. Nonetheless, from my experience I think that we're still every-slightly-so favored in this matchup - often winning by the skin of our teeth. Starshaping/Golden Sister are mandatory late-game, and not bricking by not drawing/generating either is also basically a loss. Hard mulligan for all 1/2 drops, and keep a single Sunburst for Gangplank if your hand is already looking great.
~

Warmother's (25/75)

A very unfavored and binary matchup (see below as to why) that has luckily become rarer recently. Mulligan for Removal/Invoke cards.
~

Trundle/Asol (75/25)

This deck was basically created on Day 1 specifically to destroy Trundle/Asol. Sadly though, even at 75/25 the matchup is worse than it should be due to the nature of Invoke RNG - if one player draws into Cosmic Inspiration and the other didn't the match is over, full stop + the occasional shenanigans involving The Great Beyond uninteractibly going face and non-stop Living Legends value. Mulligan for Sunburst, Vengeance and pings.
~

Discard Aggro (80/20)

I don't know why this deck is considered competitive - maybe because our matchup here is basically as favored as TF/Swain except without any gameplay nuance required on our part. Mulligan for 1-2 drops. Keep Solari Priestess/Sunburst if hand is good. Only necessary statistical losses to bad early draws against an aggro archtype.
~

Fiora/Shen (70/30)

Another draw dependent, but quite favored matchup. Quite difficult to play though - you need to balance maintaining some modicum of tempo whilst also being able to deal with their crucial threats. Mulligan for 1-2 drops ESPECIALLY Pale Cascade/Pings, and Removal.
~

Scouts (60/40)

Basically the Pirate Aggro matchup but a tad bit slower and with no burn - giving you more leeway to make up for bad draws both early and late.
~

Leona/Lux (80/20)

Basically the Trundle/Asol matchup except with no 'must remove ASAP' threats giving you more leeway to make up for bad draws. Celestial RNG and especially Cosmic Inspiration still give them a chance to win as usual.
~

Shyvana Dragons (50/50?)

I surprisingly, haven't faced too much of this deck yet personally, but looking at it's cards compared to ours, I think the matchup would be about 50/50 (an otherwise favourable looking matchup affected a bit by their high tempo removal and guaranteed Cosmic Inspiration in the form of Kadregrin).
~

Ashe/Sejuani (70/30)

This matchup is dependent on whether we draw removal for Ashe somewhat on curve, how much tempo they manage to build early on and whether we draw good enough to afford to play around Reckoning. Mulligan for Sunburst, Solari Priestess, Pings and Diana (only if you've already drawn support) as our other standard early drops are all pretty ineffective against theirs.
~

Endure (85/15)

Probably our most favored meta-deck matchup, and unfortunately rarer recently. Their win conditions - Kalista, Blighted Caretaker tempo, Neverglade Collector and They Who Endure simply don't stand a chance against our toolbox. Most losses come from unanswered Blighted Caretaker tempo. Mulligan for Spacey Sketcher, Sunburst and Pale Cascade.
~

Deep (0/100)

The biggest downside and sheer impossible matchup of this archtype. Maokai manages to pack even less interactivity/inevitability than we do, and the nature of our deck gives us no chance of out-tempoing Deep early OR late. Auto-concede.
~

Diana/Nocturne (75/25)

A simpler aggro matchup than the others. Mulligan for 1-2 drops - especially Spacey Sketcher and Diana, as well as Sunburst.
~

Tahm-Kench/Soraka (70/30?)

Another matchup that I haven't faced too much of just yet. Mulligan hard for Divergent Paths and Solari Priestess - Once we remove their uninteractive element trump-card in the Landmark win-condition, if we can survive their early tempo, the rest of the match should be a cinch given our heal/health-ignoring conditionless removal for their Champions.
~

Conclusion

Thanks for reading up to this point, and pardon my formatting, the ridiculous length and the sheer pomposity of it all.
I still think Invoke is flexible to the point of being broken and the only reason the matchup spread is so good. I also think that with the release of this guide - more people will come to recognise this archtype and the element of surprise affecting enemy mulligans against an assumed more aggro, Nightfall-focused Diana archtype will be lost. People will also know to play around less common cards such as Sunburst, and I expect winrates to fall somewhat across the board.
To conclude this guide, I'd like to say that this is this is not a healthy deck. At the deepest level, this deck is fundamentally about removing agency from your opponent and giving it to yourself, as well as securing the critical boon of having inevitability over your opponent in a game with the nature of LoR. If all decks were like this, LoR would completely cease to be fun.
What else do I think is unhealthy right now? - Simple: anything removing interactivity from your opponent - ESPECIALLY as a win condition; Maokai, Star Spring, Cosmic Inspiration, Lee Sin. The avenues through which these cards can be interacted with are way too limited right now.
A lot of the metagame nowadays is about having an uninteractable win condition, or focusing damage to face so fast the opponent has no chance to react - another form of non-interactivity. Here's hoping that the meta in the near future heads back in the direction of the close but fair midrange board battles we all came to love back in vanilla LoR.
~
(slinx4)
submitted by poklipart to LoRCompetitive [link] [comments]

What is the quickest solution to finding a 4 digit number asking only yes/no questions?

A friend and I were watching a Korean game show called "The Genius", and in it they had a particularly brilliant competitive maths game.
The premise was fairly simple - Each contestant had to pick a four digit number. They then were allowed to ask questions to each other one after the other, or use a turn to guess what the opponent's number was. The only additional rule was that 0 was treated as even for the purposes of questioning.
After watching this, my friend and I tried to come up with a solution to guarantee finding your opponent's number in the fewest possible questions, but it very quickly got extremely complex. However, we're both fairly sure that there's a clever mathematical answer to guarantee it in a low amount of questions.
After working out the number you need to use a turn to guess, so once the number is worked out a +1 needs to be added. (This isn't important if you know the number, but I figure can be relevant if you can get down to say 3 potential options, since just guessing all 3 is as efficient as working out which one it is)
The obvious first approach we tried was binary searching the numbers for each digit. With this method, each digit could be found in a maximum of 4 questions (10 to 5 to 3 to 2 to 1), so we knew that we needed to try and beat 16 questions.
We then realised that if we treated the first two digits and the second two digits as two digit numbers, it would only take a maximum of 7 questions to find each digit pair (100 to 50 to 25 to 13 to 7 to 4 to 2 to 1), so we were down to 14 questions.
Following the theme, we tested binary search on all 4 digits, but realised it would take 14 questions (10000 to 5000 to 2500 to 1250 to 625 to 313 to 157 to 79 to 40 to 20 to 10 to 5 to 3 to 2 to 1), resulting in 14. It was no more efficient than the two sets of digits, and was also harder to calculate.
I then tried a set of 3 digits and a single, (1000 to 500 to 250 to 125 to 63 to 32 to 16 to 8 to 4 to 2 to 1) + 4 for the remaining digit, and again this was 14.
I then proposed a different solution - Could we potentially get more information by adding the digits together? I tried it on a number he had picked, asking questions to do a binary search on the sum of pairs of digits. Assigning the letters abcd to the four digits, I worked out a+b, c+d, b+c, and a+d. I figured doing this would allow me to arrange the numbers correctly once I had crunched it out.
Since the number was a 4 digit number, I knew the maximum the total for all four numbers could reach was 36. That meant each pair was a maximum of 18, and sets of pairs had to add up to no more than 36.
I started off by binary searching the sum of the first two digits, which would take a maximum of 5 turns (18 to 9 to 5 to 3 to 2 to 1), and repeated for the second two. That would take 10 turns, but give me the sum of all digits, the sum of the first two, and the sum of the second two.
At this point I was adamant that I could potentially figure out the number using this information alone, but I was unsuccessful. I was able to use logic to narrow down the possible values for the outer two and inner two numbers by figuring out the total of all four digits, figuring out the number pair combinations a/b and c/d could be to satisfy that, and then working out the potential pairs of values that the inner two and outer two numbers could add up to.
In our example test, a+b was 5, c+d was 14. From this I knew the total was 19. The first pair of numbers had to be 0 and 5, 1 and 4, or 2 and 3. The second pair had to be 9 and 5, 8 and 6, or 7 and 7. Using the logic of adding the highest number of one pair to the highest number of the other pair and then cycling through the values, I worked out that the outer two numbers had to add up to 14, 13, 12, or 11, and the inner two numbers had to add up to 8, 7, 6, or 5.
Binary searching these could be done in two searches each, bringing the total to... 14. :(
This is where it turned a bit weird though - After doing some logic on the resulting numbers (the outer pair was 12 and the inner pair was 7), I came up with three potential answers that satisfied every single constraint.
5077 4168 3259
These three numbers are amazing. The first two digits add up to 5, the third and fourth add up to 14, the first and fourth add up to 12, and the second and third add up to 7.
Unfortunately from here there was no choice but to guess all three, no amount of questioning could lower it from three to two questions. And so our final total was 17, no more efficient than just binary searching the numbers in the first place.
And so, I ask you this - Is there a more efficient, human doable way to discover the four digit number than binary searching the first pair and second pair of digits?
I feel like there has to be, but I'm not knowledgeable enough to know!
submitted by Jademalo to mathematics [link] [comments]

No gods, no kings, only NOPE - or divining the future with options flows. [Part 3: Hedge Winding, Unwinding, and the NOPE]

Hello friends!
We're on the last post of this series ("A Gentle Introduction to NOPE"), where we get to use all the Big Boy Concepts (TM) we've discussed in the prior posts and put them all together. Some words before we begin:
  1. This post will be massively theoretical, in the sense that my own speculation and inferences will be largely peppered throughout the post. Are those speculations right? I think so, or I wouldn't be posting it, but they could also be incorrect.
  2. I will briefly touch on using the NOPE this slide, but I will make a secondary post with much more interesting data and trends I've observed. This is primarily for explaining what NOPE is and why it potentially works, and what it potentially measures.
My advice before reading this is to glance at my prior posts, and either read those fully or at least make sure you understand the tl;drs:
https://www.reddit.com/thecorporation/collection/27dc72ad-4e78-44cd-a788-811cd666e32a
Depending on popular demand, I will also make a last-last post called FAQ, where I'll tabulate interesting questions you guys ask me in the comments!
---
So a brief recap before we begin.
Market Maker ("Mr. MM"): An individual or firm who makes money off the exchange fees and bid-ask spread for an asset, while usually trying to stay neutral about the direction the asset moves.
Delta-gamma hedging: The process Mr. MM uses to stay neutral when selling you shitty OTM options, by buying/selling shares (usually) of the underlying as the price moves.
Law of Surprise [Lily-ism]: Effectively, the expected profit of an options trade is zero for both the seller and the buyer.
Random Walk: A special case of a deeper probability probability called a martingale, which basically models stocks or similar phenomena randomly moving every step they take (for stocks, roughly every millisecond). This is one of the most popular views of how stock prices move, especially on short timescales.
Future Expected Payoff Function [Lily-ism]: This is some hidden function that every market participant has about an asset, which more or less models all the possible future probabilities/values of the assets to arrive at a "fair market price". This is a more generalized case of a pricing model like Black-Scholes, or DCF.
Counter-party: The opposite side of your trade (if you sell an option, they buy it; if you buy an option, they sell it).
Price decoherence ]Lily-ism]: A more generalized notion of IV Crush, price decoherence happens when instead of the FEPF changing gradually over time (price formation), the FEPF rapidly changes, due usually to new information being added to the system (e.g. Vermin Supreme winning the 2020 election).
---
One of the most popular gambling events for option traders to play is earnings announcements, and I do owe the concept of NOPE to hypothesizing specifically about the behavior of stock prices at earnings. Much like a black hole in quantum mechanics, most conventional theories about how price should work rapidly break down briefly before, during, and after ER, and generally experienced traders tend to shy away from playing earnings, given their similar unpredictability.
Before we start: what is NOPE? NOPE is a funny backronym from Net Options Pricing Effect, which in its most basic sense, measures the impact option delta has on the underlying price, as compared to share price. When I first started investigating NOPE, I called it OPE (options pricing effect), but NOPE sounds funnier.
The formula for it is dead simple, but I also have no idea how to do LaTeX on reddit, so this is the best I have:

https://preview.redd.it/ais37icfkwt51.png?width=826&format=png&auto=webp&s=3feb6960f15a336fa678e945d93b399a8e59bb49
Since I've already encountered this, put delta in this case is the absolute value (50 delta) to represent a put. If you represent put delta as a negative (the conventional way), do not subtract it; add it.
To keep this simple for the non-mathematically minded: the NOPE today is equal to the weighted sum (weighted by volume) of the delta of every call minus the delta of every put for all options chains extending from today to infinity. Finally, we then divide that number by the # of shares traded today in the market session (ignoring pre-market and post-market, since options cannot trade during those times).
Effectively, NOPE is a rough and dirty way to approximate the impact of delta-gamma hedging as a function of share volume, with us hand-waving the following factors:
  1. To keep calculations simple, we assume that all counter-parties are hedged. This is obviously not true, especially for idiots who believe theta ganging is safe, but holds largely true especially for highly liquid tickers, or tickers will designated market makers (e.g. any ticker in the NASDAQ, for instance).
  2. We assume that all hedging takes place via shares. For SPY and other products tracking the S&P, for instance, market makers can actually hedge via futures or other options. This has the benefit for large positions of not moving the underlying price, but still makes up a fairly small amount of hedges compared to shares.

Winding and Unwinding

I briefly touched on this in a past post, but two properties of NOPE seem to apply well to EER-like behavior (aka any binary catalyst event):
  1. NOPE measures sentiment - In general, the options market is seen as better informed than share traders (e.g. insiders trade via options, because of leverage + easier to mask positions). Therefore, a heavy call/put skew is usually seen as a bullish sign, while the reverse is also true.
  2. NOPE measures system stability
I'm not going to one-sentence explain #2, because why say in one sentence what I can write 1000 words on. In short, NOPE intends to measure sensitivity of the system (the ticker) to disruption. This makes sense, when you view it in the context of delta-gamma hedging. When we assume all counter-parties are hedged, this means an absolutely massive amount of shares get sold/purchased when the underlying price moves. This is because of the following:
a) Assume I, Mr. MM sell 1000 call options for NKLA 25C 10/23 and 300 put options for NKLA 15p 10/23. I'm just going to make up deltas because it's too much effort to calculate them - 30 delta call, 20 delta put.
This implies Mr. MM needs the following to delta hedge: (1000 call options * 30 shares to buy for each) [to balance out writing calls) - (300 put options * 20 shares to sell for each) = 24,000 net shares Mr. MM needs to acquire to balance out his deltas/be fully neutral.
b) This works well when NKLA is at $20. But what about when it hits $19 (because it only can go down, just like their trucks). Thanks to gamma, now we have to recompute the deltas, because they've changed for both the calls (they went down) and for the puts (they went up).
Let's say to keep it simple that now my calls are 20 delta, and my puts are 30 delta. From the 24,000 net shares, Mr. MM has to now have:
(1000 call options * 20 shares to have for each) - (300 put options * 30 shares to sell for each) = 11,000 shares.
Therefore, with a $1 shift in price, now to hedge and be indifferent to direction, Mr. MM has to go from 24,000 shares to 11,000 shares, meaning he has to sell 13,000 shares ASAP, or take on increased risk. Now, you might be saying, "13,000 shares seems small. How would this disrupt the system?"
(This process, by the way, is called hedge unwinding)
It won't, in this example. But across thousands of MMs and millions of contracts, this can - especially in highly optioned tickers - make up a substantial fraction of the net flow of shares per day. And as we know from our desk example, the buying or selling of shares directly changes the price of the stock itself.
This, by the way, is why the NOPE formula takes the shape it does. Some astute readers might notice it looks similar to GEX, which is not a coincidence. GEX however replaces daily volume with open interest, and measures gamma over delta, which I did not find good statistical evidence to support, especially for earnings.
So, with our example above, why does NOPE measure system stability? We can assume for argument's sake that if someone buys a share of NKLA, they're fine with moderate price swings (+- $20 since it's NKLA, obviously), and in it for the long/medium haul. And in most cases this is fine - we can own stock and not worry about minor swings in price. But market makers can't* (they can, but it exposes them to risk), because of how delta works. In fact, for most institutional market makers, they have clearly defined delta limits by end of day, and even small price changes require them to rebalance their hedges.
This over the whole market adds up to a lot shares moving, just to balance out your stupid Robinhood YOLOs. While there are some tricks (dark pools, block trades) to not impact the price of the underlying, the reality is that the more options contracts there are on a ticker, the more outsized influence it will have on the ticker's price. This can technically be exactly balanced, if option put delta is equal to option call delta, but never actually ends up being the case. And unlike shares traded, the shares representing the options are more unstable, meaning they will be sold/bought in response to small price shifts. And will end up magnifying those price shifts, accordingly.

NOPE and Earnings

So we have a new shiny indicator, NOPE. What does it actually mean and do?
There's much literature going back to the 1980s that options markets do have some level of predictiveness towards earnings, which makes sense intuitively. Unlike shares markets, where you can continue to hold your share even if it dips 5%, in options you get access to expanded opportunity to make riches... and losses. An options trader betting on earnings is making a risky and therefore informed bet that he or she knows the outcome, versus a share trader who might be comfortable bagholding in the worst case scenario.
As I've mentioned largely in comments on my prior posts, earnings is a special case because, unlike popular misconceptions, stocks do not go up and down solely due to analyst expectations being meet, beat, or missed. In fact, stock prices move according to the consensus market expectation, which is a function of all the participants' FEPF on that ticker. This is why the price moves so dramatically - even if a stock beats, it might not beat enough to justify the high price tag (FSLY); even if a stock misses, it might have spectacular guidance or maybe the market just was assuming it would go bankrupt instead.
To look at the impact of NOPE and why it may play a role in post-earnings-announcement immediate price moves, let's review the following cases:
  1. Stock Meets/Exceeds Market Expectations (aka price goes up) - In the general case, we would anticipate post-ER market participants value the stock at a higher price, pushing it up rapidly. If there's a high absolute value of NOPE on said ticker, this should end up magnifying the positive move since:
a) If NOPE is high negative - This means a ton of put buying, which means a lot of those puts are now worthless (due to price decoherence). This means that to stay delta neutral, market makers need to close out their sold/shorted shares, buying them, and pushing the stock price up.
b) If NOPE is high positive - This means a ton of call buying, which means a lot of puts are now worthless (see a) but also a lot of calls are now worth more. This means that to stay delta neutral, market makers need to close out their sold/shorted shares AND also buy more shares to cover their calls, pushing the stock price up.
2) Stock Meets/Misses Market Expectations (aka price goes down) - Inversely to what I mentioned above, this should push to the stock price down, fairly immediately. If there's a high absolute value of NOPE on said ticker, this should end up magnifying the negative move since:
a) If NOPE is high negative - This means a ton of put buying, which means a lot of those puts are now worth more, and a lot of calls are now worth less/worth less (due to price decoherence). This means that to stay delta neutral, market makers need to sell/short more shares, pushing the stock price down.
b) If NOPE is high positive - This means a ton of call buying, which means a lot of calls are now worthless (see a) but also a lot of puts are now worth more. This means that to stay delta neutral, market makers need to sell even more shares to keep their calls and puts neutral, pushing the stock price down.
---
Based on the above two cases, it should be a bit more clear why NOPE is a measure of sensitivity to system perturbation. While we previously discussed it in the context of magnifying directional move, the truth is it also provides a directional bias to our "random" walk. This is because given a price move in the direction predicted by NOPE, we expect it to be magnified, especially in situations of price decoherence. If a stock price goes up right after an ER report drops, even based on one participant deciding to value the stock higher, this provides a runaway reaction which boosts the stock price (due to hedging factors as well as other participants' behavior) and inures it to drops.

NOPE and NOPE_MAD

I'm going to gloss over this section because this is more statistical methods than anything interesting. In general, if you have enough data, I recommend using NOPE_MAD over NOPE. While NOPE in theory represents a "real" quantity (net option delta over net share delta), NOPE_MAD (the median absolute deviation of NOPE) does not. NOPE_MAD simply answecompare the following:
  1. How exceptional is today's NOPE versus historic baseline (30 days prior)?
  2. How do I compare two tickers' NOPEs effectively (since some tickers, like TSLA, have a baseline positive NOPE, because Elon memes)? In the initial stages, we used just a straight numerical threshold (let's say NOPE >= 20), but that quickly broke down. NOPE_MAD aims to detect anomalies, because anomalies in general give you tendies.
I might add the formula later in Mathenese, but simply put, to find NOPE_MAD you do the following:
  1. Calculate today's NOPE score (this can be done end of day or intraday, with the true value being EOD of course)
  2. Calculate the end of day NOPE scores on the ticker for the previous 30 trading days
  3. Compute the median of the previous 30 trading days' NOPEs
  4. From the median, find the 30 days' median absolute deviation (https://en.wikipedia.org/wiki/Median_absolute_deviation)
  5. Find today's deviation as compared to the MAD calculated by: [(today's NOPE) - (median NOPE of last 30 days)] / (median absolute deviation of last 30 days)
This is usually reported as sigma (σ), and has a few interesting properties:
  1. The mean of NOPE_MAD for any ticker is almost exactly 0.
  2. [Lily's Speculation's Speculation] NOPE_MAD acts like a spring, and has a tendency to reverse direction as a function of its magnitude. No proof on this yet, but exploring it!

Using the NOPE to predict ER

So the last section was a lot of words and theory, and a lot of what I'm mentioning here is empirically derived (aka I've tested it out, versus just blabbered).
In general, the following holds true:
  1. 3 sigma NOPE_MAD tends to be "the threshold": For very low NOPE_MAD magnitudes (+- 1 sigma), it's effectively just noise, and directionality prediction is low, if not non-existent. It's not exactly like 3 sigma is a play and 2.9 sigma is not a play; NOPE_MAD accuracy increases as NOPE_MAD magnitude (either positive or negative) increases.
  2. NOPE_MAD is only useful on highly optioned tickers: In general, I introduce another parameter for sifting through "candidate" ERs to play: option volume * 100/share volume. When this ends up over let's say 0.4, NOPE_MAD provides a fairly good window into predicting earnings behavior.
  3. NOPE_MAD only predicts during the after-market/pre-market session: I also have no idea if this is true, but my hunch is that next day behavior is mostly random and driven by market movement versus earnings behavior. NOPE_MAD for now only predicts direction of price movements right between the release of the ER report (AH or PM) and the ending of that market session. This is why in general I recommend playing shares, not options for ER (since you can sell during the AH/PM).
  4. NOPE_MAD only predicts direction of price movement: This isn't exactly true, but it's all I feel comfortable stating given the data I have. On observation of ~2700 data points of ER-ticker events since Mar 2019 (SPY 500), I only so far feel comfortable predicting whether stock price goes up (>0 percent difference) or down (<0 price difference). This is +1 for why I usually play with shares.
Some statistics:
#0) As a baseline/null hypothesis, after ER on the SPY500 since Mar 2019, 50-51% price movements in the AH/PM are positive (>0) and ~46-47% are negative (<0).
#1) For NOPE_MAD >= +3 sigma, roughly 68% of price movements are positive after earnings.
#2) For NOPE_MAD <= -3 sigma, roughly 29% of price movements are positive after earnings.
#3) When using a logistic model of only data including NOPE_MAD >= +3 sigma or NOPE_MAD <= -3 sigma, and option/share vol >= 0.4 (around 25% of all ERs observed), I was able to achieve 78% predictive accuracy on direction.

Caveats/Read This

Like all models, NOPE is wrong, but perhaps useful. It's also fairly new (I started working on it around early August 2020), and in fact, my initial hypothesis was exactly incorrect (I thought the opposite would happen, actually). Similarly, as commenters have pointed out, the timeline of data I'm using is fairly compressed (since Mar 2019), and trends and models do change. In fact, I've noticed significantly lower accuracy since the coronavirus recession (when I measured it in early September), but I attribute this mostly to a smaller date range, more market volatility, and honestly, dumber option traders (~65% accuracy versus nearly 80%).
My advice so far if you do play ER with the NOPE method is to use it as following:
  1. Buy/short shares approximately right when the market closes before ER. Ideally even buying it right before the earnings report drops in the AH session is not a bad idea if you can.
  2. Sell/buy to close said shares at the first sign of major weakness (e.g. if the NOPE predicted outcome is incorrect).
  3. Sell/buy to close shares even if it is correct ideally before conference call, or by the end of the after-market/pre-market session.
  4. Only play tickers with high NOPE as well as high option/share vol.
---
In my next post, which may be in a few days, I'll talk about potential use cases for SPY and intraday trends, but I wanted to make sure this wasn't like 7000 words by itself.
Cheers.
- Lily
submitted by the_lilypad to thecorporation [link] [comments]

The classic WSB story - lost it all.

Going to keep this simple. EDIT: this isn’t simple and I should write a short story on this.
I am generally risk averse. I hate losing $100 at the casino, I hate paying extra for guac at chipotles, I will return something or price match an item for a few dollars of savings. I am generally frugal.
But, I somehow had no issues losing 10k in options...
How I started
I remember my first trades like they were yesterday. I was trading the first hydrogen run-up in 2014 (FCEL, BLDP, PLUG) and made a few hundred dollars over a couple weeks.
I quickly progressed to penny stocks / biotech binary events and general stock market gambling mid-2014. I was making a few % here and there but the trend was down in total account value. I was the king of buying the peak in run-ups. I managed to make it out of 2014 close to break-even to slightly down.
WSB Era
March 2015 was my first option trade. It was an AXP - American Express - monthly option trade. I saw one of the regular option traders/services post a block of 10,000 calls that had been bought for 1.3 and I followed the trade with 10 call options for a total of $1300.
I woke up the next day to an analyst upgrade on AXP and was up 50% on my position. I was addicted! I day-dreamed for days about my AXP over night success. I think around that time there was some sort of Buffet buyout of Heinz and an option trade that was up a ridiculous amount of %%%. I wanted to hit it BIG.
I came up with the idea that all I needed to reach my goal was a few 100% over night gains/ 1k>2k>4k>8k> etc. I convinced myself that I would have no problems being patient for the exact criteria that I had set and worked on some other trades.
Remember, the first win is always free.
I was trading options pretty regularly from March 2015 until August 2016. During my best week I was up 20k and could feel the milli within reach. I can remember the exact option trade (HTZ) and I was trading weeklies on it.
For those who have been in the market long enough, you will remember the huge drawdown of August 2015.
I lost half my account value on QCOM calls (100 of them) that I followed at the beginning of July and never materialized. I watched them eventually go to 0. It was another 10,000 block that was probably a hedge or sold.
In August 2015 there were some issues with China and all of us woke up to stocks gapping down huge. Unfortunately my idea of buying far dated calls during the following days/weeks after the crash went sideways. I quickly learned that an increase in volatility causes a rise in option prices and I was paying a premium for calls that were going to lose value very quickly (the infamous IV crush).
I kept trading options into the end of 2015 and managed to maintain my account value positive but the trading fees for the year amounted to $30,000+. My broker was loving it.
I tried all the services, all the strategies. I created rules for my option plays: 1. No earnings 2. Only follow the big buys at a discount (10,000 blocks or more). 3. No weekly options 4. Take profit right away 5. Take losses quickly 6. etc.
I had a whole note book of option plays that I was writing down and following. I was paying for option services that all of you know about - remember, they make money on the services and not trading.
I even figured out a loop-hole with my broker: if I didn’t have enough money in my account, I could change my ask price to .01 and then change it to market buy and I would only need to accept a warning ⚠️ for the order to go through. I was able to day trade the option and make money, who cares if I didnt have enough? After a few months of this, I got a call from my broker that told me to stop and that I would be suspended if I continued with this.
By the way, I was always able to satisfy the debit on the account - so it wasn’t an issue of lack of funds.
Lost it all. Started taking money from lines of credits, every penny that I earned and losing it quicker and quicker.
I was a full on gambler but I was convinced that 8 trades would offset all the losses. I kept getting drawn in to the idea that I could hit a homerun and make it out a hero.
I eventually hit rock bottom on some weekly expiring FSLR options that I bought hours before expiration and said to myself - what the f are you doing? I resolved to invest for the long term and stop throwing tendies away.
The feeling was reinforced during the birth of my first born and I thought - what a loser this kid will think of me if he knew how much I was gambling and wasting my life. It was a really powerful moment looking at my kid and reflecting on this idea.
I decided at that point I was going to save every penny I had and invest it on new issues with potential.
Fall 2016
TTD, COUP and NTNX IPO ‘ed I decided I was going to throw every dollar at these and did so for the next few months. I eventually started using margin (up to 215%) and buying these for the next 6 months. They paid out and managed to make it over 100k within the year.
The first 100k was hard but once I crossed it, I never fell below this magic number.
2017 - I did some day trading but it was mostly obsessing over the above issues. I did gamble on a few options here and there but never more than 1k.
2018 - SFIX was my big winner, I bought a gap up in June 2018 and my combined account value had crossed 400k by August 2018. I was really struggling at crossing the 500k account value and experienced 3 x 30-40% drawdowns over the next 2 years before I finally crossed the 500k barrier and have never looked back.
I still made some mistakes over the next few months - AKAO & GSUM come to mind. Both of these resulted in 20k+ losses. Fortunately my winners were much bigger than my losers.
I thought about giving up and moving to index funds - but i was doing well - just experiencing large drawdowns because of leverage.
2019 big winners were CRON SWAV STNE.
2017 / 2018 / 2019 all had six digit capital gains on my tax returns.
At the beginning of 2020 I was still day trading on margin (180-220%) and got a call from my broker that they were tightening up my margin as my account was analyzed by the risk department and deemed too risky. Believe it or not this was right before the covid crash. I brought my margin down to 100-110% of account value and even though the drawdown from covid hit hard, I wasn’t wiped out.
I stayed the course and bought FSLY / RH during the big march drawdown and this resulted in some nice gains over the next few months.
I am constantly changing and testing my investment strategy but let me tell you that obsessing over 1 or 2 ideas and throwing every penny at it and holding for a few years is the best strategy. It may not work at some point but right now it does.
I still day trade but I trade with 10k or less on each individual position. It allows me minimize my losses and my winners are 1-7%. I am able to consistently make between 3-700$/ a day on day trades using the above strategy. I still take losses and still dream about hitting it big with an option trade but dont feel the need to put it all on the line every month / week.
I finally crossed into the two , club. I know people are going to ask for proof or ban but I am not earning anything for posting and the details about some of the trades should be proof enough that I kept a detailed journal of it all. I have way more to write but these are the highlights.
Eventually I will share how I build a position in a story I love. I still sell buy and sell to early but I am working on improving.
TL:DR - I gambled, lost it all and gambled some more lost more. I made it out alive. I have only sold calls/puts lately.
The one common denominator in all successful people is how much they obsess over 1 or 2 ideas. Do the same. All the winners on this sub have gone all in on one idea (FSLY / TSLA ). Stick with new stories or ones that are changing and go all in...wait a second, I didnt learn anything.
submitted by jojo2021 to wallstreetbets [link] [comments]

AMZN Trade Retrospective: Collecting a $.37 Credit for the Potential to Make Another $50

AMZN Trade Retrospective: Collecting a $.37 Credit for the Potential to Make Another $50
There are different ways to trade in a choppy environment. Here’s a deep dive on how I attempted to use weekly options to trade a potential bounce in AMZN, and collected $.37 initially, for the possibility of making $50 more, even though the trade ended up being only an $.81 winner.

The Entry

Last Thursday, 9/24, when $AMZN was trading at about $3000 a share, I was looking for a cheap way to play a bounce in the stock. During that time, my bias in the markets had begun to shift to a more bullish stance after seeing how the market had difficulty grinding lower. With that in mind, I wanted to play a potential bounce in tech. But I knew I didn’t want to pay a debit at all to play for a bounce that might not even happen, given how uncertain and choppy the markets had been, but I still wanted to set myself up to capture some large gains if AMZN did indeed bounce. Therefore, the strategy that made the most sense to me, was a Call broken wing butterfly.
Given that I’m a very short-term options trader who loves trading weeklies, I was trying to look for a cheap butterfly for the upcoming week that I could put on for a net credit. After exploring the options chain, I came across the +1/-2/+1 3300/3350/3450 call broken wing butterfly for the Oct 2 series. This fly, at the time (on Sept 24), was trading for a total of $.37 credit. Meaning, by putting on that butterfly, I would get paid $.37, and the following scenarios could happen:
  1. If AMZN decided to tank or hang out sideways and never get up close enough to the butterfly to expand the spread in my favor, then I’d walk away pocketing the $.37 credit
  2. If AMZN slowly crept up to reach exactly 3350 by expiration, I’d not only get to keep the credit, but also be able to sell the butterfly back out for $50. Of course, it doesn’t need to reach exactly 3350 by expiration. If AMZN slowly worked its way up to near 3300, then the butterfly would expand very nicely as well.
  3. If AMZN blew past 3400 by expiration, I’d see a loss, up to a maximum of $50 / spread (if $AMZN moves past 3450). That’s because the 3300/3350 long call vertical of the fly provides 50 points of coverage before I essentially start losing money from the 3350/3450 short vertical, up until that 3450 kicks in to cap off further upside losses.
So that is a rough outline of the potential scenarios that would happen with this trade.
Given the choppy market conditions, I was ok with risking $50/spread (point #3), in order to not lose money if I’m wrong on direction (point #1), while at the same time, keeping myself open to the possibility of the butterfly expanding in my favor (point #2) for some potentially very large gains.
But satisfying point #3 is tricky. I needed more data points suggesting that $AMZN wouldn’t surge higher early on in the trade. Because if $AMZN did surge higher early on in the trade, then while the 3300 long call would rise in value, those two 3350 short calls would also rise in value, and because there’d still be some time value left, they could be very juiced up and eat away at the profits of that 3300 long call, so much so that the 3450 long call won’t even be able to offset those losses, especially given how far out of the money that 3450 call is.

AMZN on 9/24, daily timeframe
Looking at the chart above on 9/24, we can see that AMZN was trading at around $3000/share. In order to reach $3300 (where the first long call of the broken wing butterfly is), the stock would need to
  1. Breach the 38% fib retracement (~AMZN=3131) of the move from the 9/2 high to the 9/21 low,
  2. Breach the 20MA and 50MA
  3. Breach the 50% fib retracement (~AMZN=3211)
  4. Breach the 61.8% fib retracement (~AMZN=3292)
before finally reaching the 3300 long call. All of these levels, I felt, should provide some resistance for AMZN to have to chew thru over the following week, before it even gets to the long call. And by that time, if AMZN did reach 3300, then the 3300 long call would still have a lot of extrinsic value left (somewhere around $20 on the last day), while the 3350 short calls would be very cheap (each around $5), so the entire spread could be roughly worth $10. Which would be great, because that means I’d be getting paid $.37 to make another $10.
So with all of the above considered, I chose to take on that upside risk, for a chance to make potentially $50 (realistically I try to aim for just half of the max profit: $25, and start harvesting profits and peeling off the flies at around $5-$10), and that day on 9/24, entered the Oct2 3300/3350/3450 call broken wing butterfly for a $.37 credit.
After entry, on Friday 9/25 and Monday 9/28, AMZN made steady progress upwards, from 3000 to 3175, breaching the 31.8% retracement and tagging the 20MA and 50MA from below.

AMZN on 9/28, daily timeframe
but this move wasn’t large and fast enough to expand the value of the 3350 short calls. In fact, theta did a great job draining those short calls, while the 3300 long call did a good job retaining its premium, so the butterfly had already expanded a bit in my favor, and I was sitting at about a small $1.00 profit.

The Adjustment

However, on Tuesday and Wednesday, AMZN began to stall out. By the end of Wednesday 9/30, when it looked like AMZN was putting in a topping tail, I decided that AMZN might not be able to make it near 3300 by expiration Friday, so I wanted to take in a bit more credit while I still could, before theta drained more of that 3300 long call. At the time, the spread was trading for almost $2.
That’s when I made a slight adjustment to the spread and sold the 3300/3310 call vertical.

AMZN on 9/30, daily timeframe
This essentially rolled the 3300 long call up to 3310, and I was able to collect a small $.44 credit for it. However, this adjustment did open me up to an additional $10 of risk to the upside, because now, the long call vertical portion of the butterfly is only $40 wide (instead of $50). Still, with only 2 days left for AMZN to go higher, I felt comfortable taking on a bit more upside risk knowing that theta is going to be working hard to drain those 3350 short calls if AMZN did decide to surge higher. And at that moment, I actually wanted AMZN to move more towards my fly. My deltas were still positive, and the risk graph showed that a move towards the short strikes of the fly would expand it by another $4-5 by Thursday.
So after this adjustment, the trade stood at a $.81 credit, and the profit potential on the fly was now $40 instead of $50. Which is still pretty good.

The Tease

On Thursday, AMZN showed some strength and closed above the 50% fib (3211), which meant that if on Friday, AMZN worked its way up to around 3300, the fly could potentially be worth $5-10. Things were looking good (on any continued bullishness, the next target for AMZN was the 61.8% fib retracement at ~3300). So I left the trade alone without making any more adjustments.

AMZN on 10/1, daily timeframe

The Flop

Unfortunately, on Thursday night, news broke out that Trump was diagnosed with Coronavirus, and the market fell lower. By the open, AMZN was already trading at around 3150, roughly 150 points below the fly. The spread had instantly lost all of its value, so I basically let it expire worthless and walked away pocketing the $.81 credit.

https://preview.redd.it/mpwrkjpk6xq51.png?width=4096&format=png&auto=webp&s=8dd7f4da7b000b2266ab57a3c23c1863f9423704
While the trade did not work out as well as I had liked, the important thing to note is that I was able to get paid even when the trade didn’t go in my favor. With options, there are ways to trade an underlying to a certain target without ponying up a debit, albeit at the cost of introducing tail risk, while offering the possibility of very large upside. This may be a style of trading that one can consider employing when the outlook of the markets is uncertain, as long as the trader is willing to make the necessary adjustments to control risk.
Which leads me to the following section:

FAQ

What if AMZN decided to surge very early on during the trade? What if AMZN had surged to 3300 with 4-5 DTE, hence juicing up the short calls and causing the butterfly to take on large negative deltas?
Even though the position would be very theta positive, I would pony up the debit to cap off the upside risk by buying the 3400/3450 call vertical, hence turning the 3300/3350/3450 broken wing butterfly into the 3300/3350/3400 balanced butterfly. From there on out until expiration, I would look for ways to reduce the debit incurred from that adjustment.

But what if AMZN tanked afterwards? You could end up getting whipsawed.
I’d rather be safe than sorry and make the necessary adjustments to avoid getting run over, because I don’t like playing the hope card. I could always undo the adjustment and look for ways to collect back more credit (at the cost of introducing risk elsewhere), depending on my new directional bias on AMZN at the time.

Your maximum loss is so large, $5000. I’d never make that bet, I would never risk $5000 to make $5000.
This style of trading is not for everyone. There are different ways to perceive risk. I don't really think of risk as binary as “max gain vs max loss”. If the trade goes against me, I’m not going to open myself up to the possibility of eating the maximum loss. I’m going to manage that risk and make sure that I don’t lose any money at all on the trade. Basically, I’m not going to just put on the trade, walk away to the prayer room, and come back at expiration and hope that AMZN expired at 3350.

Why not just join thetagang and slap on iron condors / credit spreads in this environment? You could’ve collected more credit by selling a 50 point wide put vertical with your bounce thesis.
Different traders have different styles. I personally don’t like pure premium selling strategies. I’d rather have long options in front of the shorts to open myself up for some large upside and convexity in the P/L curve, rather than limit myself to the concavity of pure premium selling strategies. Having long options in front of the shorts also helps me sleep better at night.

It’s hard to read this. Is there a more visual explanation?
Here’s a video on it: https://www.youtube.com/watch?v=8uq76fZ3EME

TL;DR - I used weekly options to trade a potential bounce in AMZN, and got paid $.37 initially to do so, for the possibility of making $50 more. While the trade did not pan out, I walked away pocketing $.81 for being wrong.
submitted by OptionsBrewers to options [link] [comments]

The Challenges of Designing a Modern Skill, Part 3

Okay, Wendy’s or Walgreens or whoever, I don’t care who you are, you’re listening to the rest.

Introduction to Part 3

Welcome back one last time to “The Challenges of Designing a Modern Skill,” a series where we discuss all aspects of skill design and development. In Part 1, we talked about OSRS’s history with skills, and started the lengthy conversation on Skill Design Philosophy, including the concepts of Core, Expansion, and Integration. This latter topic consumed the entirety of Part 2 as well, which covered Rewards and Motivations, Progression, Buyables, as well as Unconstructive Arguments.
Which brings us to today, the final part of our discussion. In this Part 3, we’ll finish up Section 3 – Skill Design Philosophy, then move on to chat about the design and blog process. One last time, this discussion was intended to be a single post, but its length outgrew the post character limit twice. Therefore, it may be important to look at the previous two parts for clarity and context with certain terms. The final product, in its purest, aesthetic, and unbroken form, can be found here.

3-C – Skill Design Philosophy, Continued

3-12 - Balancing

What follows from the discussion about XP and costs, of course, is balancing: the bane of every developer. A company like Riot knows better than anyone that having too many factors to account for makes good balance impossible. Balancing new ideas appropriately is extremely challenging and requires a great respect for current content as discussed in Section 3-5 – Integration. Thankfully, in OSRS we only have three major balancing factors: Profit, XP Rate, and Intensity, and two minor factors: Risk and Leniency. These metrics must amount to some sense of balance (besides Leniency, which as we’ll see is the definition of anti-balance) in order for a piece of content to feel like it’s not breaking the system or rendering all your previous efforts meaningless. It’s also worthy to note that there is usually a skill-specific limit to the numerical values of these metrics. For example, Runecrafting will never receive a training method that grants 200k xp/hr, while for Construction that’s easily on the lower end of the scale.
A basic model works better than words to describe these factors, and therefore, being the phenomenal artist that I am, I have constructed one, which I’ve dubbed “The Guthix Scale.” But I’ll be cruel and use words anyway.
  • Profit: how much you gain from a task, or how much you lose. Gain or loss can include resources, cosmetics, specialized currencies, good old gold pieces, or anything on that line.
  • XP Rate: how fast you gain XP.
  • Intensity: how much effort (click intensity), attention (reaction intensity), and thought (planning intensity) you need to put into the activity to perform it well.
  • Risk: how likely is the loss of your revenue and/or resource investment into the activity. Note that one must be careful with risk, as players are very good at abusing systems intended to encourage higher risk levels to minimize how much they’re actually risking.
  • Leniency: a measure for how imbalanced a piece of content can be before the public and/or Jagex nerfs it. Leniency serves as a simple modulator to help comprehend when the model breaks or bends in unnatural ways, and is usually determined by how enjoyable and abusable an activity is, such that players don’t want to cause an outrage over it. For example, Slayer has a high level of Leniency; people don’t mind that some Slayer tasks grant amazing XP Rates, great Profits, have middling Intensity, and low Risk. On the other hand, Runecrafting has low levels of Leniency; despite low Risk, many Runecrafting activities demand high Intensity for poor XP Rates and middling Profits.
In the end, don’t worry about applying specific numbers during the conceptual phase of your skill design. However, when describing an activity to your reader, it’s always useful if you give approximations, such as “high intensity” or “low risk,” so that they get an idea of the activity’s design goals as well as to guide the actual development of that activity. Don’t comment on the activity’s Leniency though, as that would be pretty pretentious and isn’t for you to determine anyway.

3-13 - Skill Bloat

What do the arts of weaving, tanning, sowing, spinning, pottery, glassmaking, jewellery, engraving, carving, chiselling, carpentry, and even painting have in common? In real life, there’s only so much crossover between these arts, but in Runescape they’re all simply Crafting.
The distinction between what deserves to be its own skill or instead tagged along to a current skill is often arbitrary; this is the great challenge of skill bloat. The fundamental question for many skill concepts is: does this skill have enough depth to stand on its own? The developers of 2006 felt that there was sufficient depth in Construction to make it something separate from Crafting, even if the latter could have covered the former. While there’s often no clean cut between these skills (why does making birdhouses use Crafting instead of Construction?), it is easy to see that Construction has found its own solid niche that would’ve been much too big to act as yet another Expansion of Crafting.
On the other hand, a skill with extremely limited scope and value perhaps should be thrown under the umbrella of a larger skill. Take Firemaking: it’s often asked why it deserves to be its own skill given how limited its uses are. This is one of those ideas that probably should have just been thrown under Crafting or even Woodcutting. But again, the developers who made early Runescape did not battle with the same ideas as the modern player; they simply felt like Firemaking was a good idea for a skill. Similarly, the number of topics that the Magic skill covers is so often broken down in other games, like Morrowind’s separation between Illusion, Conjuration, Alteration, Destruction, Mysticism, Restoration, Enchant, Alchemy (closer to Herblore), and Unarmored (closer to Strength and Defense). Why does Runescape not break Magic into more skills? The answer is simple: Magic was created with a much more limited scope in Runescape, and there has not been enough content in any specific magical category to justify another skill being born. But perhaps your skill concept seeks to address this; maybe your Enchantment skill takes the enchanting aspects of Magic away, expands the idea to include current imbues and newer content, and fully fleshes the idea out such that the Magic skill alone cannot contain it. Somewhat ironically, Magic used to be separated into Good and Evil Magic skills in Runescape Classic, but that is another topic.
So instead of arguments about what could be thrown under another skill’s umbrella, perhaps we should be asking: is there enough substance to this skill concept for it to stand on its own, outside of its current skill categorization? Of course, this leads to a whole other debate about how much content is enough for a skill idea to deserve individuality, but that would get too deep into specifics and is outside the scope of this discussion.

3-14 - Skill Endgame

Runescape has always been a sandbox MMO, but the original Runescape experience was built more or less with a specific endgame in mind: killing players and monsters. Take the Runescape Classic of 2001: you had all your regular combat skills, but even every other skill had an endgame whose goal was helping combat out. Fishing, Firemaking, and Cooking would provide necessary healing. Smithing and Crafting, along with their associated Gathering skill partners, served to gear you up. Combat was the simple endgame and most mechanics existed to serve that end.
However, since those first days, the changing endgame goals of players have promoted a vast expansion of the endgame goals of new content. For example, hitting a 99 in any non-combat skill is an endgame goal in itself for many players, completely separate from that skill’s combat relationship (if any). These goals have increased to aspects like cosmetic collections, pets, maxed stats, all quests completed, all diaries completed, all music tracks unlocked, a wealthy bank, the collection log, boss killcounts, and more. Whereas skills used to have a distinct part of a system that ultimately served combat, we now have a vast variety of endgame goals that a skill can be directed towards. You can even see a growth in this perspective as new skills were released up to 2007: Thieving mainly nets you valuable (or once valuable) items which have extremely flexible uses, and Construction has a strong emphasis on cosmetics for your POH.
So when designing your new skill, contemplate what the endgame of your skill looks like. For example, if you are proposing a Gathering skill, what is the Production skill tie-in, and what is the endgame goal of that Production skill? Maybe your new skill Spelunking has an endgame in gathering rare collectibles that can be shown off in your POH. Maybe your new skill Necromancy functions like a Support skill, giving you followers that help speed along resource gathering, and letting you move faster to the endgame goal of the respective Production skill. Whatever it is, a proper, clear, and unified view of an endgame goal helps a skill feel like it serves a distinct and valuable purpose. Note that this could mean that you require multiple skills to be released simultaneously for each to feed into each other and form an appropriate endgame. In that case, go for it – don’t make it a repeat of RS3’s Divination, a Gathering skill left hanging without the appropriate Production skill partner of Invention for over 2 years.
A good example of a skill with a direct endgame is… most of them. Combat is a well-accepted endgame, and traditionally, most skills are intended to lend a hand in combat whether by supplies or gear. A skill with a poor endgame would be Hunter: Hunter is so scattered in its ultimate endgame goals, trying to touch on small aspects of everything like combat gear, weight reduction, production, niche skilling tools, and food. There’s a very poor sense of identity to Hunter’s endgame, and it doesn’t help that very few of these rewards are actually viable or interesting in the current day. Similarly, while Slayer has a strong endgame goal it is terrible in its methodology, overshadowing other Production skills in their explicit purpose. A better design for Slayer’s endgame would have been to treat it as a secondary Gathering skill, to work almost like a catalyst for other Gathering-Production skill relationships. In this mindset, Slayer is where you gather valuable monster drops, combine it with traditional Gathering resources like ores from Mining, then use a Production skill like Smithing to meld them into the powerful gear that is present today. This would have kept other Gathering and Production skills at the forefront of their specialities, in contrast to today’s situation where Slayer will give fully assembled gear that’s better than anything you could receive from the appropriate skills (barring a few items that need a Production skill to piece together).

3-15 - Alternate Goals

From a game design perspective, skills are so far reaching that it can be tempting to use them to shift major game mechanics to a more favourable position. Construction is an example of this idea in action: Construction was very intentionally designed to be a massive gold sink to help a hyperinflating economy. Everything about it takes gold out of the game, whether through using a sawmill, buying expensive supplies from stores, adding rooms, or a shameless piece of furniture costing 100m that is skinned as, well, 100m on a shameless piece of furniture.
If you’re clever about it, skills are a legitimately good opportunity for such change. Sure, the gold sink is definitely a controversial feature of Construction, but for the most part it’s organic and makes sense; fancy houses and fancy cosmetics are justifiably expensive. It is notable that the controversy over Construction’s gold sink mechanism is probably levied more against the cost of training, rather than the cost of all its wonderful aesthetics. Perhaps that should have been better accounted for in its design phase, but now it is quite set in stone.
To emphasize that previous point: making large scale changes to the game through a new skill can work, but it must feel organic and secondary to the skill’s main purpose. Some people really disliked Warding because they felt it tried too hard to fix real, underlying game issues with mechanics that didn’t thematically fit or were overshadowing the skill’s Core. While this may or may not be true, if your new skill can improve the game’s integrity without sacrificing its own identity, you could avoid this argument entirely. If your skill Regency has a Core of managing global politics, but also happens to serve as a resource sink to help your failing citizens, then you’ve created a strong Core design while simultaneously improving the profitability of Gathering skills.

3-16 - The Combat No-Touch Rule

So, let’s take a moment to examine the great benefits and rationale of RS2’s Evolution of Combat:
This space has been reserved for unintelligible squabbling.
With that over, it’s obvious that the OSRS playerbase is not a big fan of making major changes to the combat system. If there’s anything that defines the OSRS experience, it has to be the janky and abusable combat system that we love. So, in the past 7 years of OSRS, how many times have you heard someone pitch a new combat skill? Practically no one ever has; a new combat skill, no matter how miniscule, would feel obtrusive to most players, and likely would not even receive 25% of votes in a poll. This goes right back to Section 3-5 – Integration, and the importance of preserving the fundamentals of OSRS’s design.
I know that my intention with this discussion was to be as definitive about skill design as possible, and in that spirit I should be delving into the design philosophy specifically behind combat skills, but I simply don’t see the benefit of me trying, and the conversation really doesn’t interest me that much. It goes without saying that as expansive as this discussion is, it does not cover every facet of skill design, which is a limitation both of my capabilities and desire to do so.

3-17 - Aesthetics

I don’t do aesthetics well. I like them, I want them, but I do not understand them; there are others much better equipped to discuss this topic than I. Nonetheless, here we go.
Since the dawn of OSRS, debates over art style and aesthetics have raged across Gielinor. After all, the OSRS Team is filled with modern day artists while OSRS is an ancient game. What were they supposed to do? Keep making dated graphics? Make content with a modernized and easily digestible style? Something in-between?
While many players shouted for more dated graphics, they were approached by an interesting predicament: which dated graphics did they want? We had a great selection present right from the start of OSRS: 2002, 2003, 2004, 2005, 2006, and 2007. People hungry for nostalgia chose the era that they grew up in, leading to frequent requests for older models like the dragon or imp, most of which were denied by Jagex (except the old Mining rock models). But which era was OSRS supposed to follow?
Jagex elected to carve their own path, but not without heavy criticism especially closer to OSRS’s conception. However, they adapted to player requests and have since gone back and fixed many of the blatant early offenders (like the Kingdom of Kourend) and adopted a more consistent flavour, one that generally respects the art style of 2007. Even though it doesn’t always hit the mark, one has to appreciate the OSRS artists for making their best attempt and listening to feedback, and here’s to hoping that their art style examination mentioned in June 2020’s Gazette bears fruit.
But what exactly is the old school art style? There are simple systems by which most players judge it in OSRS, usually by asking questions like, “Would you believe if this existed in 2007?” More informed artists will start pointing out distinct features that permeated most content from back in the day, such as low quality textures, low poly models, low FPS animations, a “low fantasy” or grounded profile that appeals somewhat to realism, reducing cartoonish exaggerations, and keeping within the lore. Compiled with this, music and sound design help that art style come to life; it can be very hard on immersion when these don’t fit. An AGS would sound jarring if its special attack sounded like a weak dagger stab, and having to endure Country Jig while roaming Hosidius suddenly sweeps you off into a different universe.
But coming back to skill design, the art, models, and sound design tend to be some of the last features, mostly because the design phase doesn’t demand such a complete picture of a skill. However, simple concept art and models can vastly improve how a skill concept is communicated and comfort players who are concerned about maintaining that “old school feel.” This will be touched on again later in this discussion under Section 5-2 – Presentation and Beta Testing.

3-18 - Afterword

Now we’ve set down the modern standards for a new skill, but the statements that started this section bear repeating: the formula we’ve established does not automatically make a good or interesting skill, as hard as we might have tried. Once again, harken back to the First Great Irony: that we are trying to inject the modern interpretation of what defines a skill upon a game that was not necessarily built to contain it. Therefore, one could just as easily deny each of the components described above, as popular or unpopular as the act might be, and their opinion could be equally valid and all this effort meaningless. Don’t take these guidelines with such stringency as to disregard all other views.

5-0 - The OSRS Team and the Design Process

If you’ve followed me all the way here, you’re likely A) exhausted and fed up of any conversation concerning new skills, or B) excited, because you’ve just struck an incredible skill idea (or perhaps one that’s always hung around your head) that happens to tick off all the above checkboxes. But unfortunately for you B types, it’s about to get pretty grim, because we’re going to go through every aspect of skill design that’s exterior to the game itself. We’ll be touching on larger topics like democracy, presentation, player mindsets, effort, and resource consumption. It’ll induce a fantastic bout of depression, so don’t get left behind.

5-1 - Designing a Skill

Thus far, Jagex has offered three potential skills to OSRS, each of which has been denied. This gives us the advantage of understanding how the skill design process works behind the scenes and lets us examine some of the issues Jagex has faced with presenting a skill to the players.
The first problem is the “one strike and you’re out” phenomenon. Simply put, players don’t like applying much effort into reading and learning. They’ll look at a developer blog highlighting a new skill idea, and if you’re lucky they’ll even read the whole thing, but how about the second developer blog? The third? Fourth? Even I find it hard to get that far. In general, people don’t like long detail-heavy essays or blogs, which is why I can invoke the ancient proverb “Ban Emily” into this post and it’ll go (almost) completely unnoticed. No matter how many improvements you make between developer blogs, you will quickly lose players with each new iteration. Similarly, developer blogs don’t have the time to talk about skill design philosophy or meta-analyse their ideas – players would get lost far too fast. This is the Second Great Irony of skill design: the more iterations you have of a lengthy idea, the less players will keep up with you.
This was particularly prominent with Warding: Battle Wards were offered in an early developer blog but were quickly cut when Jagex realized how bad the idea was. Yet people would still cite Battle Wards as the reason they voted against Warding, despite the idea having been dropped several blogs before. Similarly, people would often comment that they hated that Warding was being polled multiple times; it felt to them like Jagex was trying to brute-force it into the game. But Warding was only ever polled once, and only after the fourth developer blog - the confusion was drawn from how many times the skill was reiterated and from the length of the public design process. Sure, there are people for whom this runs the opposite way; they keep a close eye on updates and judge a piece of content on the merits of the latest iteration, but this is much less common. You could argue that one should simply disregard the ignorant people as blind comments don't contribute to the overall discussion, but you should remember that these players are also the ones voting for the respective piece of content. You could also suggest re-educating them, which is exactly what Jagex attempts with each developer blog, and still people won’t get the memo. And when it comes to the players themselves, can the playerbase really be relied on to re-educate itself?
Overall, the Second Great irony really hurts the development process and is practically an unavoidable issue. What’s the alternative? To remove the developer-player interface that leads to valuable reiterations, or does you simply have to get the skill perfect in the first developer blog?
It’s not an optimal idea, but it could help: have a small team of “delegates” – larger names that players can trust, or player influencers – come in to review a new, unannounced skill idea under NDA. If they like it, chances are that other players will too. If they don’t, reiterate or toss out the skill before it’s public. That way, you’ve had a board of experienced players who are willing to share their opinions to the public helping to determine the meat and potatoes of the skill before it is introduced to the casual eye. Now, a more polished and well-accepted product can be presented on the first run of selling a skill to the public, resulting in less reiterations being required, and demanding less effort from the average player to be fully informed over the skill’s final design.

5-2 - Presentation and Beta Testing

So you’ve got a great idea, but how are you going to sell it to the public? Looking at how the OSRS Team has handled it throughout the years, there’s a very obvious learning curve occurring. Artisan had almost nothing but text blogs being thrown to the players, Sailing started introducing some concept art and even a trailer with terrible audio recording, and Warding had concept art, in game models, gifs, and a much fancier trailer with in-game animations. A picture or video is worth a thousand words, and often the only words that players will take out of a developer blog.
You might say that presentation is everything, and that would be more true in OSRS than most games. Most activities in OSRS are extremely basic, involve minimal thought, and are incredibly grindy. Take Fishing: you click every 20 seconds on a fishing spot that is randomly placed along a section of water, get rid of your fish, then keep clicking those fishing spots. Boiling it down further, you click several arbitrary parts of your computer screen every 20 seconds. It’s hardly considered engaging, so why do some people enjoy it? Simply put: presentation. You’re given a peaceful riverside environment to chill in, you’re collecting a bunch of pixels shaped like fish, and a number tracking your xp keeps ticking up and telling you that it matters.
Now imagine coming to the players with a radical new skill idea: Mining. You describe that Mining is where you gather ores that will feed into Smithing and help create gear for players to use. The audience ponders momentarily, but they’re not quite sure it feels right and ask for a demonstration. You show them some gameplay, but your development resources were thin and instead of rocks, you put trees as placeholders. Instead of ores in your inventory, you put logs as placeholders. Instead of a pickaxe, your character is swinging a woodcutting axe as a placeholder. Sure, the mechanics might act like mining instead of woodcutting, but how well is the skill going to sell if you haven’t presented it correctly or respected it contextually?
Again, presentation is everything. Players need to be able to see the task they are to perform, see the tools they’ll use, and see the expected outcomes; otherwise, whatever you’re trying to sell will feel bland and unoriginal. And this leads to the next level of skill presentation that has yet to be employed: Beta Worlds.
Part of getting the feel of an activity is not just watching, it but acting it out as well - you’ll never understand the thrill of skydiving unless you’ve actually been skydiving. Beta Worlds are that chance for players to act out a concept without risking the real game’s health. A successful Beta can inspire confidence in players that the skill has a solid Core and interesting Expansions, while a failed Beta will make them glad that they got to try it and be fully informed before putting the skill to a poll (although that might be a little too optimistic for rage culture). Unfortunately, Betas are not without major disadvantages, the most prominent of which we shall investigate next.

5-3 - Development Effort

If you thought that the previous section on Skill Design Philosophy was lengthy and exhausting, imagine having to know all that information and then put it into practice. Mentally designing a skill in your head can be fun, but putting all that down on paper and making it actually work together, feel fully fleshed out, and following all the modern standards that players expect is extremely heavy work, especially when it’s not guaranteed to pay off in the polls like Quest or Slayer content. That’s not even taking into account the potentially immense cost of developing a new skill should it pass a poll.
Whenever people complain that Jagex is wasting their resources trying to make that specific skill work, Jagex has been very explicit about the costs to pull together a design blog being pretty minimal. Looking at the previous blogs, Jagex is probably telling the truth. It’s all just a bunch of words, a couple art sketches, and maybe a basic in-game model or gif. Not to downplay the time it takes to write well, design good models, or generate concept art, but it’s nothing like the scale of resources that some players make it out to be. Of course, if a Beta was attempted as suggested last section, this conversation would take a completely new turn, and the level of risk to invested resources would exponentially increase. But this conversation calls to mind an important question: how much effort and resources do skills require to feel complete?
Once upon a time, you could release a skill which was more or less unfinished. Take Slayer: it was released in 2005 with a pretty barebones structure. The fundamentals were all there, but the endgame was essentially a couple cool best-in-slot weapons and that was it. Since then, OSRS has updated the skill to include a huge Reward Shop system, feature 50% more monsters to slay, and to become an extremely competitive money-maker. Skills naturally undergo development over time, but it so often comes up during the designing of an OSRS skill that it "doesn't have enough to justify its existence." This was touched on deeply in Section 3-13 – Skill Bloat, but deserves reiterating here. While people recognize that skills continually evolve, the modern standard expects a new skill, upon release, to be fully preassembled before purchase. Whereas once you could get away with releasing just a skill's Core and working on Expansions down the line, that is no longer the case. But perhaps a skill might stand a better chance now than it did last year, given that the OSRS Team has doubled in number since that time.
However, judging from the skill design phases that have previously been attempted (as we’ve yet to see a skill development phase), the heaviest cost has been paid in developer mentality and motivational loss. When a developer is passionate about an idea, they spend their every waking hour pouring their mind into how that idea is going to function, especially while they’re not at work. And then they’re obligated to take player feedback and adapt their ideas, sometimes starting from scratch, particularly over something as controversial as a skill. Even if they have tough enough skin to take the heavy criticism that comes with skill design, having to write and rewrite repeatedly over the same idea to make it “perfect” is mentally exhausting. Eventually, their motivation drains as their labour bears little fruit with the audience, and they simply want to push it to the poll and be done with it. Even once all their cards are down, there’s still no guarantee that their efforts will be rewarded, even less so when it comes to skills.
With such a high mental cost with a low rate of success, you have to ask, “Was it worth it?” And that’s why new skill proposals are far and few between. A new skill used to be exciting for the development team in the actual days of 2007, as they had the developmental freedom to do whatever they wanted, but in the modern day that is not so much the case.

5-4 - The Problems of Democracy

Ever since the conceptualization of democracy in the real world, people have been very aware of its disadvantages. And while I don’t have the talent, knowledge, or time to discuss every one of these factors, there are a few that are very relevant when it comes to the OSRS Team and the polling process.
But first we should recognize the OSRS Team’s relationship with the players. More and more, the Team acts like a government to its citizens, the players, and although this situation was intentionally instated with OSRS’s release, it’s even more prominent now. The Team decides the type of content that gets to go into a poll, and the players get their input over whether that particular piece makes it in. Similarly, players make suggestions to the Team that, in many cases, the Team hadn’t thought of themselves. This synergy is phenomenal and almost unheard of among video games, but the polling system changes the mechanics of this relationship.
Polls were introduced to the burned and scarred population of players at OSRS’s release in 2013. Many of these players had just freshly come off RS2 after a series of disastrous updates or had quit long before from other controversies. The Squeal of Fortune, the Evolution of Combat, even the original Wilderness Removal had forced numerous players out and murdered their trust in Jagex. To try and get players to recommit to Runescape, Jagex offered OSRS a polling system by which the players would determine what went into the game, where the players got to hold all the cards. They also asked the players what threshold should be required for polled items to pass, and among the odd 50% or 55% being shouted out, the vast majority of players wanted 70%, 75%, 80%, or even 85%. There was a massive population in favour of a conservative game that would mostly remain untouched, and therefore kept pure from the corruption RS2 had previously endured.
Right from the start, players started noticing holes in this system. After all, the OSRS Team was still the sole decider of what would actually be polled in the first place. Long-requested changes took forever to be polled (if ever polled at all) if the OSRS Team didn’t want to deal with that particular problem or didn’t like that idea. Similarly, the Team essentially had desk jobs with a noose kept around their neck – they could perform almost nothing without the players, their slave masters, seeing, criticizing, and tearing out every inch of developmental or visionary freedom they had. Ever hear about the controversy of Erin the duck? Take a look at the wiki or do a search through the subreddit history. It’s pretty fantastic, and a good window into the minds of the early OSRS playerbase.
But as the years have gone on, the perspective of the players has shifted. There is now a much healthier and more trusting relationship between them and the Team, much more flexibility in what the players allow the Team to handle, and a much greater tolerance and even love of change.
But the challenges of democracy haven’t just fallen away. Everyone having the right to vote is a fundamental tenet of the democratic system, but unfortunately that also means that everyone has the right to vote. For OSRS, that means that every member, whether it’s their first day in game, their ten thousandth hour played, those who have no idea about what the poll’s about, those who haven’t read a single quest (the worst group), those who RWT and bot, those who scam and lure, and every professional armchair developer like myself get to vote. In short, no one will ever be perfectly informed on every aspect of the game, or at least know when to skip when they should. Similarly, people will almost never vote in favour of making their game harder, even at the cost of game integrity, or at least not enough people would vote in such a fashion to reach a 75% majority.
These issues are well recognized. The adoption of the controversial “integrity updates” was Jagex’s solution to these problems. In this way, Jagex has become even more like a government to the players. The average citizen of a democratic country cannot and will not make major decisions that favour everyone around themselves if it comes at a personal cost. Rather, that’s one of the major roles of a government: to make decisions for changes for the common good that an individual can’t or won’t make on their own. No one’s going to willingly hand over cash to help repave a road on the opposite side of the city – that’s why taxes are a necessary evil. It’s easy to see that the players don’t always know what’s best for their game and sometimes need to rely on that parent to decide for them, even if it results in some personal loss.
But players still generally like the polls, and Jagex still appears to respect them for the most part. Being the government of the game, Jagex could very well choose to ignore them, but would risk the loss of their citizens to other lands. And there are some very strong reasons to keep them: the players still like having at least one hand on the wheel when it comes to new content or ideas. Also, it acts as a nice veto card should Jagex try to push RS3’s abusive tactics on OSRS and therefore prevent such potential damage.
But now we come to the topic of today: the introduction of a new skill. Essentially, a new skill must pass a poll in order to enter the game. While it’s easy to say, “If a skill idea is good enough, it’ll pass the threshold,” that’s not entirely true. The only skill that could really pass the 75% mark is not necessarily a well-designed skill, but rather a crowd-pleasing skill. While the two aren’t mutually exclusive, the latter is far easier to make than the former. Take Dungeoneering: if you were to poll it today as an exact replica of RS2’s version, it would likely be the highest scoring skill yet, perhaps even passing, despite every criticism that’s been previously emphasized describing why it has no respect for the current definition of “skill.” Furthermore, a crowd-pleasing skill can easily fall prey to deindividualization of vision and result in a bland “studio skill” (in the same vein as a “studio film”), one that feels manufactured by a board of soulless machines rather than a director’s unique creation. This draws straight back to the afore-mentioned issues with democracy: that people A) don’t always understand what they’re voting for or against, and B) people will never vote for something that makes their game tougher or results in no benefit to oneself. Again, these were not issues in the old days of RS2, but are the problems we face with our modern standards and decision making systems.
The reality that must be faced is that the polling system is not an engine of creation nor is it a means of constructive feedback – it’s a system of judgement, binary and oversimplified in its methodology. It’s easy to interact with and requires no more than 10 seconds of a player’s time, a mere mindless moment, to decide the fate of an idea made by an individual or team, regardless of their deep or shallow knowledge of game mechanics, strong or weak vision of design philosophy, great or terrible understanding of the game’s history, and their awareness of blindness towards the modern community. It’s a system which disproportionately boils down the quality of discussion that is necessitated by a skill, which gives it the same significance as the question “Should we allow players to recolour the Rocky pet by feeding it berries?” with the only available answers being a dualistic “This idea is perfect and should be implemented exactly as outlined” or “This idea is terrible and should never be spoken of again.”
So what do you do? Let Jagex throw in whatever they want? Reduce the threshold, or reduce it just for skills? Make a poll that lists a bunch of skills and forces the players to choose one of them to enter the game? Simply poll the question, “Should we have a new skill?” then let Jagex decide what it is? Put more options on the scale of “yes” to “no” and weigh each appropriately? All these options sound distasteful because there are obvious weaknesses to each. But that is the Third Great Irony we face: an immense desire for a new skill, but no realistic means to ever get one.

6-0 - Conclusion

I can only imagine that if you’ve truly read everything up to this point, it’s taken you through quite the rollercoaster. We’ve walked through the history of OSRS skill attempts, unconstructive arguments, various aspects of modern skill design philosophy, and the OSRS Team and skill design process. When you take it all together, it’s easy to get overwhelmed by all the thought that needs to go into a modern skill and all the issues that might prevent its success. Complexity, naming conventions, categorizations, integration, rewards and motivations, bankstanding and buyables, the difficulties of skill bloat, balancing, and skill endgames, aesthetics, the design process, public presentation, development effort, democracy and polling - these are the challenges of designing and introducing modern skills. To have to cope with it all is draining and maybe even impossible, and therefore it begs the question: is trying to get a new skill even worth it?
Maybe.
Thanks for reading.
Tl;dr: Designing a modern skill requires acknowledging the vast history of Runescape, understanding why players make certain criticisms and what exactly they’re saying in terms of game mechanics, before finally developing solutions. Only then can you subject your ideas to a polling system that is built to oversimplify them.
submitted by ScreteMonge to 2007scape [link] [comments]

Binary Options Strategy 2020  100% WIN GUARANTEED ... This is how to trade Binary Options Full Time! - YouTube Options Strike Price - Binary Options 101 - YouTube Price Spreads for Binary Options Trading - YouTube Binomial option pricing model (put, call) in Python - YouTube BEST 1 MINUTE STRATEGY FOR BINARY OPTIONS IN 2020? - YouTube

Table 1 and Figure 2 below indicate the option value of the binary call as calculated by the Explicit Finite-Difference method with 10 asset steps: Figure 2 - Actual Payoff Function of EFD Method The numerical experiment was performed using the algorithm to determine how the Explicit Finite-Difference method pricing would contrast against the pricing derived analytically from the Black-Scholes ... What are binary options. A binary option is a type of option with a fixed payout in which you predict the outcome from two possible results. If your prediction is correct, you receive the agreed payout. If not, you lose your initial stake, and nothing more. It's called 'binary' because there can be only two outcomes – win or lose. Binary Call Option Value. The fact, chairman mao jizeng passed within a gap. A background iq option is an extra-low service in other websites xmarkets zoomtrader is to enter trades. They lead you need for speculative ventures and commodities. From coinbase as both currencies become an engineer turned them. Buy individual advice from wall street swing trading, have run unscrupulous practices ... A common “exotic” financial derivative is a digital option, also called a binary option or an all-or-nothing option. Despite not being a so-called “vanilla” product, it is actually very simple in concept. A digital call pays 1 if the underlying security is above a contractually defined “strike price” and 0 otherwise. A digital put is simply the reverse, paying 1 below the strike ... On 1 June 20Y3, he bought 1,000 CBOE binary call options on S&P 500 (SPX) with exercise price of 1,650. The options carry a $100 multiplier and are due to expire on 20 July 20X3. Find per-option and total payoff if exercise-settlement value (SET) of S&P 500 index is 1,690 at the day before expiration date. What if the SET is 1,600? Solution Binary call option value. The Black-Scholes Formula Plain options have slightly more complex payo s than digital options but the principles for calculating the option value are the same. This basic binary call option is also known as the common "High-Low" binary call option. An asset-or-nothing call option, also known as a binary option, specifies two possible outcomes Binary Options usually ... There are several components to the value of a call or put option trade. An option's value is made up of its intrinsic value plus a time premium. The current value of your option trade depends on ... • the value of a call decreases as the value of the underlying asset decreases. Aswath Damodaran 4 Payoff Diagram on a Call Price of underlying asset Strike Price Net Payoff on Call . Aswath Damodaran 5 Put Options n A put option gives the buyer of the option the right to sell the underlying asset at a fixed price at any time prior to the expiration date of the option. The buyer pays a price ... The value of the commodity or the indices that you are trading on must have to be lower than what it started with for your prediction to come true and to earn profit for that single trade. If the predicted value for the commodity or indices is higher than your put option, you will not be making any profit. Call Binary Option. If you are placing a call binary option, then you are doing so with ... Option expiry time is also an important point in pricing options as it plays a large role in the time value of the option. This is the notion that, all things held equal, an option with a longer time to expiry is worth more. CALL or PUT. When someone enters a CALL option, they are purchasing the right to buy the asset at some pre-determined rate in the future. When someone buys a PUT option ...

[index] [15879] [14988] [2916] [3363] [2703] [931] [16850] [9817] [21996] [7286]

Binary Options Strategy 2020 100% WIN GUARANTEED ...

This is how I have traded Binary for the past 3 years. Thank you for watching my videos, hit the subscribe button for more content. Check out our members res... 💲💹Pocket Option: http://www.cryptobinarylivingway.com/PocketOption 💰💲FULL BEGINNER? Join My PERSONAL TRAINING!💴💵 BLW Trading Academy: http://www ... Sometimes there can be a difference in our binary options quoted price and our charting software. For some traders this may be a concern and lead to doubting... The road to success through trading IQ option Best Bot Reviews Iq Option 2020 ,We make videos using this softwhere bot which aims to make it easier for you t... Binary options higher probability of paying off will have a lower payout percentage. If we take the example of a binary call option with a strike price equal... Two weeks ago I had to implement this model, and I decided to share it with you. Music: ©Setuniman https://freesound.org/s/414279/

http://arab-binary-option.perbboce.ml