Thursday, March 16, 2017

Powershields Reference Post

Powershields Reference Post
- - -

• To get a normal powershield, you must transition from no shield on one frame to a full digital press on the next frame. This is difficult on unmodded controllers. It is easier on controllers with lubricated trigger tracks/weakened springs or controllers that have been plugged into the system with the triggers mostly depressed. PSs are easiest on controllers with no spring.

• PSs CANNOT be buffered shields. For example, if you wavedash and continue to hold down the trigger after the airdodge then it will buffer a shield on the first frame after your landing lag but this shield will NOT be able to powershield. To PS on the first frame out of endlag you'd need to press shield on that frame exactly.

• The window for reflecting a projectile is 2 frames. That is, the projectile must collide with the reflect sphere (an area within the digital shield sphere) on the first or second frame that it is active.

• The window for PSing a physical attack is 4 frames. The hitbox must collide with the PS sphere (i.e. your normal, digital shield) on frames 1-4 of an active shield.

Note: Because the reflect sphere is slightly smaller than the shield itself, PSing projectiles can be inconsistent unless the shield appears over the hitbox. If you try to hold a stationary shield and allow the hitbox to collide naturally then you run a small risk of the hitbox colliding with the shield, but not deep enough on that frame to connect with the reflect sphere. This is partly why some PS setups (including crouching under lasers as puff/sheik/marth) seem so much more consistent.



• PSing a projectile reflects a projectile AND skips any hitlag/shieldstun/pushback. You are immediately free to do any normal OoS action.

• PSing a physical attack has a normal amount of hitlag/shieldstun. There is added shield pushback. After shieldstun, Guardoff (dropping shield animation) can be canceled with any A or B attack or jump. In order to cancel Guradoff, you must have no more than 3 frames of shield active after shieldstun.

• PSs do not take shield damage.


- - -




There are several exploits that allow for slightly different versions of a PS. They are listed and detailed below:

• ADT-PS (Analog Digital Transition Powershield)
execution: one frame of lightshield before digital shield.
The reflective sphere on frames 2 and 3 (frames 1 and 2 of hard shield) mimics the size of the initial lightshield. However, The window for powershielding physical attacks is reduced from 1-4 to 3-4, and the shield actually has downtime vs physical attacks after the digital press.

• DAT-PS (Digital Analog Transition Powershield)
execution: digital press at the same time as light press, then release digital press within the PS window.
DAT-PS increases the size of the PS sphere to match a lightshield on frame 3. The analog press may be 1f earlier than digital press but the window for a PS remains the appropriate 2 or 4 frames from when the shield activates. As the size increases after the PS window for a projectile, this is not an effective technique for reflecting projectiles, despite a false visual cue that in real time makes it appear as if the reflect sphere gets bigger on frame 2 if a projectile is reflected on that frame. It does not. DAT-PS increases shieldstun to match that of the lightshield after it expands on frame 3.
This technique has niche but profound uses and is relatively easy.
https://www.reddit.com/r/smashbros/comments/2rkczd/light_powershield_possibly_new_technique/


• Z-PS (Z-Powershield)
execution: while holding A, press a digital shield and Z on the same frame.
Z-PS increases the size of the PSsphere proportionally to the lightest lightshield on frame 1. It has shieldstun in accordance with the lightshield (about double) and dramatically increased pushback. It has wide application, but the need for a buffered A press makes it awkward to use in competitive play.
https://smashboards.com/threads/official-ask-anyone-frame-things-thread.313889/page-17#post-18512581


• ADT-Z-PS (Analog Digital Transition Z-Powershield)
execution: input analog shield, then a digital press and Z on the following frame. This negates the need to buffer an A press to perform a ZPS. Additionally, ADT-Z-PS has a 1-2 frame window of downtime from physical attacks after the digital press. However, this is not very significant in context and the bigger drawback is that it will remain to some level inconsistent due to the nature of one frame links.


• <3 AKA Heart Shielding (PS storing)
If you perform a frame 1 PS on a physical attack that does <3% OR miss a frame 1 PS against a projectile that does <3% because it collided with shield but not the reflect sphere then the next physical attack to connect with the shield will be PSed, even if it is transitioned to lightshield. This happens most frequently vs a falco with stale lasers.
Because Yoshi’s shield acts differently/buggy, he has different conditions and greater potential for expanded PS storing.
https://www.reddit.com/r/SSBM/comments/4yd5kb/heart_shield_powershield_storing/
https://smashboards.com/threads/some-yoshi-information.351635/page-9#post-21400329


- - -

None of this information is original to me. I am simply gathering some common, not-as-common, and should-be-common knowledge into one place for ease of reference. TY to @taukhan for proofreading. The bulk of this can be further researched from kadano’s post here https://smashboards.com/threads/official-ask-anyone-frame-things-thread.313889/page-17#post-18512581

Tuesday, February 28, 2017

Analysis for maXy

https://flowfeedback.com/feedback/8xvPpNDza7a2fTvsY
Analysis of maXy(puff) vs MINT(fox) at Schism
will do your sets for $10/game

Saturday, February 11, 2017

Analysis for Denko

Denko vs Ichigo (2/2/17) Analysis

Notation by in game timer

Game 1
BF
8:00 both players open waiting for the other to overextend in neutral
then aerial in at the same time lol. The better decision was to approach from the ground ready to shield/WD back. Same objective (take center) but safer.
7:55 bad spacing, can’t be under falling falcon without a shield at least.
7:51 this bait and bair were fine, the bair was just started a little too slow
7:49 you need to be super careful commiting to falling upair after bair. Good players will shield and then punish it hard.
7:42 don’t ever pound here. It’s tempting but really predictable and easy to punish. Instead try to purchase space to land by threatening bairs.
7:38 instead of downthrow you should upthrow falcon. Better followups.
7:35 nice nair. Because you started it before falcon was even able to start an aerial it was going to win. A little slow to follow it up though.
7:31 good mash out of dthrow
7:30 you are habitually crossing up the shield every time you hit it. Crossups are generally unsafe until you condition your opponent not to expect it. Additionally, they’re not really worth the risk much of the time anyway. Instead try to prioritize a low as possible bair.
7:12 you get a heavy punish on an input error. Should be an easy edgeguard.
Good job grabbing ledge, but now because you know he has to upB instead of fairing him on his way up you can actually just chase him. If you had just ledgedashed and then jumped then you could have baired him off into another edgeguard.
7:08 great recognition of the situation and grab. Backthrow was correct at this % and I like your prioritization of the ledge. Again though, an immediate fair has a low reward. In this case because there wasn’t time for anything more I would have done a dair -> grab -> another backthrow.
7:03 you ran forward when it wasn’t safe. By this time you can recognize that this falcon player always jumps at you when you’re in the corner, even if he waits a moment before he does it. That means you can just wait for him to come and then grab him.
6:57 Falcon players like to stomp on top platform between stocks. Never approach this stomp from below. Instead, go to the side platform and come up beside him with bair. Aim for his elbows. Just be ready to fade away if he chooses to drop below so that you can’t get upaired.
6:55 this dair could have been grabbed on reaction very easily.
6:54 you had time for a second fair. Instead you do a weird fsmash. Not sure why.
6:50 because falcon is off and below the stage and has to use his jump right away it is safe to chase after him with a fair without grabbing ledge first. You should look at that. First bit of homework for you: Watch this set back and pause every time you get an edgeguard situation. Then write down the best possible thing you could have done to cover ALL of his options or (what happens immediately after them) (double jump, aerial, fast fall, upB, airdodge). You’ll notice explicit patterns.
6:44 Again, this dair is grabbable. You have a habit to aerial out of shield. Try to do the best thing in every scenario until it becomes your new habit.
6:39 See how when you are out of his range he always waits to see if you will jump forward then gets impatient and jumps at you himself? This is exploitable.
6:30 very risky fair. Use bair to purchase space to land. That’s your only goal when off stage.
6:25 good spacing.
6:15 ah, big opportunity but flubbed movement keeps you from getting the grab.
bair afterward is a little slow.
6:14 Good jump but he is ready to shield. In this situation you have to sometimes fast fall and grab or he will shield every time and win this scenario 100% of the time.
6:10 remember that you can crouch cancel these nairs.
6:05 nice crouch, nice rest.
6:00 you shield, but then change your mind and try to jump. Don’t change your mind! Have some confidence. If you’re right you’re right and if you’re wrong you’re wrong, but if you change your mind then something awkward and bad for puff will happen pretty much every time.
5:57 was a grab, not a fair. Fair is hella slow compared to grab.
5:55 lol looks like a shieldpoke on your tiny little puff ears
it wasn’t the best roll anyway. Just back off a bit and give falcon the chance to mess up, no need to rush into center stage when he’s in control.
5:49 you overextend with your inv and get hit out of it
your goal is to use inv to get center or a clean hit that will lead into a combo, not both.
5:44 your rising bair is too slow again. You need to practice that.
5:42 blegh, good pressure but bad spacing.
you keep trading afterward because you two keep having the same idea, haha
5:30 good reaction to his tech
5:27 you get pretty lucky with that pound. He happened to be really late to react to your going off stage.
5:20 he’s unsure of how to get you off the ledge and gets distracted into letting you crossup.
5:18 this time you waited for his approach but misspaced when it came.
5:17 this fair is really bad. It doens’t threaten anything but a weird preemptive empty jump and is very punishable on whiff. Instead turn your back and purchase space.
5:15 unsafe pound.
5:10 lmao baited.
5:04 wow. He either read you like crazy or messed up his spacing and got crazy lucky.
4:58 poor DI after the knee. DI down and away, mash either jump or rest, if you see the knee connect with you then DI up and in for survival.
You also need to use pounds on your way back (as long as you’re out of range to get punished) so that you don’t have to make a predictable straight line for the ledge right away. You almost died there.
4:50 falcon shieldgrabs too late and gets uptilted. If he was early enough then you easily could have died instead.

Stocks Taken:
Bair whiff punish on ledgeknee
Falcon wanders into a pound
uptilt rest
uptilt rest

Stocks Lost:
Poor spacing vs falling bair
Roll into a shieldpoke knee
Unsafe recovery pound.


Ok so I have a good understanding for the themes here.
A group things that you have to fix, one at a time.

1) Don’t cross up every time you hit a shield. There’s no good reason to most of the time. Instead, space a low and spaced bair if you think they will WD/drop shield, an upair if you think they will jump, or a fast fall empty land grab if you think they will keep shielding. Those three will get you much much more reward much much more often. There’s no way around it, just practice. Play a session where you just play your normal game other than focussing on doing this instead of unnecessary crossups. Sometimes a crossup is the right call, but most of the time it isn’t.

2) Practice your aerials in frame counter
https://www.youtube.com/watch?v=HMUC9rH4NSQ
your rising aerials are too slow and it’s making a difference. You can also practice just general movement and fastfalling as soon as possible in frame counter. I would tell you the numbers for exercises but a) I forgot where I wrote them down (lol) and b) it’d be like cheating, you’ll learn way way more looking them up yourself so I’m going to make you do that. Anything that you want to practice in frame counter, go into debug mode and count in frame-by-frame how a frame perfect jigglypuff would do it. Then write those numbers down, plug them into frame-counter mode and practice. It’s AMAZING. By far the best way to practice your movement/timing. If you have any questions about setting up frame-counter or debug mode just message me and I can walk you through it. I’ll tell you though that I learned the most I ever learned when I started using debug mode to try and look for what a theoretical god puff would do in common situations.

3) As I mentioned, watch the vid over and pause every time there’s an edgeguard. Write down the “correct” response.
Generally: Falcon is going to go high or low. You’re going to go to either the ledge or to right outside of his aerial range. Then you’re going to punish his commitment.
Double jump/DownB: reposition to cover his next option
UpB: bair so that if he fades forward the base of your bair hits him, if he fades back then the tip does.
Aerial, bair/fair his endlag
airdodge/upB that you can’t reach: punish his landing lag.

4) When you are recovering only pound when it is not punishable. Generally, try to stay high just above his FH upair range when possible. Postpone making a commitment to ledge or to top/side platform until the last possible moment. Wait to see if he will slip up his spacing and give you one of the two for free. Remember that your only goal is to touch the ground safely. Don't get baited into trying to hit him. Turn your back so that he backs off a little bit, touch ground /then/ get back into the game.

The MU:
Most of this MU is made much simpler by keeping a specific spacing goal. Try to stay either
a) in the air outside of his nair range or
b) right at the edge of it on the ground.
It’s not always easy because he can DD but that’s the goal. When you are there your goal is to punish any big commitments that he makes on reaction. You will shield grab/CC/or aerial whiff punish any approach that he does.
Remember, if you crouch liberally and are confident in resting whiffed grabs then falcon has to jump to approach you. Additionally, his aerials are all slow and reactable. That means that you can either CC them or shieldgrab them all on reaction if you are on the ground, so being just inside of his SHFFL range isn’t a problem for you. If you are outside of it /then/ jumping is safe. In order to punish this kind of range falcon has to start DDing aggressively and start doing fakeout SHFF empty land grabs. Your answer to that is to use WDs to fine-tune your spacing when on the ground and, if flacon should truly refuse to commit to approaches, start to use spaced bairs or frame advantage to inch him into the corner where FH fairs will shut down his range and force him into an easy edgeguard.
In the next game I’ll point out instances to use parts of this broad strategy.


Game 3
FoD
8:00 Falcon starts the game off really aggressively, whiff a whiff punish so he gets a ledgecancel bair into a CC stomp upair knee. Ouch. You just didn’t react to the hits fast enough to get hard enough away DI to escape the combo. Not a big deal. Will come with experience thinking about it.
7:50 you bait falcon to the ledge and get a fair. In this case upair would have been the better punish, if you were looking for a whiff punish there.
7:49 you give up a bit more space than you need to but that ends up being nice because he runs right into it with a grab. You should have grabbed after he whiffed then gotten an edgeguard. Instead he gets another ASDI down punish.
7:41 the upair uptilt wasn’t a bad idea, he got a lucky timing to get through
7:40 sick nair. This is a good example of using frame advantage to gain space. Because he committed to dash forward FH he couldn’t turn around, jump, and aerial the right space before your nair came out. You don’t whiff punish his counter attack cleanly enough but that’s ok because you got the information that he’s just going to keep stomping when he’s uncomfortable. Try to whiff punish it with an upair next time, since you know it’s coming.
7:37 ah, speak of the devil
nice combo extension with the upair bair
7:25 good patience coming back, no rush.
great spacing outside of his aerial to get that grab. Exactly what you want to be doing.
7:21 A little premature punish vs the aerial.
7:18 there’s the stomp, grab.
you can space the upair to poke shield.
7:15 great spacing outside of his aerial again. See how simple it can be? You miss the grab, but that’s just execution.
7:08 def should have rested after the DI in on dthrow.
7:06 again, spacing outside of his counter attack aerial and getting a punish
7:01 would have been another but your bair is slow
6:51 terrrrrrrrible pound
6:46 don’t worry about fairing, just stay safe until you’re back at your position goal. FH fair is not safe.
6:42 upair was way too slow
6:35 miss an edgeguard opportunity by fixating on ledge, should have been spacing outside of jump and then kicking it. You get the pickup after getting upBed though.
6:29 sloppy crossup gets you hit and bad combo DI gets you hit some more. DI down and away to avoid the combo.
6:23 supppper unsafe nair. Doesn’t work vs falcon, his upair is too big.
6:20 you go into the air inside of his SHFF and get kneed. See how if you stayed on the ground it would have been a shieldgrab?
6:12 bair is too slow again
6:11 don’t doubt yourself. if you want to shield then shield.
6:10 bad spacing to jump.
6:06 bunch of meh fairs. If you instead focus on position then you’ll get interactions with more punish potential.
6:03 there it is. See how because you jumped from the plat you were outside of the SHFFL range and your aerial whiff punished?
6:01 good grab, wrong throw.
Nice read.
6:56 not the best DI. Could have been bad. DI out harder.
you establish your space but then give it up a little hastily, then the recording ends at 5:49 so I can’t comment on the end of the match. In any case it’s likely that you just took too much damage from the scuffles after misspacing and lost?

Stocks Taken
upair whiffed grab, WoP
bair, edgeguard
techroll read rest

Stocks Lost
ASDI downs your fair, dthrow knee
whiff upair, get kneed
whiff bair, get kneed

The takeaway: look at how much potential you get when you just focus on staying at the outside of falcon’s aerial range. You don’t have to worry about what to do in scuffles if instead of entering into them you just back off, give your character a moment to get into position, and start whiff punishing again. Once you force falcon to respect your punish game off of whiff punishins /then/ your FH aerials will start to be very effective.

systems model of creativity notes

Systems Model of Creativity
https://books.google.com/books/about/The_Systems_Model_of_Creativity.html?id=DuFsBgAAQBAJ&printsec=frontcover&source=kp_read_button#v=onepage&q&f=false

raw notes:

Herbert Simone: given the correct information(problem), a computer can solve a problem (such as the elliptical paths of the planets) in a tiny fraction of the time it takes a human. That is, if creative discovery is just fast problem-solving then computers are also capable of creativity. You run into presented problems vs discovered problems.

creativity: flow and the psychology of discovery and invention
systems model of creativity

musicians improvising vs executing highights DLPFC region of the brain, poker but not chess, having to make decisions with  insufficient knowledge, discovering problems

art that discovers a problem through its process rather than executing a pre-conceived plan, is universally rated as more valuable/original

Where is creativity?

creativity is, unfortunately not a process that can be identified or possessed. Rather, it is a word used to describe the event in which an agent for change makes innovations within a field, located between the individual, the field, and the cultural domain.
This is Csiksezentimihalyi’s counter to Herbert Simon’s computer program capable of replicating creative scientific problems. Simon claims that if it was creative for Newton to draw up a formula, the exact same formula from a computer is also creative. Csiksezentimihalyi claims that this is not the case because creativity is defined by its context. Thus, van Gogh is more “creative” than a van Gogh forger. In this way, according to Csiksezentimihalyi, the object itself is no judge. In the contemporary, Giotto is boring and of questionable worth. In context, that is, after we are instructed of its art historical or religious or aesthetic value, then its creative value changes. “In the contemporary” requires a context of its own that lies totally outside of the work itself. The field and the domain are necessary for creativity to exist. It— we could say like art— is invented. Social construct.
Consider Mendel’s contributions to genetics. They were not appreciated, not even by Mendel himself, until 40 years after his initial experiments at which time the theory of natural selection and variation had a need for them. Where is creativity? In the experiments, in Mendel, or in the framework that needed them? It is inseparable from any, as long as they are together.

It is important too to recognize that we don’t want just new ideas— we want new good ideas. And it is the realm of the realm to determine goodness.

Enter, the gatekeeper. The popes made art history. A handful of gallerists made art history. Greenberg made art history. Saatchi/Gagosian are making art history. The establishment is the domain. The establishment creates taste. The establishment rules on what goodness is. Leonardo traveled in accordance to which patron had the more money.
Sometimes, in the event of a radically new field, the domain widens a bit.

creativity (as a construction) emerges over time.

the system, the construction, the architecture of related ideas forming a canon is domain.
It is cultural heritage and in evolutionary terms, the convenient packaging of extra-biological ideas worth passing on, called memes.

Florentine artists weren’t exceptional, Florence was. Nor was this accidental. It was a conscious, calculated policy decision on the part of the oligarchy in en effort to build a new Athens. It supported its arts financially, educationally, culturally, and this paid off. It set a task and the artists fulfilled.

Questions: What are the ways in which information is stored and how does the structuring of the information affect creativity?

innovation, creatvity, problem finding, originality
formulating something in a strikingly better way

following domain of the domain
creativity is a subjective judgement. It is not an objective quality. Judges will assess it differently according to their past experience and even their personal idosyncricities, even when judged to be experts on creativity themselves. The creativity is a product of a social sphere.
So the personal creativity (a second individual individually “discovering” Einstein) must be paired with the Persuasion to be recognized. “In science the credit goes to the man who convinces the world, not to the man to whme the idea first occurs.” If you define creaativity as making innovations within a field that are then incorporated into the domain, that is, within the systems model, persuasion is an intregal part. So close in fact that Csikszentmihalyi says they are inseparable as far as the empirical is concerned, for if you can’t persuade then who is to say that you have at all? If we are to pin it down this seems the simplest way. To call it creative is to call it a worthy change to fit into the domain. Creative is here maybe better called genius. It must affect change.

This model appears problematic to me in that it is circular. It MUST exist in the way that it does and it identifies creativity as only that which works. “Of course, one might disagree with this definition of creativity.”

Without rules there cannot be exceptions and without tradition there can be no novelty.

creativity in 4 stages
1 preceding hard work and research, preparation
2 idle time alone, incubation
3 moment of insight, illumination
4 hard work and elaboration to bring to fruition, verification

“anything that is attractive has to have an aspect of obviousness”

“If you cannot persuade the world that you had a creative idea, how do we know that you actually had it? And if you do persuade others, then of course you will be recognized as creative.” So freakin problematic. Downplays inherent motivation/reward as somehow necessarily secondary to external. It’s not even the fruits, it’s the gregariousness of the thing that determines creativity. And if that’s not what creativity is then well hell, let’s just move the goalposts. He assumes that it’s the only useful or measurable creativity. I don’t accept that.

Likens it to evolution. Variation that is tested then if successful integrated by being passed on.

imminent individuals are flexible and curious
they are introverted exactly when they need to be
they are extroverted exactly when they need to be
and intensely
optimal?
their behavior is determined not by rigidity but by the demands of the domain/moment/value


if a structure is not capable of recognizing let alone accepting change then change can’t exist. Thus, the more expedient action is to alter not the input but the receptivity. Agents will follow. So we practice mindfulness and ask more questions.

ah, Csikszentmihalyi’s model is static
by having a defined field and domain it is so heavy-handed. Will make its point at the cost of nuance. It is a bit determinist, which is fine. It seems that the entire purpose and utility is to direct conversation away from the idea of the promethean genius. That is obv progressive, but a bit… I don’t think it’s the best response.
Strongly prefer nietzsche’s 500 hands. It’s more artful, appropriate, and so much less stupid. Nietzsche makes each a servant of the other in paradox, which is what it seems that Csikszentmihalyi means to suggest but tramples on.

Tuesday, January 31, 2017

GTO and Exploitative Play

GTO and Exploitative play

Today I’m going to expand on a dichotomy that has been largely subtextual but would be more useful made explicit.

GTO
In poker there is a concept known as GTO, or Game Theory Optimal. In a GTO model every decision is optimal because it cannot be exploited by an opposing player even if he has full knowledge of your gameplan because the risk-reward is entirely and mathematically accounted for.

The best illustration of GTO is the Prisoner’s Dilemma. The Prisoner’s Dilemma is easily solvable. A payoff matrix reveals that without knowledge of the opponent’s decision you should always betray them. This is the best decision precisely because it has the highest reward attached to an outcome that cannot be made worse (punished) by the opponent. Interestingly, humans have a cognitive bias toward cooperative behavior even though cooperation is in this case a mathematically losing strategy. This demonstrates the importance of actually making the matrix to determine the GTO in even this simple scenario.

Now, what if we turn to Rock Paper Scissors?
GTO for RPS is to throw rock 1/3 times, paper 1/3 times, and scissors 1/3 times in a random order. This has the highest reward attached to an outcome that cannot be punished. But if your opponent deviates from GTO then this is a little bit problematic. Let’s say that you face an opponent that abandons randomization and always throws rock. GTO demands that you ignore him and continue to randomize your throws. In the case that there is some knowledge of the opponent, GTO is in practice suboptimal depending on how you define optimal. This is of course what makes the a theoretical GTO so interesting in poker, a game in which results are measured in profit over time. Game Theory Optimal carries the highest profit with the least amount or risk but this does not not necessarily equal Most Profitable— in fact GTO only breaks even. Thus, if maximum profit is the goal then GTO is suboptimal in any case in which the opponent is not also playing GTO! By refusing to open yourself up to exploitation, you cannot exploit an opponent.

Exploitative Play
Exploitative play is, in a nutshell, recognizing risk in an opponent’s gameplan and compensating for it. Let’s say I recognize that my opponent throws rock every hand. Even though rock-only is exploitable, GTO cannot exploit it. In order to exploit rock-only I have to abandon GTO and adopt a more paper-heavy strategy. Once I do, provided that my opponent does not deviate from rock-only, Paper-heavy has an increased profit that is exactly as profitable as it proportionally favors paper. HOWEVER, in abandoning GTO to exploit my opponent’s strategy, I have adhered to a new strategy that is equally exploitable. It is entirely possible for my opponent to counter-adjust and switch to Scissors-only. That is the risk attached to abandoning GTO in search of profit. Your opponent may punish you at least as severely as you sought to punish them.

In summary:
GTO is maximizing profit by eliminating risk.
Exploitative play is further maximizing profit while inviting risk.


So what does this mean for Melee?


Potentially a lot. As I’ve repeatedly discussed, mixups are closely related to RPS. There is an inherent GTO. Adhering to or abandoning GTO for a more exploitative strategy is a judgement call that we always make deliberately, intuitively, or out of ignorance. It might be appropriate, it might not be. It’s a matter for individual assessment.

Here is what we should remember:

* GTO goes even unless you gain an unfair advantage, at which point GTO will always win over time precisely because it eliminates risk. It's specifically designed not to lose.

* Similarly, an optimized GTO model is more profitable than an underdeveloped GTO model.
If you are playing with rock (1pt), paper (1pt), and scissors (1pt) but your opponent is playing with rock (1pt), paper (1pt), and nail-clippers(.25 pts) then you win over the long term even without any exploitative play because you're using better options.

* Exploitative play requires that you understand your opponent’s strategy. You might consider it Attacking your Opponent’s Understanding. Maybe your opponent’s brain honestly believes that rock-only is optimal. Or maybe he’s just leading with rock to try and bait a paper switch. In a fighting game in which prepared reactions can trump a mixup scenario altogether there's a huge difference. In order to be successful, exploitative play requires 1) information and 2) acumen, otherwise it is not strategy, it’s just blind hope and high-risk variance.



Further reading:
https://arxiv.org/pdf/1404.5199v1.pdf and
http://poker.cs.ualberta.ca/publications/IJCAI03.pdf

Tuesday, January 10, 2017

Faster Improvement

Faster Improvement

Before you start, you have to accept a few base assumptions.
* In a game, “skill” is just your ability to execute winning tactics/strategy.
* In this context, “improvement” is synonymous with “learning skills.”
* Skill acquisition is a function of the accumulation of focus-intensive work, not directly of time.

Remember the Four Stages of Competence? It works nicely with these.
It, integrated as a cyclical model found in Improved Drastic Improvement, is the best methodology known to me. At its heart, this model simply asks that you:

* Identify a problem.
* Identify the solution.
* Practice the solution until it’s in your unconscious.
* Repeat.

Over time these correct solutions accumulate to form your unconscious gameplan. Each skill as learned individually measurably contributes to your results, ideally building on one another to create a juggernaut. With enough of the curated skills worked to the unconscious level, winning is inevitable.

But what does this process actually look like?

As of now, I think it’s best manifested in the following manner.

1) Record a (netplay/tournament/seriouslies) set.
2) Immediately break and identify the ONE most important lapse in your execution/strategy.
3) Identify the best tournament-viable solution to that lapse.
4) Practice executing this solution until it’s locked into your unconscious.
5) Repeat.

I’m sure that sounds a bit repetitive at this point, but in the past year or two the increase in popularity and infrastructure within the community has created a thriving netplay scene. Because you can record the equivalent of a tournament set vs a worthy human opponent at will it is now PRACTICAL to use the above model as your primary way to play, not just to color or elaborate on infinite friendlies. A like-minded friend that is able and willing to go through this process with you is obviously still a godlike asset, but it is a boon that this very rare kind of individual is no longer a requirement to go through it.

Before closing, I would like to make a few points.
* Working on exactly one issue at a time allows you to focus harder on it, increasing the efficacy of your practice.
* This issue could be technical/strategical/mental/health/attentional/etc. Anything that is an issue is an issue.
* Some issues might be completely solved in a matter of minutes, hours, or weeks. It might take some studying with debug mode, google, another smasher, or even a book to find the best possible solution for you. Who knows?
* The most effective practice with the smallest time commitment is two or three half-hour sessions spread throughout the day. Remember, intensity of focus is more important than time spent.
* What you choose to prioritize and what you honestly believe is the best solution will culminate into your personal style. It’s silly to worry about that because it will happen naturally.