Wednesday, June 28, 2017

10 Different Bairs

10 Different Bairs

Puff’s bair has decent frame data (hits 9-12, IASA at 31, AC at 25, 10f Lcancel lag) and good range (I know I know, it looks ridiculous when isolated but honestly that’s pretty standard for a good move in melee). But when you combine the raw frame data with Melee’s 1f of hitbox drag and Puff’s aerial mobility you end up with a move with great utility/flexibility.

In years past, Hungrybox would usually answer questions on smashboards with “Bair more.” The phrase wasn’t exactly wrong, but it was definitely unhelpful and contributed to a bad culture for learning. In order to bair better (or as Puff’s opponent, to punish bairs better), it’s important to recognize the different tasks that bair is used for.

To that end I’ve compiled a list of 10 fundamental uses for bair in neutral and attached a brief description to each.


1) retreating AC bair
This is basically a quick retreat that drags a huge hitbox behind you. Puff can choose mid-jump how much stage to give up based on risk-assessment. It is difficult to punish on reaction, but carries a stage cost. It can be punishable with a read. Note that with practice Puff can SH bair DJ bair. This makes it a little bit better on defense and offense.

2) advancing AC bair
This is a quick and flexible poke. Because Puff can be active as soon as f18 after hit (or SH bair DJ bair), it is particularly great for combos past a certain %. Puff can choose whether to continue to fade forward or begin to retreat as the hitbox comes out. Advancing AC bair is easily punished with FHs, CC, some OoS options, etc.

3) neutral rising FH bair
This is a flexible, relatively noncommittal poke. It’s generally used as a method to protect the ascent to FH height from an immediate attack from the opponent, at which point puff has more mobility (can immediately choose to go up, down, left, or right). It is difficult to punish, but can be blown up with by preemptive FH dair/nair or by running forward after the bair and pressuring Puff to retreat.

4) retreating rising FH bair
This is a quick retreat that is safer than an AC bair with a bit higher stage-cost and a different end position (puff at FH height rather than grounded). It is very difficult to punish directly but very easy to punish indirectly (just walk forward and take the free stage).

5) descending low bair
A spaced, low-as-possible, L-canceled bair is a great poke because it is almost never directly punishable except out of CC. It carries low reward until the % at which bair forces tech, at which point it earns at least hefty stage on hit. On some characters a low bair is safe on shield even when non-spaced provided that puff buffers a crouch. A descending bair is punishable with a preemptive attack/FH (think: dtilts), CC (difficult), or an indirect threat (example: Puff tipper low bairs Fox’s shield. He can’t get anything immediately OoS, but he can WD or SH and force her to shield/jump away/etc).

6) descending high bair fadeback
If puff chooses to bair at head-level as opposed to feet-level then she has time to mix up her drift pattern before landing. This kind of bair is especially strong against many character’s shields and can often shieldpoke, but because of the extra endlag from falling farther it is very bad against CC or on whiff.

7) descending high bair crossup
This bair is identical to the above but instead of fading back puff does a crossup. It is universally less safe and you normally don’t see it unless the puff player is a) bad or b) cutting corners because they’ve conditioned a level of respect. It’s worth noting here that a descending upair works in a similar way.

8) in place aerial DJ bair
When Puff is in the air, she can use a DJ bair to protect her air-space from attack. This is whiff punishable and can often be traded with.

9) fadeback aerial DJ bair
When Puff is in the air, she can use a fadeback DJ bair to protect her air-space from an attack more safely at the cost of a little bit more stage. The fadeback makes it effectively immune to trades and very difficult to whiff punish without a read on the drift.

10) low DJ bair
If Puff is descending (from any position, be it a FH, SH, or otherwise), she can DJ and bair at the last moment. This creates a situation very similar to a rising FH bair, with a couple notable differences. First, the bair does not hit as low. It won’t hit many crouching characters and will whiff completely vs shorter characters. Second, it comes out faster by virtue of not having any jumpsquat. This bair is often used to the FH effect and/or as a read on an opponent's FH, especially in the corner.


The next time you play or watch Puff, keep an eye out for these techniques and watch how they accent puff’s other options. Additionally, watch for any opportunities that her opponent takes or loses to punish any obvious bairs via option coverage or reads.

Friday, June 23, 2017

How to Combat Puff Planking (and Save Melee Kappa)

How to Combat Puff Planking (and Save Melee Kappa)

In order to promote understanding, jump-start meta development and accurately demonstrate planking’s strengths and weaknesses, I’ve put together the relevant frame data for Puff’s ledge options and outlined a methodology for creating counter-play.

First, we need to examine Puff’s options.
All frame data assumes perfect execution with ledge invincibility, thus any vulnerability listed is minimum.
Keep in mind that melee is programed with a mechanic that prevents characters from regrabbing ledge before f29 after letting go (with some exceptions), guaranteeing Jigglypuff a minimum of 8f of vulnerability.

• Empty Refresh (with or without sing), not active, vulnerable 30-38
• Fair Refresh, active on 9, vulnerable 30-42
• Bair -> Sing Refresh, active 15, vulnerable 30-50
• Ledgedash, actionable on 21, vulnerable 30+
• Poke w/ Fair L cancel, active 9, vulnerable 30-44
• Poke w/ Bair L cancel, active 34, vulnerable 30-44
• Poke w/ AC Bair, active on 16, vulnerable 30-44
• Poke w/ Pound, active 14-29, vulnerable 30-47
• Poke w/ AC Upair, active 11-14, vulnerable 30-47
• Poke w/ AC Dair, active 7-29, vulnerable 30-47
• Ledgejump, actionable on f40, vulnerable 30-39+
• Hax/Softdash (Let go, DJ, hold forward to grind up against the stage until you pop over, waveland backward, buffer a FF. Multiple frame perfect inputs, not tournament viable), not actionable, not vulnerable, regrabs on f30
• <100 Getup Attack, active 20-24, inv 1-30, vulnerable 31-56
<100 Roll, inv 1-29, vulnerable 30-49
• <100 Empty Getup, inv 1-30, vulnerable 31-33
• >99 Getup Attack, active 43-59, inv 1-39, vulnerable 40-69
>99 Roll, inv 1-53, vulnerable 54-79
• >99 Empty Getup, inv 1-55, vulnerable 56-59
• Let go->DJ->delayed attack, variable frame data
• Nothing, actionable at any point after first 8f, vulnerable on 38+

As is clearly demonstrated by the above, puff has a large number of options from the ledge with similar levels of viability. However, as her opponent, the number is not actually overwhelming because the principles (when/where she can attack and when/where she can be punished) are very similar. Most of these options are slight variations on each other. The options themselves are not especially good or tricky, it’s her potential to change the timing. Let’s say that you decide to put an attack over the ledge area at frame 31. Fantastic, that can punish most everything. However, Puff’s aerial mobility is profound enough that she could be in any of several places and depending on what you’ve committed to this could avoid or change your punish. Additionally, with an empty let-go -> double jump, Puff can sacrifice invincibility in order to alter the timing of her attack. She can burn her 29f, stay out of the way, then should you attack where she would have been on frame 31, she can whiff punish you by delaying the attack until let’s say f40. This is the primary sense to the mixup. It’s not What so much as When.

Thus, there are three main timings of vulnerability to attack
1) Just before Puff grabs ledge (not possible for her to be inv or to have a hitbox out and grab)
2) About 30f after Puff’s commitment from ledge (whiff punish)
3) 38f after Puff grabs ledge (start of vulnerability should she do nothing)

So the question for the opponent to consider is: what are the primary tools that the character has to engage with the above three timings? What are the options that have a high enough reward attached to a low enough risk to make this very simple version of the neutral game unprofitable for Jigglypuff? Are there options that will straight up win? It is also important to consider how much cheating you can do via reactions. That is, if this is like a game of rock paper scissors, can you use an approximate human reaction time of 20f and delay your throw enough to win without guessing? Can you cover multiple options at the same time or stagger them (i.e. if A whiffs then immediately do B and it’ll cover a second option).

I’m not going to do this work for you 'cause it's a time investment, but I will say that finding the solutions is as simple as making a list of proposals, checking to make sure that they can collectively cover everything, then eliminating any needless crossover until you end up with a group of 3-5 or so that will cover everything with the best risk-reward. Generally, any angled projectile invalidates planking on its own. Long range pokes work well for 1 and 3. FH aerials work well for 2. Additionally, you can use shield/CC/WD to space around attack timings and threaten to grab ledge and eject the Puff into an invincible bair of your own. There's a lot of flexibility.

Here is a group of non-character-specific counter-play ideas


• Shoot projectiles timed to beat regrab.
• Wait on side plat, react with fall-through bair.
• Wait on side plat, fake SH fallthroughs until you can react with fall-through bair.
• Run up CC.
• Run up shield.
• Run up shield pivot, WD onto ledge.
• Walk up f/dtilt.
• Backwards FH, wait to DJ or bair or grab ledge on reaction.
• Run off double jump aerial.
• SHFFL ledgecancel -> DJ

Again, none of these or any additional, character specific options will be enough by themselves. However, in every MU a group of 3-5 should be more than sufficient to turn an irritating stalling tactic into a high risk high reward scenario, if not a more winning low risk medium/high reward one. At that point, beating a specific Jigglypuff is just a matter of matching the ratio of your mixup to the ratio of her refresh options and/or practiced execution.


Notes:
• Although it is pretty frame-tight, Sing regrab can sweetspot under spaced fox dtilt (but NOT mid dtilt).
• Punishes don’t always have to be direct. Taking center may be just as valuable as getting a hit. It depends on context/priorities.
• Remember, with the new 20xx replays it is extremely easy to practice scenarios exactly like this one.
• 8 minutes = 686 puff fair ledgegrab refreshes. There is no reason that puff should give you hundreds of opportunities to punish any of three timings without your engaging and profiting from the situation. Rather than a pain in the ass, look at this as an opportunity to gimp the ungimpable character or at least to get good damage after forcing her off.
• Obviously combating marth or sheik or another character’s planking will have different frame data, but the process for developing the best mixup will be the same.
• As the Puff player, the same process should be gone through with an eye not for sustaining the situation but for getting stage control or a nice punish. Puff's planking is a gimmick. It relies on an opponent's bad understanding to work. Puff can do better.

Monday, June 19, 2017

Flow/Peak

Flow/Peak

Flowstate is a name by Mihály Csíkszentmihályi (a chief creativity researcher) for what is popularly called “being in the zone.”

Flowstate is characterized by high levels of focus on the task at hand and low self-consciousness. It’s that state where you get sucked into the task to the degree that you lose track of self and time and experience your highest performance. For this reason, achieving flowstate is the goal of traditional sports psychology.

Contributors to flowstate:

• Structure
• Realistic, achievable goals
• Unambiguous progress-markers
• Task-focus
• High challenge level
• High skill level
• Intrinsic motivation
• Feeling in control

Detractors:
• Anxiety
• Self-focus
• Feeling out of control

Checks:
• Intense focus
• Little to no self-reflection
• Merging of action and awareness
• Don’t care/notice time passing
• Could you hold a conversation? If so you're not even close.

While similar, flowstate is sometimes distinguished from what we’d call “total absorption,” in which your awareness of everything but the task fades out completely. The easiest way to tell the difference is whether or not you feel like you’re coming to your senses when leaving the state. If you were so unaware of your senses that they seem come back to you as if to fill a void, that’s absorption. If you were aware of them but they weren’t as central as your hyper focus, that’s flowstate. At first this distinction seems like hair-splitting, but consider the following: Who is more likely to succeed? 1) A competitor that is focused on the game and his mind is committed to thought/the screen or 2) a competitor that is focused on the game while 100% present in the moment. Different question: Who is more likely to get hit by a car? A car crash isn’t likely, but surprises are all but guaranteed. Being mindfully focussed rather than totally absorbed preserves your capacity to deal with them.

All this being said, flowstate as a concept is a somewhat problematic because, to my understanding,
a) its theory and research is nascent.
b) its chief context (traditional sports psych) is largely badly conceived.
c) cognitive science isn’t developed enough to provide mechanical answers to the pragmatic questions. Flowstate itself is too nuanced.

The primary problem with the flowstate concept is the subtext that it is outside of your control. The belief is that you cannot control the state itself, you can only lay out the conditions as best you can and then trust that it will show up. This is easily illustrated by describing the primary problem with traditional sports psychology in general. In summary: the primary purpose of sports psychology is to provide psychological skills training that will achieve flowstate on game-day. Because flowstate has been thought to be impossible to achieve with anxiety, these psychological skills usually involve a huge amount of self-correction i.e. self-monitoring. As a result of trying to eliminate anxiety, they actually introduce a secondary flow detractor, self-focus. Traditional sports psychology has been demonstrated to have very weak efficacy. That is to say, if conditions are perfect then the athletes go straight to flow. But should conditions be any less than perfect the athletes are just as distracted by their self-corrective focus than they would have been on their anxiety and while they are better able to handle themselves psychologically on the long-term, they show NO performance increase, sometimes even showing a decrease. This matches my personal experience with traditional sports psych skills.

There is an alternate perspective that rejects the notion that flowstate is uncontrollable; peak. Peak and flowstate are effectively interchangeable terms, but the circles that use peak tend to be less micromanagey. They do not subscribe to the methodology in which you control conditions in order to prime for the optimal mental state in order to prime for the optimal behavior in order to prime for optimal performance. Rather, they choose to skip first few steps and focus on optimal behavior or things directly related to it. Chief among these is MAC, a method increasingly used by olympic athletes and champion sports teams. In the MAC model, psychological skills are used not to eliminate anxiety, but to instruct athletes on how to avoid changing their behavior if anxiety is present. This is obviously more consistent than the alternative. In this way, peak is controllable and, should the skills be effectively learned, very consistent. It becomes an artificial or approximate flowstate that may or may not bleed into a deeper flow/superfluidity, but always has a very high baseline.

The takeaway here should be that a high-performance state is not up to chance. It is not reserved for experts or geniuses. Rather, where and how your focus rests can be easily cultivated just like any other skill or habit. The smallest effort to that effect will go a long way!

Monday, June 12, 2017

Legality and Controllers

Legality and Controllers

In the following article I will be laying out a rational examination of the issue of legality concerning Melee controllers. I say rational (not objective) intentionally, for reasons that will be expanded on below.

While I’ll try to be brief and clear, I also want to be comprehensive. If you take issue with a line of reasoning or have an issue or solution that I didn’t think to add to a section, please email me. I always sincerely appreciate corrections/discussion.


Contents
:
Function and Fairness
Dashbacks/PODE
Notches
Arduinos
Box Controllers



Function and Fairness

Before talking about controllers, it’s important to frame the conversation with two concepts; Function and Fairness.

First, what is the function of a controller? As was laid out in Mental Game and Execution, the performative aspect of Melee as an esport concerns translating our ideas into the game itself via execution. Essentially there’s a clean progression from our conception of the game to our strategy to our mental performance to our physical execution to our controller/hardware to the software to the results. If any one area suffers then results suffer. As competitors, our want for hardware is simple; ideally, a controller should seamlessly translate our fingers’ commands to the game exactly as they were input. We want any blame to lie on us.

This demand is a reflection of our need for fairness. Fairness is a competitive hallmark.
If a game isn’t fair then it isn’t competitive.

Fairness requires that
i) the only meaningful variable between competitors is the skill being measured.
ii) other differences are either negligible or irrelevant.
iii) rules are maintained to preserve fairness.*
Basically, “Anything that you can do, I can do too.”

Competitive games measure skills objectively, but which skills they mean to measure is a subjective judgement shaped by rules. Frequently, difference in subjective judgements is at the heart of a rule disagreement. In these cases there is no right or wrong answer. Either position can be defensible, depending on your personal priorities. But subjectivity does not imply irrationality. The two are not identifiable.

Any discussion on the legality of controllers needs to primarily (if not exclusively) address this framework in order to be relevant or productive.

Throughout this post I will make frequent reference to the relevant competitive hallmarks as well as subjective judgements because I think it is the only to make clear, rational, and responsible arguments. I read recently that for humans, a contrasting idea is processed in the brain exactly like a physical threat. We experience an impulsive, emotional reaction/rejection. But it’s definitely possible and necessary to examine ideas (especially our own) rationally for coherencethat is, examining them outside of that emotional reaction even when they’re subjective/value driven. My opinion on several of these topics changed as I examined them against the context, and I hope that you can be as forthcoming while reading/reflecting.


* One blanket statement that follows from the third fairness principle, the rational NEED for rules. Rules that don’t exist can’t be followed. While I am in no way advocating for needless bureaucracy, it is an inherent responsibility for TOs to address issues of unfairness. TOs ABSTAINING FROM MAKING RULINGS IN THE FACE OF A FAIRNESS ISSUE IS A DISSERVICE TO COMPETITIVENESS/COMMUNITY. Those rulings need not and should not be impulsive or immediate. A rational, well-reasoned, well-documented, as-scientific-as-possible explanation is of course the top priority. But refusing to engage with a (meaningful) issue by definition damages fairness and competitiveness. That being said, the word “meaningful” can be a bit fuzzy can’t it? Hence the difficulty and the need for cogency.



Dashbacks/PODE

Dashback is an in-game mechanic that interacts inconsistently with controllers because of how they are manufactured/break in. Having good dashbacks in game is low in skill and high in dependence on luck/controller. This hardware issue is competitively highly undesirable as it introduces variance as a small, additional layer between your understanding and results. If everyone had access to good dashback controllers or some method to consistently create them this would be a non-issue because it’d be an even playing ground. However that’s not the reality. The playing ground is decidedly uneven. Reportedly, only some 1-3% of controllers are top performing dashback controllers, and only after some amount of use. Finding one essentially becomes a controller lottery that hugely favors top players that are popular enough to utilize better resources. Thus, good dashbacks are not a reflection of skill at all, they’re a reflection of luck/resources. They are by definition unfair and hurt competitive validity.

So the relevant question is how relevant dashbacks are. Traditionally, the answer would be “somewhat.” Good dashbacks have historically been a small advantage enjoyed by a handful of players, but not an advantage that is overly influential OR (and this is important) one with a clear solution. However, opinion and circumstance has shifted recently. Dashback controllers have become a hot topic in part due to Armada dropping out of Dream Hack Austin “because his controller was not properly malfunctioning.” While a resource detailing a statistical advantage enjoyed by dashback controller users does not exist (and would be overly difficult to make), Armada for one considers this mechanic to be important enough that he refused to compete without said advantage. Mew2King, Hax, and other top players have expressed similar sentiment, and this is not unreasonable.

The biggest recent change (besides awareness) to the dashback issue is that it can be solved with software. However, individuals have taken issue with the idea that solutions create a new controller disparity, the means involved, and over the integrity of the game.

Disparity: When the smashbox was announced, some individuals claimed that the advantage offered by created a new controller disparity in which players will have to purchase a special controller or mod for good dashbacks. This particular argument against special controllers does not hold water simply because as described above the disparity already exists and is exactly the fairness issue in question. If anything a new market of dashback controllers lessens the disparity.

Means: This is a somewhat relevant concern. Dashback can be solved for by making certain that the stick doesn’t register on its way to smashturn from center, but how this is done could be potentially problematic. Creating software on a per-controller basis with Arduinos is a viable solution, but arduinos are problematic for other reasons (detailed below). Arduinos radically lessen but do not totally solve the disparity problem. The more complete solution would need to be implemented through the game itself via a Magus Code or similar variant. Currently, the resources exist to run a tournament in which each console has code running to universally equalize dashbacks. This cleanly eliminates the disparity for all participants. However, some individuals dislike this solution as it alters the game that we play.

Integrity: Part of the appeal to Melee is that it’s an older game that hasn’t had any patches. Every issue that’s come up has been solved for by the meta or lived with. Software fixes (namely 20XX:TE) exist, but haven’t found widespread use in tournament. Melee = Vanilla Melee and always has. For someone that subjectively values that integrity more than competitiveness, running a code (that would be considered a small software patch to compensate for hardware inconsistency) to equalize dashbacks is undesirable. For another individual that subjectively values competitiveness more than the integrity of the game, that trade is acceptable. Thus, which side of the fence you’re on depends on your priorities, including how big of a deal you think dashbacks are.

Personally, I value the game’s fairness regarding to dashbacks more than I value the game’s integrity. Were I still actively TOing I would implement the Magus code, as the integrity issue is the only argument against the code that I am aware of.



Notches

Similarly to dashbacks, manufacturing inconsistency renders some controllers better than others at shielddropping with consistency due to the angle of the gate relative to the stick box (this is not as well-known, but the upward 45s are similarly inconsistent at upB without double-jumping). While good shielddrop notches ARE a distinct in-game advantage, unlike dashbacks, getting them is not prohibitive. The problem is easily remedied by filing your controller gate to the desired value.
I’ve personally filed give or take a dozen controllers. It takes a file/exacto knife, one of the two most recent versions of 20XX or a Magus ISO, and 10 minutes (alternatively: a friend that has these). For this reason shielddrop notches are not meaningfully unfair because by way of notching their controller, everyone has reasonable access to them. The playing field is effectively even. There is some talk of implementing a software fix and while that change would also be viable it is not especially necessary.

The trickier problem comes from additional notches, such as those for perfect wavedash/firefox angles etc. These are also a distinct competitive advantage via consistency. Additionally, these mods are utterly intentional. Unlike dashbacks or shielddrop notches, no controller has them from the box. But while you can pay someone to do these mods for you, in this case disparity is actually irrelevant because reproducing them is relatively easy for anyone with a file and 20XX. I’ve done 4 myself. They take considerably longer than shielddrop notches because it’s more involved but the process is not really any more difficult. In this case, the question is whether the competitive advantage offered by the boon in consistency performing very valuable and otherwise difficult tactics (perfect wavedashes/firefox angles) is dramatic/relevant enough to merit a ban. The answer depends on the degree to which you value precision as a skill compared to consistency of execution. For some individuals, the left-thumb techskill involved in locating precise angles without the help of additional notches is a highly valued skill. For others it is not.

Personally, I do not have a problem with additional notches specifically because
a) the in-game advantage is not severe. It offers consistency, which I value, not anything that is otherwise impossible. It does not overcome decision-making. I personally come from the perspective that muscle-memory related techskill is more like an arbitrary barrier/time-investment than a valuable skill to test. I don’t devalue techskill, but I wouldn’t prioritize it over creativity or consistency.
b) it is easy to do yourself should you desire it.



Arduinos

Arduinos create software solutions on a per-controller basis. By virtue of being cheap and relatively simple to install, they are not a disparity issue. The interesting problem with Arduinos is that they force two issues: how much we value techskill as execution and cheating.

Arduinos can fix a controller’s dashback, 1.0 value dash, and shielddrop inconsistency in one fell swoop. They level every playing field and are in that way very good for fairness.
However, as mentioned in the section on dashbaks/PODE, dashbacks, 1.0 dashes, and shielddrops by a memory card or like game-bound software patch (as opposed to a controller-bound patch) is cleaner and inarguably better in terms of fairness. The only reasonable argument in support of Arduinos but against a Magus-like code is inherently sentimental in that it rejects the cleaner solution with more universal results in order to preserve the game’s integrity.

Beyond fairness, though, a few Hax Arduino functions (those for easier perfect firefox/WDs as well as for the downward) are described as effective input-map alterations. These are philosophicallythough not pragmaticallyproblematic in comparison to any mods mentioned so far because they go beyond patching for hardware inconsistency and instead aim to address in-game consistency, to the degree that it is a pronounced competitive advantage somewhere beyond notches. It seems to me that Hax is using the input-map alterations to intentionally push boundaries, most likely to justify some of the B0XX’s functionalities.

Then there is the issue of cheating. The biggest problem with an Arduino is that it can do anything. It doesn’t have to only have dashback etc fixes, it could just as easily have perfect multishine/waveshine/angles/SHFFLnair/etc macros programmed to a specific button or combination of buttons. You could easily program instant ICs jump or tilt desync to X, for example. That kind of abuse is why macros have always been banned. While the use of Arduinos does not make macros legal, their use does bring macros as well as some gray areas to the surface by virtue of being difficult to detect/distinguish from more legitimate use.

In my opinion, Arduinos with the proposed changes are not necessarily ban-worthy. They mean to patch for consistency (level the playing field) and buff angle consistency (which, again, is fine unless you value left thumb techskill more than consistency). However, as a preventative measure, it would be better to ban Arduinos categorically and implement the hardware consistency solutions through an alternative means (i.e. Magus code).



Box Controllers

Note: everything written here on this topic pertains only to the Smashbox. I cannot comment on the B0XX because it has different design principles and is rumored to have some drastic divergences. Without more detail it is not possible to review its competitive validity.

Box controllers have been a huge source of controversy almost solely because of Analog -> Digital inputs. Because the Smashbox replaced an analog stick with a group of digital buttons, the team had to design a method to achieve precise angles. That system is interesting, but like the Hax Arduino raises the issue of extreme software consistency because with digital inputs far beyond being less likely (like a notch) it is not actually possible to miss an angle. As a consequence of its design, the problem with a smashbox is identical to that of notches/angle consistency. How much do you value left-thumb techskill compared to consistency? If you think that left-thumb techskill is a very meaningful skill then the smashbox, which doesn’t require said skill in the same way, is undesirable. If you value consistency over left-thumb techskill, then the smashbox’s digital inputs are desirable. In fact, if the precise control offered by a smashbox is severe enough then like a dashback controller it creates a disparity that is unfair in proportion to its prohibitive cost (but to be fair relative to other games $200 is not especially prohibitive).

In my estimation, advantages offered by a smashbox are not severe enough to merit banning. Once more, the design is to offer consistency, not anything that is otherwise impossible. The other advantages (dashback consistency/1.0 dashes/shielddrops/etc) are obviously not unique to the smashbox, so it would be disingenuous to ban it for those reasons. The only defensible reasons to ban the smashbox are 1) if the community values left-thumb techskill over consistency (as well as in this case accessibility) or 2) if the smashbox offers some other competitive advantage that breaks the game.

Because smashboxes are not yet readily available, tests for game-breaking smashbox tech are not possible, but the most frequently referred to is broken burst MSDI. On that note: I cannot cite it because as far as I know it was not published, but I read of a comparative test between keyboard MSDI and practiced Wank DI that got effectively identical results. Wank DI is actually preferable, as all of its inputs are approximately in the same direction. It will be interesting to see what tests surface after release.

To prevent abuse from smashbox-based variants, I was briefly consulted by Smashbox Dustin re: a ruleset. Using one discussed by a group of content creators/frame data nerds as a basis, I gave him the following:
• No Macros.
• 1 to 1 mapping only (This means that you can only have 1 A button, 1 B button, 2 jump buttons, etc).
• Analog may be converted to Digital BUT it must include only ONE North, South, East, West set attached to ONE stick OR wasd. NO additional 45 degree buttons. Modifiers are acceptable.
• No Digital to Analog Conversion.


In my opinion the Smashbox is a very interesting and potentially hugely beneficial development. I personally value its offer of consistency and accessibility to currently alienated players (including FGC members as well as those with hand issues etc) more than I value left-thumb techskill. Should any tech appear that is game-breaking then a ban can be discussed but a preemptive ban would be lazy and sensationalist rather than well-reasoned.

Tuesday, June 6, 2017

Ruleset Development as Related to Melee

Ruleset Development as Related to Melee

Contents:
Rules and Hallmarks of Competition
Bans in the FGC
Culture and Subjectivity
Rulesets Cannot be Perfect
Melee Community
TLDR



Rules and Hallmarks of Competition

Games are played with certain inherent limitations. For a traditional sport, these mostly include physical limitations such as laws of physics, human endurance, and the like. For an esport, these include the limitations of the game such as programed physics, character limitations, and the like. Sometimes these limits can be bent to our competitive advantage. Other times they need to be bolstered in order to preserve fairness or competitive validity. This is the purpose of a ruleset.

Rules are artificial, additional limitations that we put on a game so that it will be a better measure of competitive skill. Skill can be understood as expertise or ability. It can be very clear-cut (running fast in a footrace) or more inclusive (using teamwork, strategy, tactics, and other practiced physical and mental sub-skills to beat another team in basketball). The word Competitive, though, is a bit loaded. To better understand it we need to consider the following hallmarks of competition, which are well articulated here.

To paraphrase, in competitive games

1) Skill is measured fairly.
All variables are constant except the skill being tested are constant. You are allowed to do everything that your opponent can. Any difference in conditions other than skill is either negligible or considered inconsequential in comparison. In a footrace, every competitor runs the same distance in the same conditions. Although they may wear different shoes, shoes are not generally consequential in comparison with the skill of running fast.

2) There are universal rules designed to preserve fairness.
These rules will emphasize rather than change what the game measures. This may be a fine line. Swimsuits are all fair game until their sole influence adequately determines results, at which point they are bannable. In a less ambiguous example, doping is banned as it effectively eliminates the skill being measured and substitutes it with “access to steroids.”

3) Results are unambiguous.
The more skillful player wins. The less skillful player loses. Any draws contribute to rankings by way of a ladder etc or are solved by sudden death etc.

Some games are naturally very competitive. Others are not. There are two main indicators of natural un-competitiveness. First, if a game is overly simple and the skill-cap too low, its competitiveness suffers. TicTacToe is so easy to play perfectly that it is not competitive. Second, a game can be too high variance. A game where luck determines results more than skill is not competitive. Uno is not a competitive game. That being said, there are some games (including Poker) in which variance does not outweigh skill so much as accent it. Generally speaking, a game that is high in luck and low in skill (Rock Paper Scissors) can be competed in but is at least not a good competitive game. By default settings, Melee is such a game. However, over its competitive lifespan, the Melee community has used an evolving set of rulesets/bans to all but eliminate variance and emphasize skill. It is now a very competitive game.


Bans in the FGC

Traditionally, FGC rulesets (best of 3, 90 seconds or what have you) are simple and in the game itself. Those kinds of rules are in place already and rarely need to be altered. The purpose of a rule-change or ban is to make the game more competitive than it is natively. To avoid abuse, rule-changes are treated seriously and their reasoning is well-documented. On occasion a tactic or character appears that changes the competitive landscape enough to warrant banning. The guidelines in place for bans are spelled out in Playing To Win.
PtW says that any ban must be
• Enforceable (detectable)
• Discrete (completely defined/unambiguous)
• and Warranted (dominates the game)

Because these criteria (particularly Warranted) are so strict, the FGC rarely bans in-game techniques or characters that don’t make the game unplayable. Rather, they accept that strong tactics are strong tactics. As long as it is possible to beat a character without the exclusion of other tactics, that technique, while strong, does not break the game. If it makes the game unfun or uninteresting then in truth the game was always unfun or uninteresting; you just didn’t know before. But more often the rampant use of a strong tactic drives the development of counter-tactics and counter-counter tactics, actually radically deepening the level of play. For this reason preemptive bans are unacceptable. They will not ban anything reflexively before it is demonstrated to be meaningful, dominant problem that breaks a game’s competitiveness. This approach is very fitting with the FGCs harsh, Get Good, adaption-centric culture. It is also quite reasonable, considering that any tactics that are not clearly bannable are likely to disappear with time, be it through meta-evolution, update patches, or a new title in the series.

Interestingly, in the same chapter PtW describes a concept it calls a “Soft Ban.” A soft ban is a sort of social stigma against using “unfair” characters or techniques. In the book, it describes top players in Japan refusing to use Akuma or Old Sagat, despite their being borderline bannable. Because they weren’t used except by uppity players that were quickly humiliated by veterans, these characters never dominated results and thus a ban wasn’t warranted. Relying on a social taboo wasn’t necessarily perfect, but in this case it was sufficient to keep the game worth playing and preserving the potential for meta-development. It is interesting to see this anecdote in a book called Playing to Win, a book that places extreme emphasis on removing moralistic judgement from your approach to games. It reminds us that these games are not played in a vacuum. They exist within a culture of players with complex value systems. In the smaller, insular communities that PtW was written for, soft bans are not just acceptable but preferable to hard bans. They are, for the time being, good enough. When they aren’t good enough then presumably a hard ban comes into effect OR the game is patched/updated at which point the problem vanishes.

Note: I do not think that the FGC criteria as defined by PtW is the best system, but it is definitely foundational.


Culture and Subjectivity

Ultimately, competitive games are an objective measure of skill. However, what skills they are designed to measure is a subjective judgement. This is easy to demonstrate with SSBM. As a community, we have decided to play with 4 stocks, 8 minutes, and no items. This is a radically different set of in-game rules than the default 2 minute time match with random items. Some casuals complain that competitive melee is not a measure of “real” skill because it does not account for other aspects of the game such as the situational awareness that comes from coping with items or some stages. That opinion seems scrubby to us, but it is not untrue. The current competitive ruleset does measure specific skills at the necessary exclusion of others. It measures a smaller set of skills that we collectively decided to be more desirable. In particular, we reject most variance. Most of our ruleset is specifically designed to root out randomness in favor of consistency. But in an alternate universe, a competitive melee scene based on an alternate set of values could exist. In universe B where competitive melee is played out in coin battles there would be an entirely different culture that develops entirely different rulesets that give rise to entirely different strategies, tier lists, and rankings that are all still highly-competitive. That would be fine. The difference between universe A and B is what skills we prefer to measure, what we value seeing and experiencing, what we think is good. While not quite arbitrary, this is obviously a subjective judgement.

It is also important to recognize that Melee is unique game with specific and individualized possibilities. Those particular possibilities excite and inspire the community that is drawn to it. If you ask competitors what they love about melee you will receive a number of different responses, but all of these will be particular to the game and are the basis of the culture that surrounds it. Melee, like any expressive medium, becomes a meeting point for people, their values, their experience, and the particular idiosyncrasies and limitations of each. The best possible ruleset is the best possible catalyst for this meeting as competition. It is not just possible but probable that its subjectively valued skillset will have a different identity than those of other games and subcultures. Which is all to say that melee is a unique game that attracts a unique culture and that its ruleset requires unique consideration. Furthermore, the creation of its ruleset is synonymous with the identification of values.



Rulesets Cannot Be Perfect

Very competitive games demand constant evolution. In order to get an edge, players constantly push the boundaries of what is possible. Consequently, there is a huge amount of uncertainty as to what techniques or strategies will remain at the top and rulesets can no more predict this than the people that make them. Rulesets cannot be perfect. The relationship between the game's rules and the community is at least as fluid as the meta. Sometimes bans (soft or hard) are appropriate, sometimes they aren’t. Because the skills that we choose to measure are subjective, any elaborations that we make upon the game’s limitations (i.e. rules/bans) are also subjective. In the FGC there’s an implicit understanding that rulesets are not perfect so much as serviceable. They are free to evolve and likely to fall into disuse as new patches or new games come out. Because Melee has been around for so long, we tend to forget this.



Melee Community

Because Melee is not a traditional fighter, some of the guidelines to traditional fighter rulesets do not apply so readily. In particular, liberal stage bans in the smash community have undoubtedly been positive, as have the other format changes that FGC criteria would have a hard time justifying. It is important to recognize that while a traditional fighter can rely on the meta, patches or new games to clean up its messes before any rule changes, Melee can only rely on meta and rule changes. Luckily the game has been able to sustain an astonishing amount of counter-play development, but some rule changes have been beneficial. The strict PtW guidelines are are fantastic, but certainly not unassailable gospel. Rather, it would be to our benefit to better identify in hot topics the exact points of disagreement as they relate to competitiveness and skill assessment while recalling the usefulness of soft bans.

Many disagreements over rulesets are disagreements over what skills or priorities you value. If the skill that we mean to measure is subjectively determined (but remember, objectively assessed) then that means that the ideal boundaries (rules) of the game are potentially different per individual. There are those subjective skills as represented in the current ruleset, then there are competitors that just want to win, competitors that want the game to measure different skills, spectators that want the game to measure the same or different skills, and spectators that just want their guy to win. All of these individuals have different priorities, articulate or not, rational or not. For example: how much emphasis should be placed on which sub-skill? How much of melee skill is physical execution vs tactics (controller controversies)? How much is conventional vs unconventional win conditions (timeout controversies)? Etc. But as should be clear by the groundwork laid out in this article, all disagreements can be better argued if not outright resolved if they can be better articulated in relation to the above.


TLDR:
• Rules are artificial limitations that we put on a game so that it better measures what skills we subjectively value while preserving fairness. (Swimming suits are legal until the impact of suits overtakes the swimming skill, but racing cars is different because developing driving and developing cars are both subjectively valued. Different sports use different criteria.)
• There are hard bans (rules) and soft bans (social taboos). Both are effective, but hard bans (PARTICULARLY PREEMPTIVE BANS) are undesirable because it likely impedes unexpected meta-development.
• By applying rules to a game you make it an objective measure of a subjectively determined skillset. That is ok and actually really cool when done well because it creates a meeting point for people and their individual values/experiences.
• Because the players push boundaries there are always unknowns that the community can't be ready to make a judgement about. Thus no ruleset is perfect. It is an approximate and flexible set of limitations.
• Disagreements over rules are disagreements over values. They are best resolved when those values are clearly articulated in a specific-as-possible context.