polarbair

you hear about Fightcade

BLM | Really into Pokemon | FGC TO and wiki editor | Small-time streamer



nodthenbow
@nodthenbow

Long essay below the fold


First, because it would be irresponsible to not talk about accessibility when talking about difficulty, lets talk about accessibility.

Accessibility

Accessibility is a topic that is too often brought up as the antithesis of difficulty, but such a sweeping statement over-simplifies the issues. This will be focused on the competitive pvp aspects of fighting games, but first lets lay some ground work. In single player games accessibility simply allows more people to experience the game in the way the target audience is supposed to experience it. Two common counterarguments to this are (1) some people will play the game “wrong” and worsen their experience, and (2) an appeal to the designer’s vision for intended difficulty. Counterargument 1 is not an argument against accessibility itself, it is only pointing out that poor communication to the player from the game may result in a worse experience. Therefor if the game communicates its features properly, the accessibility features would not worsen the experience for players. The second counter argument is based on a misconception of design intention. I’ll explain this using an example. In Samurai Kirby, a reaction test minigame in a few Kirby games, there is a difficulty curve that is designed to go from requiring a trivial reaction, to nearing the limits of how fast a human can react, to requiring an outright impossible reaction for the average person. On a typical setup (CRT + SNES) the fastest a human could react would result in about a 12 frame reaction. The opponent’s reaction times are listed below:
At 60 frames per second, in stage order:

  • Waddle Doo; reacts in 82 frames.
  • Wheelie; reacts in 50 frames.
  • Chef Kawasaki; reacts in 21 frames.
  • King Dedede; reacts in 16 frames.
  • Meta Knight; reacts in 11 frames.

We will assume that this difficulty curve was intended. Given that assumption it should be evident that human reaction time played a part in determining the amount of frames the opponents react in. Given that, we can say that someone playing this game with a reaction time that peaks at 10 frames, or 17 frames, would not experience the intended difficulty curve. This is because they would be able to react to the last fight, or because they would not be able to react to the last two fights. This means that if you accept that a designer's intention for the difficulty is based on some person's ability (real or not), then if the player does not match that ability they are unable to play the game with the intended difficulty unless there are accessibility features that correct for the difference in ability.

Fighting games ideally should be fully accessible, but due to the nature if it being a competitive game there are restrictions on how some accessibility features can be implemented. In some cases this would require more complicated implementations of accessibility features, but in some other cases it is going to be the case that there will be designs and features that are going to be inherently inaccessible to some. This is not to say that all accessibility features will have issues in fighting games. A trivial example of a feature without notable issues would be a colourblind filter. It is much harder to show accessibility features that are impossible to implement due to the design of the game, which is in part due to how flexible such features can be. One of those impossible pairings is designing integral parts of a system around reactions and having a feature that allows for people with impaired reactions to have equal footing. The reason this happens is due to the way reaction times work. People’s reaction time scales with the amount of things they are trying to react to at the same time. This is called Hick’s Law, so look that up if you want to learn more about it. The incompatibility is that any system that increases the minimum reaction time required for an impaired person can be used by an unimpaired person to allow them to react to more things at the same time, which is a competitive advantage. The only solution is to change the design or not include that accessibility feature. These impossible situations are uncommon though, but the mistaking difficult to implement with impossible is an easy mistake to make.

One such feature that is often seen as requiring design restrictions is easy input special moves. Part of this mistake is because the most common implementation, allowing specials to be done with a single button press, does have design restrictions. A simple example is having a fast reversal on one button means that people can use it as fast as they can switch to a block. This has far reaching implications, but one of the notable ones is that due to how the easy input takes less time to input it will allow people to react with it faster, and due to Hick’s Law this allows people to react to more things at once while threatening the reversal. This is a common case in games similar to Street Fighter as jump ins are usually just fast enough to let an average person block when they are not intending to react to a jump in, and fast enough to reversal when they are intending to react to a jump in. This issue is not because the input for a special move was easier to do though, and that is why this is not an impossible feature to implement. In order to implement easy inputs for special moves you need to emulate the downsides of inputting each individual special move. For example a fix to the inputting speed increase could be solved by adding a delay after the easy input before the move comes out. That’s my spiel on accessibility in fighting games. It’s a good thing that should be implemented. On to difficulty.

Difficulty

Difficulty is a blanket term for how hard something would be for a some audience. It’s a fairly general term, which causes issues when one tries to make comparisons. For our purposes difficulty will be defined as the elements of playing a typical competitive fighting game character, such as the ones in Tekken, Street Fighter, and Guilty Gear games, that are required to be done at a level that is not trivial to the average able bodied person. The scale will generalize things that take a relatively small amount of time to develop the skills required to meet the requirements for competition a competitive environment as easy, and things that take a relatively long amount of time will be hard. As an example, jumping in a fighting game would be characterized as easy because developing the skills required to jump on par with a competitive player is simple, as the skill is a mechanical test of performing a (usually trivial) jump input. Learning when to use jumps is a separate skill which could be easy or hard depending on the context of the game that skill applies to. Difficulty is also made up of skill tests. The categories of skill tests that will be discussed are defined as:
(1) Execution. Execution is the group of skills under the umbrella of dexterity, hand eye coordination, and timing.
(2) Memorization. Memorization is the ability to remember and recall required information quickly and/or at a large volume.
(3) Awareness. Awareness skills are ones that are used to keep track of dynamic situations and conditions. Resource management is an example of an awareness skill.
(4) Reactions. Reactions are skills directly related to reacting to things in itself. Choosing what to react to is not a reactions skill, but context switching (the ability to quickly prepare to react to new things) is a reactions skill. Although there is a hard cap on how fast people can react, people do not react at their fastest without practice.
(5) Specifics. Specifics are the umbrella category that covers skills which are in parts of skills in other categories, but within the context of specific fighting game situations and tropes are better served as being seen as grouped skills due to the fact that in the context of their skill tests they require each other. An example of such a skill would be adaptation, because adaptation requires many other skills and specific knowledge to perform, while also being integral to competition.

Capability can be framed as a function of time spent learning the skill. This is true for most of the skills in fighting games too. However due to the fact that competitive fighting games are complex enough to not have known limits, to efficiently become capable at a skill, estimators must be used in order to determine how a skill should be learned and applied. For skills that require these estimators using the metric of time spent learning as a the objective measure for capability of that skill hides the fact that the time spent learning is dependent on the method used to learn. The method used to learn may not be efficient, so using time as a measure will fail to correlate with other empirical measures of capability. As an example, consider a situation where in learning to react to a four way reactable mixup one assumes that the best way to learn the skill is to practice reacting to all four mixup options individually. This assumption, that the mixup can be reacted to by learning four separate reactions, is the estimator of the skill, but it is not the most efficient way of learning. One part that can be learned is that reacting to multiple things at once has different areas of reaction times to be practiced, so learning a not fully transferable skill for the situation it will take a longer time to acquire the skill needed because a better estimator is needed before the skill can actually be learned. Either it increases the time spent learning if you decide to start timing at the start of learning or it is decreased if you start timing when the better estimator is found because part of the skill was already practiced and learned. For this reason in order to accurately judge difficulty it would seem useful to separate skills that use estimators that are likely to change from skills that do not.

Execution skills often can be measured by objective skill tests, and these objective skill tests are often entirely accurate to the skill required by the game. The assumed way an execution skill is tested by a game is often effectively equivalent to the true way that skill is tested by the game. An example would be one frame links in Street Fighter III: Third Strike, where the assumed way to execute a one frame link is to press the button at the right time, and for nearly all cases that is correct. One example where the assumed test is different from the actual test is in one frame links in Street Fighter IV. SF4 has a mechanic called plinking (and slinking) which allows a one frame window to be extended to a two frame window. However, once this mechanic is know and the assumed test is updated, then the assumed test is fairly accurate to the actual test in most cases. As a fun tangent, in some old games there is a feature called turbo or frame skip, which will skip showing a frame of the game. This feature is sometimes programmed in such a way that it skips reading inputs for that frame too, which can turn a one frame link into an impossible combo at times. This can result in the assumed skill test allowing for more consistency than the actual skill test will allow, which will make it seem much harder until you learn what is going on. Back on track, because execution is so well measured by time and because many of the skills are generally transferable (timing, hand eye coordination, ect) it can be seen as one of the easiest or one of the hardest skills to learn (depending on where you start timing the other skills). In cases like these where time is the major factor in developing the skill, as opposed to information, I tend to evaluate its worth by asking what justifies its difficulty. First part of that is to determine to what degree that skill test is a necessity so that the designs that make it necessary can be judged. In the case of one frame timing windows an input buffer can remove most of them, so other than some outliers they are not very necessary. Knowing this we move on to the second step of determining what it adds and takes away from the game by existing. In the same case this is usually added rewards for players who have developed timing skills from somewhere, some people find learning hard combos enjoyable while others find it unenjoyable, and it can cause accessibility issues. Because timing as a skill is essentially a comparison of time spent practicing and not an interactive competition, plus the accessibility issues, I find the skill check in this case to generally be against the type of competition a fighting game is often played for, in the same way that multiplying damage/health by hours played would be. In addition the obviousness of execution tests means their contribution to difficulty is often over represented.

Memorization is nearly the opposite of execution in how it easy it is to find a good estimator. Memorization is difficult not because it is hard to remember a lot of things, the difficulty is mainly being able to remember what you know when you need to know it. The skill of long term memorization is easily overcome by taking notes, so it is trivial for most cases that would be relevant to fighting games. The way that memorization is tested in most games is not a lump test of all knowledge, specific knowledge becomes relevant at different points and learning what situations lead into what is the main skill test in memorization. There are two approaches to learning situations, and both can be useful. The first way is to focus on the minimum set of required things and over time add in more complex knowledge about common or important information, and the other way is to learn as much as possible and over time learn the situations where the knowledge needs to be recalled quickly. The easy to hard scale is both depending on which approach a person uses as well as the sheer number of things that need to be learned, larger is harder. This skill’s difficulty can swing largely between characters and matchups even in the same game. The time it takes to learn the required memorization skills can sometimes carry between games, for example blocking something is usually a cue to bring your defense knowledge into working memory, but the specifics of the knowledge itself is largely context specific so new associations will have to be made. One other skill that knowledge is checked is also linked to reaction skills. By learning to link knowledge directly to stimulus. For example, instead of linking the opponent jumping to anti airing you would link it directly to doing your character’s anti air input in order to not spend extra time going from knowing what to do to doing it. This specific skill will always be context specific, and if you play a lot of games or characters sometimes it can cause mental mixups to happen. Memorization itself, as in ingraining the knowledge enough to be able to recall things without notes, is also a function of time, but it does not require dedicated time to practice because it will be learned while learning what knowledge to apply where (assuming you know enough about the situation to figure out what should be applied where). Memorization is virtually impossible to remove from fighting games, but there is a range of simplicity and complexity that can be done. Some people love complex things and some people love simple things, and fortunately most games can and want to have characters that cater to the whole range.

Awareness has a lot of crossover with memorization, but it differs in that memorization is the what and awareness is the where and when. This also means that in most awareness tests there will also be memorization tests, but there are important distinctions between them. One important distinction is that awareness tests are often dictating changes in strategy. Where memorization tests how well you can recognizes cues, awareness tests how well you can understand the conditions of the current moment. How well you can transition between situations is one of the main skills that is tested under the awareness categories, and another one is how well you can adapt to new unexpected scenarios (scrambles). One example of an awareness test in the first way is meter management. Checking your meter takes time, but it informs what you should be doing. Passing the awareness test by being able to accurately estimate your meter is rewarded by saving the time that would have been spent checking. Another example of an awareness test is the time spent in scramble situations. If you know and are prepared to transition into all the possible situations then you will never be in a scramble situation. Awareness tests are not necessary, but they are one of the defining sources of fun for fighting games. The range of difficulty that people find fun is down to personal taste, but much like memorization it is possible to cater to a range of people by making diverse character designs. One issue is that awareness is tested during player interactions, so there runs a risk that a character’s awareness difficulty might override the intended difficulty of the other player’s character. In my opinion it is better to err towards harder for this reason, because erring low can remove one of the main sources of fun.

Reactions are the simplest form of difficulty. The closer the reaction test requires you to be to the human limit of reaction speeds the harder it is. As stated before, the time learning reaction skills can vary based on how well the estimators are chosen. It’s likely impossible to remove reaction tests entirely from a competitive fighting game, and making them all trivial is usually not possible without some esoteric designs. There is also a limit to how much leeway one has within the difficulty of reaction tests. This is due to Hick’s Law and the reasoning is the same as why making accessibility tools for reactions is not possible (given enough leeway an easy reaction to n things turns can be abused for a competitive advantage by turning it into a difficult reaction to n+1 things). There isn’t a lot to be said about the worth of reaction tests because the difficulty range is very limited and they are an integral part of the genre, but erring on the easier side is probably the safe bet for purely practical reasons as most technical issues and online play conditions will tend to make reactions harder. Having some buffer room before things intended to be reactable become unreactable will often be worth it in the long run.

Specifics are the types of skills that are obvious to players as a group, and are often treated and tested as a group. Things like anti airing fit this category as it’s an obvious test, even if the skills that are being tested are not obvious. For example anti airing usually is obviously a reaction and execution test, but awareness is often also tested because a common design is to set the speed of jumps and anti air moves such that anti airs are only possible if you are prepared to react with one. Breaking down these skills into more specific tests is important in learning, but when designing such tests it is best to think of them as a whole, tweaking the individual skill requirements to serve the overall difficulty. Specifics are essentially a framing tool for other tests.

Conclusion

Difficulty is complex, people are different, and what some people find hard isn’t going to be hard to others. Judging the difficulty of fighting games accurately requires a much greater understanding of the games than what is apparent without research. To someone not in the know execution will seem like the hardest test of skill in a game, regardless of if it is or not. Although spending time practicing is a factor, learning and recognizing what the best skills to learn is often the fastest way to meet the difficulty requirements of the game and become competitive. Difficulty is not the lump sum of time spent learning either, how fast something is learned is after all dependent on how it is taught. Difficulty and the worth of difficulty is not a simple judgment, so it is important to have an understanding of it when discussing the worth of the many things it impacts.


You must log in to comment.