Small experiment about the perception of inertia.

Introduction.

The idea for this experiment was suggested in the middle of a friendly discussion with SkyArt and Blossom in Turtle's server.

SkyArt's stance was that inertia is irrelevant for casual players who are unfamiliar with high level gameplay in DMC4/5, because they could not even identify it in high level clips or understand how it works. And probably that's why it was dismissed in DMC5.

The summary of my own stance was that it's most probably (perceptively) relevant even when people don't know DMC4/5 and don't understand said mechanic. Because it still provides a more complex structure that affects the visual stimuli (like a combo or gameplay clip).

In other words, one shouldn't need to understand how inertia in DMC4 works to perceive the difference between Dante remaining relatively static and Dante moving in relation to an enemy in a way that makes sense in terms of structure. And even to enjoy said structure.

The main question for the experiment would be:

Do people generally need a trained perception (thanks to knowledge, practice and/or experience watching many high level clips) to appreciate the benefits of this mechanic (even if they don't really know how it works)?

In music, for example, there are variables that don't need a special training to be appreciated. You don't need to be a musician to perceive something is off if a drummer's beats aren't well timed. Even when there are many instruments playing at the same time and you can't really say what's exactly off.

However, for other variables you may need a more trainer perception. Like enjoying specific styles.

The same can be applied to other forms of art. There are already studies about perception in the context of, for example, dance. But it would be hard to extrapolate conclusions because of how specific is our question and the independent variable (strictly related to motion from inertial techniques).

Method and Results.

21 subjects unfamiliar with DMC4's high level gameplay, or even unfamiliar with DMC in general, were asked which clip they find more pleasant to watch or merely prefer (assuming they don't like any of them):

Clip A.

Clip B.

Clip A barely has inertia (some from gravity and Full House transmitting minor inertia to ecstasy, but no typical inertial tech), but Clip B integrates: inertial rainstorm, inertial rave backwards and flying guard. 3 well known inertial techniques that provide structure to a combo thanks to different types of movement.

Both clips also integrate normal movement with sky star at the beginning, because it's something possible without inertial techs.

The order of presentation of both clips was randomized, even though subjects could re-watch several times if they wanted to. Subjects were also generally isolated to avoid the effect of majority influence or other confounding variables.

It was clarified that mod effects, like kanjis, and the background could be ignored. And simply focus on both characters (without using words like movement, transition, structure, etc).

Null Hypothesis would be:

Inertial mechanics aren't relevant enough to cause a significant visual impact, thus the tendency would be that subjects (collectively) wouldn't prefer any clip in particular, or definitely not Clip B.

Theoretical probability would be 0,5 for Clip B (or less).

B=<A

Alternative hypothesis:

Subjects will tend to prefer Clip B.

B>A.

Results analyzed using a left tailed binomial distribution test to evaluate statistical significance:

X = number of subjects who preferred Clip B -1. Because we are interested in the probability of obtaining 15 or more, while this test evaluates the left part of the distribution (the probability of obtaining less than 15 in our case).

ensayos = number of subjects (for us).

SP = theoretical probability from the null hypothesis.

C is related to accumulative probability or exact probability (1 in our case because it's accumulative).

Since we are interested in the right tail, becase we want to know the probability of obtaining 15 or more times the choice of Clip B with a theoretical probability of 0,5:

1-0,96=0,04

Being the p-value below the standard of 0,05. Which means there would be statistical significance (even though the sample is way too small to take this too seriously).

If you are curious about this type of analysis, there are calculators online as well. Like this one:

What all this means is simply that the probability of 15 subjects out of 21 choosing Clip B is roughly 0,04 (very unlikely) with a default probability of 0,5 per Clip. Thus inertia (probably) is not irrelevant, or there are other factors that we haven't realized (called confounding variables).

Discussion and criticism.

The most obvious flaws of this small experiment are the limited sample and the application of only one pair of clips.

Limitations come partly from the fact that we couldn't apply the experiment making a general tweet and asking the question with both clips, because many people in my social media have watched already clips of mine and could have a trained perception. This means we had to know if the subject were familiar or not with DMC/high level gameplay before accepting them. Subjects also had to be isolated to avoid confounders.

The reason why we decided to use only 2 clips was to keep it simple/easy to apply. For example, one of the subjects was a coworker, and obviously I couldn't really sit with him too long watching and re-watching clips. Another reason was to avoid an attention-span confounder if people had minimal interest in DMC/the experiment.

One of the subjects was very unfamiliar with video games and replied that he preferred clip A because it was easier to understand, while clip B looked more abstract because of the fast movement. This is interesting, because it could illustrate a relevant variable where inertia has a negative effect under certain circumstances. This could be analogous to musical complexity having a negative effect for some people unfamiliar with certain musical styles, like jazz or classical music. It also suggests that it could be interesting to analyze groups in terms of factors like age and/or being unfamiliar with video games overall.

Another problem of the experiment is the ambiguity of being unfamiliar with high level gameplay/DMC. It's hard to draw a line and we couldn't dismiss any subject who had already watched any DMC related clip, because the sample would have been even smaller.

In any case..., this small experiment isn't obviously anything to be taken too seriously because of obvious limitations, but it would be interesting to see a more rigorous ones to support the implementation of complex mechanics (in action games and other genres) that provide more options for movement.

Thanks to Blossom, Quang, ADevelishMotive and Lidemi for applying the experiment to some subjects.