User:Alex the weeb/Project C: the clone assessment project

Introduction
So, what is Project C? This is an idea which was stolen from inspired by Toomai's Cloneosity project. Cloneosity is an attempt at determining how similar various "clone" characters are to each other, using boring stuff like numbers and statistics. In essence, this project has the same goal, however it will be focused exclusively on Ultimate, and features a completely overhauled methodology.

So, what's the difference?
While Cloneosity takes a rather simplistic approach at gauging how similar fighters are, I intend to go much more in depth. Essentially, Cloneosity only considers whether the animations of the moves look similar, as this is an indication of whether or not a move was built from scratch or not, but I believe this to be flawed as an overall method. While it is useful to an extent, it inevitably leads to suspicious conclusions, such as Ganondorf and Captain Falcon being more similar than Lucas and Ness, despite Ganondorf having a completely different character archetype, playstyle, design and overall feel than Falcon, while Lucas shares all of these traits very closely with Ness. Perhaps there are some facets that are shared more closely between Ganondorf and Falcon, than between Lucas and Ness, but I think that categorizing Ganondorf as a semi-clone of Falcon, while Lucas gets away with being labled a pseudo-clone, is just one example of this method being flawed.

Simply put, only considering move animations, and not the moves' technical properties and functionality, nor the fighters' attributes (which play a major role in determining their character archetype), will only get you so far. It won't tell you how much a clone's playstyle, and general feel to the player, differ from their parents. In a game like Ultimate, where there is a large spectrum of degrees of clonehood, I believe a much deeper analysis of the fighters is required.

Effectively, I hope for this project to be more accurate, more precise, and more gameplay oriented than Cloneosity. Of course, like any method of assessing clones, it will also be flawed, but I hope it will be less flawed than its parent project. I present to you, Cloneosity's very own semi-clone: Project C!

Moves
Every move will be assessed and placed in to one of the following categories, each with their own assigned point values, which are as follows:


 * Identical moves (1 point) - Moves do not vary in any of the key categories. For all intents and purposes, the moves are functionally identical.
 * Copied moves (0.75 points) - Moves are mostly identical, but vary in one of the key categories. The functionality of the moves are therefore different.
 * Common moves (0.5 points) - Moves are similar to each other, and have a clear common descent, but vary in multiple key categories, and thus aren't cloned in the conventional sense.
 * Inspired moves (0.25 points) - This is the trickiest one. The moves have significant differences from each other, but it is reasonable to conclude that some inspiration was taken for the move. Previous common descent is a good way of establishing this.
 * Unique moves (0 points) - The moves are far too different from each other to be considered cloned to any degree.

Effectively, 0.25 points are scored for each key category that doesn't differ, and 0.25 points are scored for clear evidence of inspiration.

If you want a more human visualization, you can think of identical moves as being effectively flawlessly copied by the clone, copied moves being copied by the clone, not not perfectly, common moves being the clone putting their own spin on the move, and inspired moves being the clone taking inspiration from their parent's move, but ultimately doing their own thing with it.

Certain minor moves will not be assessed. These include things like pummels, floor attacks, edge attacks etc., as most of these moves are cloned across a large chunk of the cast anyway, so will just artificially inflate scores. Final smashes will also not be assessed, as these have no competitive relevance. Grabs will be lumped together, as typically they either all vary, or none of them do, and move counts will be standardized to 3 for jabs, and 1 for everything else, so as to prevent artificial inflation or deflation of scores. How this is done will become clear in the scoring section.

Stats
Important stats will also be considered when determining how similar fighters are. This is important, because a fighter's stats have a significant impact on their playstyle. Naturally, fighters with a large number of similar moves will still be forced to play differently if they have significant stat differences. Considered stats will be divided into 2 categories: major and minor. Additionally, if stats vary, but by less than 10%, the points scored will simply be halved, rather than negated entirely.

Major
Major stats are worth 1 point each
 * Walk speed
 * Run speed
 * Air speed
 * Air acceleration
 * Fall speed
 * Gravity
 * Weight
 * Jump height (combined)

Minor
Minor stats are worth 0.5 points each * Only one of these must vary to avoid scoring a point.
 * Initial dash speed and run accel *
 * Shield size and shield stance *
 * Traction
 * Fastfall speed

Any stats which are not mentioned above (such as walk accel and air friction) are too minor to be worth assessing here.

The key categories
These are used to assess what category to place each move in.


 * Hitbox properties - The values found in attack scripts within hitbox flags. These include things like hitbox size, damage, knockback, launch angle etc. In some cases, the severity of the differences may be assessed, for example if the two moves' launch angles vary by just a couple of degrees, this may not be enough to be considered a difference in this key category. In such cases, a justification will be provided.
 * Frame data - The timing of the attack, including startup, active frames, ending lag and FAF, and in the case of aerials, landing lag and autocancel windows. As before, very minor differences may not qualify as a difference in this key category.
 * Animation - The attack's animation. This will not include particle effect changes, as these have no practical effect. Mirrored stances, or virtually imperceptible animation changes may not qualify as a difference in this key category.

In cases where the differences are too minor to qualify in an individual key category, but there are differences in multiple key categories, these differences will be considered to have half their usual weighting, putting the move into a category other than identical. For example, if there are minor differences in 2 key categories, it will be treated as a major difference in 1 key category, and thus the move will be considered copied, rather than identical.

Extra credit
Extra credits (or perhaps more accurately, extra penalties) may be applied for differences which do not fit into the above sections, but are significant enough to warrant a reduction in the final score. Such extra credits may include body type differences, passive abilities, or anything else which has a significant effect on the feel and playstyle of the fighters. Binary passive abilities (those which are common to many fighters and either are or aren't present) such as crawling and wall jumping have halved weighting. Idle animations will not be considered, as these rarely come into play. Minor differences will also not qualify, such as Luigi being slightly taller than Mario, but otherwise having the same body type.

Scoring
Below is a set of tables showing the individual point values for every move or stat, a brief explanation when applicable, and the total score, for each clone. Minor stats are indicated with small text. Only fighters covered on the Clone article will be assessed, but I plan on producing some bonus content later on.

Fighter: Luigi Parent: Mario

Fighter: Dr. Mario Parent: Mario

Fighter: Young Link Parent: Link

Fighter: Dark Samus Parent: Samus

Fighter: Falco Parent: Fox

Fighter: Pichu Parent: Pikachu

Fighter: Lucas Parent: Ness

Fighter: Ganondorf Parent: Captain Falcon

Fighter: Daisy Parent: Peach

Fighter: Lucina Parent: Marth

Fighter: Roy Parent: Marth

Fighter: Chrom Parent: Roy

Fighter: Dark Pit Parent: Pit

Fighter: Toon Link Parent: Young Link * * Toon Link was Young Link's replacement in Brawl, and served as a more extreme version of Young Link's archetype. In Ultimate, he has more in common with Young Link than Link, and thus only this pairing will be assessed.

Fighter: Wolf Parent: Fox

Fighter: Isabelle Parent: Villager

Fighter: Ken Parent: Ryu

Fighter: Richter Parent: Simon

Evaluation
Here, I have graphed the scores of each fighter, to help visualize their scores relative to each other. Echo Fighters' bars are coloured red. This is the result: As you can see, there are certainly some controversial picks here, but overall I think this has turned out pretty well. The fact that Echo Fighters naturally floated to the top of the scoreboard is certainly reassuring. So now, the question inevitably becomes: "how would you classify each of the clones"? To be honest, I don't think the 3 category model used on the Clone article is no longer nuanced enough to satisfactorily categories the clones, as of Ultimate. But, if I had to categorize them, based on the graph, this is how I would divide the categories:

Final thoughts
Obviously, the results I obtained don't quite line up with the current interpretations on the Clone article. That's ok. I very much expected that, since it uses a similar method to Cloneosity, and therefore, in my opinion, is also somewhat problematic. Maybe my take on assessing the clones of Ultimate might help, but I'd be lying if I said I expected that to be the case, since the Wiki seems pretty set in its ways about this.

If you have any questions, suggested improvements, or general comments about this project, please be sure to voice them on the talk page. Thanks for viewing!

Bonus
Disclaimer: unless stated otherwise, assume that any bonus content is opinion-based.

DLC Loaned Moves
Loaned moves are moves that resemble a move that another fighter possesses, but which belong to a non-clone character. DLC fighters have quite a few of these.