Customization in Fighting Games

Gameplay customization is something everybody asks for without really thinking it through. I usually write these video companion pieces as a short introduction, intended to be read before watching the video. For this one, let's do things in reverse order. Please give it a watch before continuing on.

If you've watched the video and generally agreed with my position, you'll recognize that:

• Customization can just completely break stuff

• Customization done wrong can lead to less player choice

• Representing customization in the UI during matches can be difficult, depending on how customization is implemented

I capped off the video by explaining how (Ultimate) Marvel vs. Capcom 3 sidestepped most of these issues and is one of the most successful customizable fighting games to date, even though I don't think people generally view it through this lens. People might not hear Marvel vs. Capcom and think, "Oh yeah— that customizable fighting game!", but building your own team is a huge part of the appeal. Since then, versus-style fighting games have become a full-on fighting game subgenre with different developers throwing their hat in the ring, so I think it would help to have tools to evaluate how well they pulled it off. Measuring success lies in these two questions:

1. Are there lots of team combinations which are competitively viable and meaningfully different from one another?

2. How much does character synergy matter?

To try and figure out the answer to question 1, you need to figure out the total number of combinations possible. Then, remove the ones that aren't competitively viable and the ones that aren't meaningfully different. It's pretty easy to figure out that first part: the total number of combinations. In the video, I cite UMvC3 as having around 19 million. Here's the math:

50 characters * 3 assists * 49 characters * 3 assists * 48 characters * 3 assists * 6 possible team orders = 19,051,200

The hard part is filtering the combinations that are garbage tier and/or basically the same as another combination. Since it's impossible to test them all out, I think you need to look at lots of high level tournament play. If very few people are playing the exact same team, you probably did a pretty good job. If you see lots of people playing identical teams, things didn't go great. While there were a few prominent teams played by more than one player (Zero/Vergil/Dante, for example), unique teams were more common, so I think UMvC3 did pretty well here. Question 2 is quite a bit more subjective, but I think pretty much everyone would agree that UMvC3 has extremely meaningful character synergy. If someone asked me "How do I fight against Magneto?", my answer would be, "How do you fight against Magneto what? Are they using Sentinel drones assist? Doom hidden missiles assist?" While the U.S. has historically dominated the UMvC3 competitive scene, the Japanese player Kusoru famously won an early UMvC3 tournament, and his victory was in part a product of nobody knowing how to fight against Rocket Raccoon log trap assist. When evaluating any potential UMvC3 successors, I think these are the two metrics needed to see how worthy they are of that title.