(I’m writing a series of journal entries to track my progress prepping for Evo this year. Check out the previous entry here.)
Today I’m going to write about the process of improving in fighting games. In general, we tend to romanticize the work people do to improve at something as a mystical invocation of human dedication and grit, but really, you can boil down the process of improving in pretty much any competitive activity into a three-step cycle:
- Generate a hypothesis
- Test that hypothesis
- Use results to inform new hypothesis
When people talk about how better tutorials won’t fix fighting games, it’s worth revisiting ye olde arcade days, where this cycle was built into the pattern of engagement, and it looked like this:
- Quarter up, watch the person playing, and look for clues on how to beat them
- Try to beat them
- Lose, get back in line, think of new ways to beat them
Now let’s look at the average online play experience:
- Log in and queue up for a match against a rando you’ve never played against before
- Play the match
- Try to learn something from it, queue up again, play another rando
If you feel like you’re stagnating, try breaking down how much time you spend in each phase, plus what you’re doing for each phase, and see if any obvious gaps stand out.
For example, I was getting a little frustrated last week because I felt like my Laura had plateaued, so I sat back and took at look at what I had been doing and realized that while I had been playing lots of games, I hadn’t been consciously looking for new tools to add to my game, and so I was losing matches because I was trying to use the tools in my comfort zone better instead of expanding my knowledge of the character. Basically, I did plenty of testing, but I wasn’t getting new material in to generate new hypotheses, so I spent an hour watching replays of high-ranked Lauras on CFN.
After an hour, I had the following set of notes:
replay review Laura notes
will do fireball in neutral range even if it doesn’t seem to lead to much
constantly pressing buttons
likes ex grab
vt -> cFP xx ex fireball -> HP bolt charge
pokes with sRH, sMP? maybe an anti poke?
got the left-right cross under stuff with sHP
neat dizzy combos starting with full charge fb
dash cancels off of sHP
good at hit confirming into MP bolt charge
uses EX fb as pressure tool
how to use f+HP?
lots of CH hit confirms
ex fireball -> dash grab
cancels into FBs a LOT if unpunished
dizzy combo: level 2 HP fireball, fHP fHP ex FB HP bolt charge
You can see that these notes are all over the place; some of them are looking at specific habits (setups, frequent pokes, fireball usage), some of them are around skills I know I’m lacking in (like confirming counter hits, which unlocks a lot of character power in SFV), and others are general observations like “constantly pressing buttons.”
Now, I don’t have the full picture, because I can’t read the players’ minds while I’m taking these notes. All of these are basically pre-hypotheses — I observe this behavior in players who are more successful than I am, try and figure out what they were thinking that led to said behavior, try out the behavior myself to see if it fits or not, and work from there to refine my set of gameplay techniques until I have an understanding of which tools work against which characters/players. I suspect stuff like EX fireball -> EX command grab is one of those things that only works once against good players, but I won’t know until I try it a bunch and figure that out!
It’s important, I think, to build the hypothesis-generating muscle in your brain, even if you get it wrong a lot. Sometimes you’ll copy a behavior and find it’s successful for different reasons/in different situations than you originally thought. That’s ideal, IMO, because you get a better idea of both how better players play, and how better players think.
What if you’re a tech monster who’s constantly watching stuff and labbing it out but still doesn’t see improvement? Well, your iteration loops might be weakest in the test phase — if you’re just playing matches against online randos, you’re not testing your tech against consistent opponents, which introduces a lot more noise into your results. If I want to test whether I should be using Laura’s EX fireball to prolong pressure strings, and I’m playing against whomever the matchmaker pairs me with, it’ll be harder for me to tell whether it’s a good idea or not than it would be against someone who I can play with consistently, since in the latter case I can just play enough games to compare results.
So: If you’re stuck in a fighting game rut, try breaking your play habits down into those three steps, look for areas of high or low focus, and organize your broader training goals around taking on small, testable hypotheses. This way you can break down your improvement by smaller, manageable, easier-to-understand chunks, and track your growth over time by what you’ve learned, not just who you’ve beaten. Give it a shot and let me know how it goes!