How do I practice playing against the player?

Patrick Miller
13 min readJul 25, 2022

[This essay was funded by my generous Patreon supporters. If you liked this and want to see more, please consider joining the crew!]

Hi Pat, I’ve got a question about how to prioritize my improvement.

I started playing fighting games when Covid hit, and I’ve been grinding hard ever since. I’ve been thinking about my strengths and weaknesses as a player. I think that I focus a little bit too much on “solid” play, as opposed to “exploitative” play.

When I say “solid play”, I’m talking about a playstyle that’s focused on guaranteed value: optimizing combos, optimizing oki, weighting your mixups to optimize risk/reward, increasing your option coverage, etc. Solid play works against any player, no matter how good they are. These are the baseline strategies that you can always fall back to, minimizing your reliance on reads.

When I say “exploitative play”, I mean a playstyle that’s focused on punishing your current opponent. This is all about gathering information, then using that information to modify your play in a way that counters your opponent. A simple example is blocking more against an opponent that uses DP too much. Or, if you notice your opponent is only playing Rock/Scissors in some situation, you can start to play Rock more.

When I start a new game, I feel like there’s a lot of low-hanging fruit for improving my solid play: learning system mechanics, anti-airs, oki, bnbs, etc. I think I’ve collected most of this lowest-hanging fruit for the games that I play.

I’m at a point where my play sessions are focused more on the game’s “high-hanging fruit”: finding bits of improvement that will add up over the long term. This consists of stuff that’s difficult, obscure, or both. For example, I could work on my consistency for a universal combo that I mess up frequently, or I could learn an easier combo that’s corner-only, lightweights-only, requires 75 meter, and isn’t burst-safe. In +R, I could work on learning an FRC timing that I’m not comfortable with. There’s also matchup-specific stuff I can work on, like practicing reacting to a certain move, fuzzy-guarding a particular mixup, or finding an answer to a tough situation.

When I’m playing, I usually try to focus on some specific learning targets. I’ll go for a combo that I usually drop, or I’ll try to react to an overhead that usually hits me. In short, I like to play with confidence, pretending that I won’t make execution errors. Of course, I’ll make unforced errors, and I’ll lose games over it, but this is fine. Maybe I want to practice a certain combo, so I’ll fish for that combo starter. Succeeding in the thing that I’m practicing is more important to me than winning the game, because the goal is to get long-term improvements to my autopilot.

On the other hand, sometimes I try to work on my exploitative play. I’m not sure of the best way to do this, but I try to direct my focus to what my opponent wants to do, and I try to keep track of how they respond to common situations. Then I try to pick appropriate strategies to beat what they want to do: maybe I’m reading their approach angles, or mashing out of a predictable gap in their pressure, etc.

Here’s my problem: I don’t think I can effectively practice both playstyles at the same time. I’ve tried, and splitting my focus has poor results. I think that when I’m practicing exploitative play, I have to stop gambling with risky execution. Instead of dropping difficult combos, I need to settle for easy combos. Instead of trying to upgrade my autopilot, I need to make the best possible use of my current autopilot.

I feel like I really need to work on my exploitative play, since I think it’s the thing that’s holding me back most as a player. But there’s all these cool combos I want to internalize, there’s all this setplay that I want to work on, etc. There’s so much work to be done in all my matchups! I really enjoy gathering matchup knowledge, improving my execution, and doing cool and difficult things. Focusing on exploitative play is not as satisfying for me, because it’s harder to see concrete improvement from it.

I usually try to play long sets when possible, so if I constantly mess something up, then I expect my opponent to notice. this can cause… structural change in our “player matchup”. the set can reach a kind of degenerate state, polarized around some key interaction.

For example, let’s say that I fight a Slayer who tries to wakeup backdash cancel every time, but he keeps getting hit by my meaty. That means that I can probably just win by looping oki off one knockdown. Against strong Slayer players, I get way less off a knockdown, and I have to plan to win a lot more neutral interactions. The Slayer who’s missing all the BDCs won’t face my “main gameplan” in the matchup, because the holes in their execution change my evaluation of how good a knockdown is, and therefore the way that I play neutral.

If that Slayer repeatedly tries to execute something that isn’t working, the matchup can enter a stable state: Slayer is doing the thing that works on paper (but messes up the execution), and the other player has the upper hand by doing the thing that *actually* works in the here and now.

I’m looking for advice on how to spend my focus. Do you have any suggestions for deliberate practice of exploitative play? How do you balance between learning your character, learning the matchup, and learning to beat your current opponent? Do you work on all of these things at the same time, or do you cycle between them?

Thank you,

Trans Rights

P.S.: This “value” vs “exploit” dichotomy is just my poker-style perspective on fighting games. Perhaps I should’ve drawn a sharper distinction between matchup knowledge and execution. Criticisms or alternative perspectives are welcome.

Hi Trans Rights,

Excellent question! Let’s go in on this.

You’ve identified that there are indeed layers to this shit; I usually refer to them as playing your character, playing the matchup, and playing your opponent. Combos, setups, and pushing your execution typically fall into the first category (barring matchup-specific combos and setups, anyway); neutral, resource management, and matchup-specific stuff fall into the second category; stuff that involves layered reads and rotating options falls into the third.

If I’m working on my character, most of that starts in the lab until I’m ready to take the new stuff to PvP. If I’m working on a matchup, it starts with reviewing videos to figure out specific things to work on, then labbing the options until they feel comfortable, then practicing that matchup with actual people.

When it comes to working on reads, it depends on what I’m practicing for. In a tournament setting, the goal is to quickly identify patterns, so my work starts by examining what i do on layer 1 — for example, mapping out my go-to options for the first knockdown in the first round of the first game against an opponent I’ve never played before, identifying the common set of responses, and mapping out the layer 2+ options based on that. In a tournament first-to-two set, I’m probably not going to have enough information to get a deep download on anything besides some relatively superficial stuff like the opponent’s patterns on air tech/reversal/frame trap situations, so the goal is really just to quickly get a read on that stuff in case I need to bet on it at some point in the set.

If I’m prepping for a specific opponent in a long set, then most of the work starts with reviewing what I know about the character matchup, studying footage of that player to see how they play the character/matchup, identifying their layer 1 choices and then prepping as much as i can to rotate different options in offense/defense/neutral so I can be ready to go deep. All of this work ends up being valuable for tournament play as well, because when I do it enough I eventually find new options and adaptations that I can bring into shorter sets, but probably wouldn’t come up if I was only playing shorter sets.

All of this stuff is worth doing, and you should be doing all of it! But, I wouldn’t worry about doing it all at the same time. I usually do the things I feel like working on at any given moment because I prioritize my own enjoyment pretty heavily, but I play in brackets pretty often, so it’s easy for me to chase stuff to work on based on whatever I lost to most recently. In general, I’d recommend having a specific thing or two to work on in a session over trying to work on ‘everything’ because if you’re working on everything, you’re kinda working on nothing.

I had no idea that this meme had a happy ending version.

So let’s answer your second question first: I usually rotate between working on things at different layers mostly based on whatever I feel like working on.

Not every opponent will give you the opportunity to work on a specific thing, though. So, when I am working on a lot of things with a character, I’ll have a notes doc where I track what I’m working on, and how I’m working on it. Stuff like “Drill TK Alpha Blade in training mode” eventually graduates to “Practice TK Alpha Blade in casuals”, and so on. If I have matchup notes I need to practice, they might start out as “Figure out how to deal with XYZ situation” and then turn into “Remember to do ABC in XYZ situation.” If I’m in a set where someone isn’t playing the matchup that I wanted to practice, then I’ll use it as a chance to practice something that isn’t matchup-specific. It’s all growth, and I’m not particularly worried with what order it comes in.

That said, I do think you can do stuff like work on risky combos while also practicing a matchup, it just means you’re probably going to lose more. Which is fine, unless you’re relying on wins to tell you whether you’re doing something correctly. I would not recommend relying on results for feedback! But I figured I’d mention this because it’s the main thing I could think of that potentially keeps you from juggling multiple improvement points. The more you get used to playing with the higher-risk stuff in your game, the less risky it gets, so there’s no real reason to not go for it when you’re just playing games.

The nice thing about refining your character and matchup-specific stuff first is that it makes it a bit clearer exactly which chunks of the game are dependent on the opponent. If your safejump timing is weak, then you’ll have to hard bait DPs more than you would have to if your safejumps are tight; if you’re picking the wrong buttons in the matchup in neutral, then you’ll lose neutral more often and have to work on multiple layers of defense more than you would if you were winning neutral half the time. Player-specific work is volatile because gambling on human psychology is volatile, so if you’re trying to optimize for tournament competition, you generally want to save the big reads for when you absolutely need to use them.

In any case, I’d recommend just working on stuff a little bit at a time based on whatever ordering makes sense to you. there’s no need to rush on this stuff and it’s a lot more enjoyable that way IMO. But, if you want to focus on the highest-value work first, I’d suggest getting to optimal character control and refining your character matchups first before going deep on downloading. Control the things you can control, and whatever’s left will be reads practice.

Now let’s get to your first question: How do you practice playing against the player?

You could do all this stuff, or you could just get a Sharingan.

I mentioned some of the study techniques I use to identify player patterns earlier in this essay — stuff like player-specific video review, practicing as their character, and looking for consistency in their patterns for hard read situations. However, I think there’s a more valuable topic to handle here, and that’s about the value of the player reads you’re getting in casual sets.

As you recognized in your email, playing with the risky stuff means you’re going to drop more stuff and fail more interactions that you could have avoided failing had you chosen to do something more reliable, and your opponent will notice that, so you aren’t playing a “real” set per se, and so your ability to use these sets to practice playing against the player is limited.

We usually call this “playing casuals.” In martial arts, we would call this “sparring.” It is a zone where there are zero consequences to losing ten games in a row (save for your pride, perhaps), and since there are no costs, it’s an excellent place for both players to practice the parts of the game that you need to work on before you feel confident bringing them into a match with stakes. It’s particularly valuable when you’re playing with people who are significantly stronger or weaker than you, because it means both players can feel free to experiment without feeling like anything significant is at stake, and can spend the time practicing with a partner instead of focusing on trying to win.

(As an aside — one of the things I’ve noticed in the pandemic babies of the FGC is that they often assume that everyone is going as hard as possible all the time, and interpret anything less than that as ‘disrespectful’. I think it’s easy to feel this way when the vast majority of our matches these days are played in faceless netplay UIs with minimal communication, but personally I am of the opinion that it’s healthier to treat each match with zero expectations about the ‘seriousness’ of a match, because it is through less-serious play that we often end up learning the most about the game we’re playing. And besides, it’s more fun that way.)

Anyway, when it comes to ‘warping the set’ I would submit that you probably shouldn’t consider any set you’re having to be ‘real’ unless it’s in a tournament, there’s money on the line, or something else that would motivate both players to be playing at their absolute highest level.

Now, depending on the person, it might not take much to bring that energy. If you ask someone to play a first to ten set, you’ll likely get different results than if you had simply played until one of you hit ten wins without defining the limit up front. If we’re playing at a setup with a rotation, the level of seriousness scales exponentially with the length of the line because now there’s a cost to losing. And if we’re playing and then you call for the last game, that last game is for the title of Supreme Champion of the Universe, and every read from the set prior is cashed in so I can walk away the winner. Or maybe it just takes a system like Ranking Points to get you going at 100%.

In general, though, people don’t play the same when something is on the line. Even the most degenerate no-neutral-havin’-ass netplay monster will slow things down once they start playing in an IRL bracket. You can be warming up with someone and then immediately play them in round one of pools and they’ll feel like a whole different player.

Of course, some people don’t have an off switch for casuals; they only go 100% all the time. I generally consider this a sign of a newer player; if everything in your game is low-confidence, then there’s no difference between your casual game and your competitive game. And if you’re new and trying to test yourself, it’s not uncommon to go all-out against everyone all the time. It’s true in martial arts, too, and eventually it’s the kind of thing you learn how to tone down as you get enough competence and competitive experience to treat the two as different. (I wrote about this in How to be a good sparring partner, if you want more.)

Personally, my casual matches are usually 50–60% intensity, netplay tournaments like REV2SDAY or WNF +R are like 80%, and money matches and in-person tournaments are 100%. If I’m playing a long set with someone and I beat them a couple games in a row I’ll switch up my approach to try experimenting with new stuff or focusing on outplaying specific situations to keep things interesting and productive for me. If someone gets hit by the same setup N times, I’ll rotate it out to try different stuff instead.

Rotations are critical for competent fighting game play. Now go home and be a ninja.

Since you generally can’t assume you know what mindset or goals your sparring partner has for the session, it’s usually not productive to think about the state of the set itself as representative or meaningful in any way outside of the fact that the game itself is the same. That doesn’t necessarily mean you can’t pick up on a player’s patterns around hard reads or rotating options or whatever, but any reads you make will have to be caveated with a big disclaimer that says “*based on mashing fifty games at 3AM.”

So work on your execution if you want, or stick to reliable stuff if you want, or mix and match as you like based on how the game or the set is going, but in general assume that unless there’s something that clearly establishes both players have an excellent reason to win, nothing in the match is ‘real’ and your efforts are best spent sharpening your execution/matchup knowledge/decision cycling. You will get some data about your opponent, but you never know how differently they’ll play when there’s something on the line.

Good luck, and thanks for writing in!

Thanks for reading! If you found this essay valuable and want to support my work, please do not hesitate to share it around on your social channels, follow me on Twitter, check out my Twitch stream, or join my Patreon.


-patrick miller



Patrick Miller

a little bit miyamoto musashi, a little bit yoga with adriene.