Playing to Win, Part 3: Not Playing to Win

Ok, ok. I’ll let you in on the secret: “playing to win” at all times is counter-productive. If you want to win over the long-term, then you can’t play every single game as if it were a tournament finals. If you did you wouldn’t have time for basic R&D, you’d never learn the quirky nuances that show up unexpectedly at tournaments, and you are likely to get stuck honing sub-optimal tactics.

Basic R&D

Playing to win and playing to learn are often at odds. If you play the game at hand to maximize you chances of winning, then you won’t take the unnecessary risks of trying out new tactics, counters, moves, patterns, or whatever. Playing it straight is the best way to win the game at hand, but at the cost of valuable information about the game that you may need later, and valuable practice to expand your narrow repertoire of moves or tactics.

Here’s a simple example from Street Fighter. Let’s say I know for a fact that one split second from now my opponent will do a particular “super move.” To win the game at hand, the smartest thing to do is just block the move, but that doesn’t teach me a whole lot. How invulnerable is his super move, anyway? Could I have stuck out an early kick that would knock him out of his super? Or could I have waited for the “super flash” to happen (signifying the beginning of his super move) and then done an invulnerable dragon punch 1 frame later? Maybe my invulnerability will last longer than his and I’ll knock him out of it. Maybe his will always win. That’s valuable information to have for the time when you have zero energy and the opponent forces you to block the super move and die. This situation will happen in the tournament, so you better know what your options are.

Very often in “casual play” I will forgo the safe option in order to try possible counters to certain moves. Even if I lose a game when a possible counter turns out not to work, the knowledge gained is well worth it, since I’ll never make that particular mistake again (I hope!). If you really want to play to win, you have to know all the options open to you at every moment, and that doesn’t happen without a lot of disastrous experiments.

This concept applies to pretty much any game, of course. “Will my 6 corsairs really beat his 12 mutalisks in StarCraft?” Or, “I know I have the flak cannon, but will the shock rifle combo work just as well around corners in Unreal Tournament?” You will never know unless you try it.

Honing Sub-optimal Tactics

Early in a game’s life, players have not yet figured out which strategies and tactics are actually the best…though many players will claim to know all. Those players may very well know better tactics than other players of their time, but games evolve. New things are discovered that obsolete old tactics. Usually, radically different and better tactics are discovered that put the old ones to shame. Sometimes, new counters are discovered that can entirely defeat the old “best” tactics. In a fighting game, you also have the concept of figuring out which characters are the best. It can take months (or years!) for players to figure out that character X, though widely thought to suck, is actually able to abuse bug/feature Y in such a way as to be nearly unbeatable.

So how does all this relate to playing to win? The hardcore “Play to Win” player will choose his one character, his set of powerful tactics, and hone them to perfection over time. He’ll know all the tricks for that character to perform those tactics. For example, in the fighting game Marvel vs. Capcom 1, he might pick Mega-man and learn the “rock ball trap.” This a pattern of attack where mega man creates a soccer ball (“rock ball” in Japan), kicks it diagonally across the screen, then fires one blue projectile in the air, then one on the ground. That’s 3 projectiles total controlling the play field. While the opponent deals with that, Mega-man has time to summon another soccer ball and repeat the pattern.

A serious Mega-man player will learn the rock ball trap variations needed against Chun Li, the different variations needed against Venom, and so on. Other players will find tricks to negate the usefulness of the rock ball trap in general, then the Mega-man player will find the counter-tricks that allows him to keep the pattern going. This will feel a lot like “Playing to Win,” but in the end, this player will do precious little winning. He will have mastered a sub-optimal tactic that in the end is not bad, but isn’t 1/10th as good as other things that other characters can do.

I think of a game as a topological landscape with lots of hills and peaks that represent different tactics/strategies/characters. The higher the peak, the more effective that strategy is. Over time, players explore this landscape, discover more and more the hills and peaks, and climb to higher locations on the known hills and peaks. Players can’t really add height to these peaks; they are only exploring what’s there. The problem is, when you reach the base of a new peak (say, the rock ball trap peak), it can be very hard to know that the pinnacle isn’t very high. It might be really difficult to climb (lots of nuances to learn to do the trap), but in the end, the effectiveness of the tactic is low compared to the monstrous mountains that are out there. You have reached a local maximum, and would do better to exploring for new mountains.

In other words, playing to win involves exploring. It involves trying several different approaches in a game to see which you are best at, which other players are best at, and which you think will end up being the most effective in the end. When you are perfecting your rock ball trap (your best chance of winning at the time), you have to realize that “playing to win” might actually involve taking up a new character you know nothing about…a character that you will eventually play 10 times better than you could ever dream of playing Mega-man.

Learning Secret Lore

Tournament play often creates critical moments of decision when you are exposed to a very strange situation in the game. In a tournament, the best players get to play each other, often with a clash of play-styles. They each have their own tricks and must find immediate answers to the tricks of their opponents. And it’s not just for fun anymore, it’s “real.” It matters. Under this pressure players find creative and unusual solutions to they tricky spots they get put into.

When these strange situations come up, will you be familiar with them? Do you know the options and the risks involved? Knowledge of “secret lore” or unusual interactions in a game often means the difference between winning and losing.

And how will you learn this secret lore? Perhaps you are preparing for a tournament, practicing, playing to win. What will you practice? You’ll practice the things you know you need to do the most in a match. You’ll practice against the things that you know you’ll face? Basically, you’ll do it all “by the book.” Consciously preparing for a tournament is pretty much the opposite of exploring “unusual situations.” In your practicing, will you seek out a player of a character you think sucks? Will you play characters you have no intention of playing in the tournament? Probably not. But what happens when a mysterious player out of nowhere shows up with that “sucky” character, and shows everyone how good that character really is? That other character you were messing around with might be just the thing you need…too bad you didn’t explore that. You were “playing to win.”

The Karmic justice of it all is that love of the game really does count for something. Those who love the game play it to play it. They mess around. They pick strange characters, try strange tactics, face others who do the same, and they learn the secret knowledge. Those who play only to win can’t be bothered with any of that. Every minute they spend playing goes toward climbing their current peak, attaining their local maximum. Perhaps they don’t even like the game enough to be bothered with anything except the most mainstream character and the most mainstream tactic with that character.

I practiced pretty hard for a tournament in Super Turbo Street Fighter that occurred on August 9th-11th 2001. Before the tournament, I decided to play only Dhalsim and to practice him a lot against whoever I could. I also happen to actually like the game, and I’d sometimes mess around with my “fun characters” of Honda and Ryu, and occasionally with my “professional” character: Bison. Dhalsim was my focus, though.

When the actual tournament came around, I would have never guessed what it all came down to. My Dhalsim did well, and it came time for me to face a well-known Japanese player who plays T-Hawk. T-Hawk is known to be terrible, especially against Dhalsim, but this was a prime example of a player who could work magic with a “sucky” character. After one game, my Dhalsim was utterly destroyed, and I needed a change of plans. I figured that my “casual play” Honda would do well, since I could sit and do nothing the entire game and be safe from T-Hawk. If he ever got near, I could head-butt and knock him away, then sit and do nothing. (See my article on The Art of War: The Sheathed Sword.) Anyway, my performance, a true exhibition of stubbornness and boringness in tournament play, paid off. I defeated the Japanese player in an utterly ridiculous character matchup that no one would ever predict actually happening in a tournament. I went on to lose another ridiculous character matchup against a different Japanese player, but that’s another story.

The unlikely moral here is that playing to win is often counter-productive. Those who love the game and play to play will uncover the unusual nuances that might be important in a tournament. Those nuances might never be important, but the “play to play” player doesn’t care. It’s all for fun, and he’s happy to accumulate whatever knowledge he can. The “play to win” player might lock himself into perfecting certain tactics/strategies/character that will eventually be obsolete, as hard as that will be to believe at the moment. Meanwhile, the player who is able to take a step back and mess around will either discover new mountains to climb, or at least take a stab at climbing some other known mountains. The joke’s on you when his mountain turns out to be 10 times higher than yours.

Postscript—

Months after writing the above article, I traveled to Japan in March 2003 as part of Team USA, representing the US in Super Turbo Street Fighter. I also played a bit of Capcom vs. SNK 2 over there. One interesting thing about Japanese players is that they stick with just one character (or one team of characters in CvS2), since their tournament format requires keeping the same character the entire tournament. In the US, we can switch characters between games, giving us an incentive to learn at least 2 to 4 different characters.

The Japanese players definitely proved to me that by sticking to one character and learning EVERYTHING about that character, you win the unwinable matches. In both Street Fighter games I played in Japan, I saw Japanese players who devoted themselves to supposedly weak characters and demonstrated the topological peaks for those characters are miles higher than I had realized. One might think that invalidates some of the points I made in this article…yet the winner of the CvS2 tournament used the same old unfair, broken characters and tactics that we’re all aware of (A-groove roll-canceling Blanka/Sakura/Bison for those who care). That same player, Tokido, won the CvS2 portion of the 2001 tournament I mentioned above, so perhaps he’s proved my point after all. He’s identified what many players agree is the highest peak of that game, and devoted himself to perfecting it. Unfortunately he’s an incredibly boring player, but nonetheless a boring player who won the US National and Japan National tournaments!

--Sirlin

Comments

[Melissia]

I have to agree with this. Recently, the Dawn of War II Beta began, and I participated in it. The beta feedback forum was filled, on day one, perhaps even hour one, with "OMG IMBA" and "[X Unit] IS OP!!!". This is because people tried to use tactics they knew from the previous game, which is balanced completely differently from the new game, and quite frankly wasn't helpful at all to Relic.

In reality, I've actually found to be one of the most balanced RTS games ever made, with just a few abuses that need to be fixed and a few weak units that could use an extra something to make them worth their cost. A lot of the reasons behind the cries of imbalance and overpowered is because people have their favorite faction, and their favorite strategy within that faction, and despise changing things up. (This is also the main reason why people say it's bland and lacks variety as well, even though most good players will always change their strategy halfway into the game to adapt to how their opponent plays).

A lot of this could be solved by... well.. just playing for fun. The game isn't even releasd yet, it's in its beta testing phase, noone knwos all the nuances ofa ny single race, nevermind the entire game. Even in well-developed games such as street fighter, noone really knows all the nuances of every single character, and people come up with new strategies which overcome the old ones very frequently. Playing one strategy one game, another the next, and another the next, and so on, would be a far, far more helpful thing to do for the beta testers, and Relic is more likely to listen to someone who has honestly played as much as they can, in every way they can, than someone who picks a faction, picks a strategy, and sticks with it.

[Flopsy]

"I think of a game as a topological landscape with lots of hills and peaks that represent different tactics/strategies/characters. The higher the peak, the more effective that strategy is. Over time, players explore this landscape, discover more and more the hills and peaks, and climb to higher locations on the known hills and peaks. Players can’t really add height to these peaks; they are only exploring what’s there. The problem is, when you reach the base of a new peak (say, the rock ball trap peak), it can be very hard to know that the pinnacle isn’t very high. It might be really difficult to climb (lots of nuances to learn to do the trap), but in the end, the effectiveness of the tactic is low compared to the monstrous mountains that are out there. You have reached a local maximum, and would do better to exploring for new mountains."

Simulated annealing?

[chs nghtbrngr]

Dawn of War: Dark Crusade has a pretty wide variety of options for Nercon players playing to win. One technique I developed very early on was the Tomb-Spyder multiply. The game, for a reason I'm not sure of, has a delete function that allows you to kill your units. Necrons are designed to come back from the dead with even more power, and that philosophy comes out in the Tomb Spyder. It can harvest dead bodies of Necron units and spawn new squads from the old. The multiply involves building three full squads of Necron warriors, which are cheap and easy to produce. (3 x 8 units per squad, so 24 units.) I set my tomb spyder to collect corpses, and before I knew I had eight new squads. I built them up to max strenght and did it again. I attacked my opponent very early with a large number of units. This probably wasn't imagined by the game designers, but it does kinda fit your article of using un-thought of tricks to win. Is it cheating? No, not in the least. I'm simply using different legal unit and game abilities to grow my forces.

It just goes to show that even RTS has tricks that can be used without hacking the game.

[The_White]

Heh, great article this one. I got a few more examples for anyone wanting more info. These are all from SSBB

Me(kirby) VS mate(Wario): Wario starts his chomp move, one that keeps him stationary for an amount of time but causes him to eat anything that touches his mouth, if it is a player then he chewes on them for a while before spitting them out. Since I was in the air and close to him the safe move was to trigger a Final Cutter (shoots me into the air with a sword drawn then slams me into the ground creating a shockwave) thus getting a small amount of free damage. In the air, Kirby has another option, Stone form, which turns him into a rock dealing large damage and knockback to anything hit by it. This was a risky tactic because I wasn't sure which move would win. So I gave it a shot... turns out his attack beat my attack and I ended up loosing the match. On the up side, I now know that the safe move is actually the better option.

Me(wario) vs mate(can't remember): I'm recovering with Wario and I know that the safe move is simply to recover but the risky option is to use his chomp to drag the opponent down (a tactic that I know for a fact with my main character, Kirby)... long story short, I failed and lost. Another lesson learnt.

Now these examples seem to say "just go the safe way" so here's another one

Me(Falco) vs mate(Ike): While I'm recovering from off the edge an item appears that when thrown, latches onto the ground and starts circling the stage dealing damage to everyone except the guy who threw it. To make matters worse he used a fire based attack to charge it up even further before throwing it. The safe option is simply to dodge the item as it makes its passes. Instead I went the risky way and threwout Falco's reflector (bounces projectile attacks) in the hopes it would work on this item. It did, the thing reversed direction so I had an extra way of avoiding it. Good but the safe option is still easier. Item rolls around the stage and comes in contact with the other character, hits him and gives me the win... my reflector had turned his item into my item o_O I had just discovered a way to compltely negate the usefulness of an attack (and in fact, make it harmful to the guy using it) with any character with a reflector by taking a risk and playing to learn rather to win. Now when I play to win, my mates are very cautious about using that item against me.

[Velen_Z]

I am a player in the Brawl community, but I consider myself on a middle ground as to the kind of player I am. I am both casual and competitive. Playing to win is fine, but I find playing for fun to be better than playing to win. Unfortunately, your book "Playing to Win" has caused the social chasm between the casual and competitive players to widen and deepen, and it is cited as used as if it were the scriptures of a dogmatic religion by the people in the competitive community who are boneheads.

It has made the casual players considered by those in the competitive community to be socially inferior.

I for example, play Bowser, who is low on the tier list. A lot of people judge the way Bowser plays by their experiences with him in Melee, as being the second worst character instead of judging him by how well the player can actually play with him. There is however a motto that is starting to rise form this.

"Don't knock a Bowser player until you see how they play."

The majority however go with their past experiences and underestimate Bowser's potential, which is part of the reason he hasn't gotten much progress in terms of growth in either Melee or Brawl: because people are not willing to try and discover things for him.