Conversation with Halo’s Max Hoberman
After Capcom's Digital Day press event, I got the chance to talk with Max Hoberman for a while about game balance. He's lead designer of the newly announced Capcom game called Plunder, and he was lead multiplayer designer on Halo 2 and 3. I explained to Max that all I seem to do is balance things these days (Puzzle Fighter HD Remix, Street Fighter HD Remix, Kongai at Kongregate.com, and my own card games: Yomi and Spellblind). I asked if he had any advice.
His first advice was that no matter how great you are, you need a post-launch patch. Sounds good to me. I asked if he had anything general to say on the subject and he explained that he likes to focus on the fun first. Even in playing board games, if he can make a move that is suboptimal, but increases everyone's fun, he'll often do it. This prompted me to ask if he cares about balance at the highest tournament level, or only about pleasing the average player. He said that due to Halo's rise in various gaming leagues, he had to care about tournament balance in Halo 2 and 3, so yes he does shoot for that.
I also asked how much or little he relies on math and how much he listens to players. He explained a few math things he does, but mostly it's not about math. Even though I have a math degree from MIT, I use almost no math (on purpose) so it was nice to hear he had the same conclusion. *Why* we think this is deserving of an entire article, so I won't try to cover it here. On the subject of whether to listen to players, he was quick and firm in his answer that you generally cannot listen. He said players almost always lack the big-picture understanding of what they are asking for, are usually biased to buffing their own favorite things, and generally make suggestions that make the game more fun for them personally, rather than the larger view the designer needs to take. Listen to them all, understand what they're getting at, then come up with your own solution, he said. We both agreed that it's a strange situation that no matter what you decide, lots of people will feel you did the wrong thing. (This is because some people suggest doing X while other people suggest DON'T do X, so you always disappoint someone.) Them's the breaks.
With no prompting from me at all, he said that balancing a fighting game is far more difficult than balancing a first-person shooter. I said that while I agree, I think most people don't. He said, "Really? Why wouldn't they agree." I explained that I thought first-person shooter players feel their genre is somehow less exalted if it's easier to balance and they might not realize the extreme difficulty involved in balancing vastly different movesets to all compete fairly against each other in a fighting game. Max said it should be pretty clear to anyone that balancing fighting games is harder, and people should get over it. (And of course, play whichever games you find fun, rather than caring so much about which were harder to balance anyway.)
Finally, we zeroed in on the concept of symmetric vs asymmetric games very quickly. Max naturally understood this distinction without me having to even talk about it. (To the readers: in asymmetric games, players each have a DIFFERENT set of moves, yet they must somehow be fair vs each other. Symmetric games such as Chess maybe have many deep strategies, but both sides have access to (almost) identical moves.) I told him about my other blog post on this topic. Max and I agreed that balancing symmetric games is easier, almost by definition. He said it's fine in a first person shooter, for example, if one weapon is not as good as the others as long as it has some use. But it's not really so fine if a fighting game characters is not as good as the others.
Max asked which games my readers came up with in the asymmetric category. He immediately said StarCraft (everyone's answer!), but couldn't think of any other good examples (other than fighting games). I was kind of stumped. I couldn't remember anything good anyone had nominated, ha. Though I do still claim Magic: The Gathering, when treated as a battle of constructed decks.
The most interesting thing (to me) that Max said is that he thinks players really prefer asymmetric games (because it's a puzzle to figure out) but that game designers generally prefer symmetric games. Max himself says he strayed away from asymmetry all he could in Halo and where it does occur in some gametypes, the teams each switch sides to even things out. Considering the difficulty that all of us have even naming good asymmetric games, I'd say Max is right that game designers tend to not make them, and I also bet he's right that designers consider it too hard. And yet...that's all I do!! All FIVE of the games I listed at the start of this article are asymmetric. And 100% of competitive games I've ever planned to do and still plan to do are also asymmetric. I guess I'm way on the other side of the tracks from other designers on this one. One of the main interesting thing in competitive games is seeing how different characters/races/moves stack up against each other.
I hope you gained some insight from Max Hoberman in all of that. I'm literally going to go fill in spreadsheets of balancing information right now, ha.
--Sirlin
March 16th, 2008 at 2:51 pm
Just out of curosity, what’s your definition of “math” applied to these games? As in, Unit A has 10 health and deals 5 damage per hit, versus Unit B who has 15 health, yadda yadda?
March 16th, 2008 at 3:55 pm
General rule of software development is listen to customer complaints but not their proposed solutions.
Perception is reality. If customers say a product is hard to use then it is. If customers say a game is unbalanced then it is. (Even if you know for a fact that it is not) If customers say the framerate is bad or the control is bad or whatever then it is, period.
The problem with listening to customer solutions is, like you said, they propose solutions without a broader context and understanding and often times the solutions don’t even address the underlying problems anyway.
March 16th, 2008 at 3:57 pm
I don’t even know where to begin on that question. Basically yeah.
Example Street Fighter variables:
# of frames of startup on T.Hawk’s dive
initial x-velocity of T.Hawk’s dive
initial y-velocity of T.Hawk’s dive
x-acceleration of T.Hawk’s dive
y-acceleration of T.Hawk’s dive
All four of those previous variables again for the recovery of T.Hawk’s dive
Damage of T.Hawk’s dive
…and hundreds of thousands more
Example Kongai variables:
Rumiko’s Ninja-port:
If hits, causes Rumiko to teleport away and sets the range to far.
energy cost: 35
damage: 15
range: close-only
speed: 4
damage type: light magic
% to hit: 95
% to crit: only base chance (I forget if it’s 3 or 5%)
Example Yomi values:
Shinryuken:
type: super
cost: Ace plus up to 3 additional aces
damage: 10 + 9 for each additional ace
block damage: 5
speed: 0.0
combo points: n/a (or 5)
combo type: can’t combo
Example Spellblind variables:
Unnamed card
This turn, counter all effects that increase a spell’s value.
Mana cost: 2
Spellpower: 4
Refresh: 3
Where do all these numbers come from? This is maybe somewhere between 0.01 and 0.1% of all the variables involved in these games. Or at least that’s a quick guess. At first glance you’d think it takes a lot of math to fill in thousand and thousands of numbers that have to somehow end up being diverse and fair.
March 16th, 2008 at 8:22 pm
“At first glance you’d think it takes a lot of math to fill in thousand and thousands of numbers that have to somehow end up being diverse and fair.”
So when you say “we don’t use a lot of math,” you’re saying you just kind of guess what most of the numbers should be rather than using some specific formula to figure everything out (to make it sound simpler than I’m sure it is)? I’m curious because that’s something that’s been mulling in my head regarding a personal project of my own.
March 17th, 2008 at 12:00 am
Hi Sirlin,
Thanks for another good article. Regarding asymmetric games, I’m working on one too. I have always been fascinated with the many ways superhuman powers combined in comics, so unsurprisingly that’s the focus of my game. I’ve gotten a good deal of help from reading your various articles on playing to win and fighting game design and balancing. You wrote an article about one game in particular (the name escapes me at the moment, the one with the guy with the gigantic scalpel) that I found incredibly useful. The idea of a common core of actions available to all players was very relevant and raised the ‘minimum fun’ bar noticeably.
March 17th, 2008 at 2:45 am
Jasonimus Prime.
The game you’re referring to is Guilty Gear, or the current iteration, “Guilty Gear XX Accent Core”
March 17th, 2008 at 2:45 am
I must have missed the article/blog entry about symmetry in games.
In regards to asymmetric games, I think Alien Vs. Predator 2 deserves mention. It’s a perfect example of taking three very different character types with very different movement styles and weapons, and throwing them all in the same game. All the species play differently, yet none of them feel over-powered. The game has a plethora of options, some of which do break multiplayer balance (exosuits/queen molting kills). However, In a normal 4 on 4 TDM with class weapons enabled, it’s very fair.
Some people who play the game will complain that certain weapons are over powered, but i disagree because every weapon has a viable counter. Other people will complain that weapon stagger or predator speed hop exploits break the gameplay, but again I disagree, as these tactics also have counters.
It’s an old game, but it stands up well even in comparsion to todays generation.
March 17th, 2008 at 3:20 am
Thanks, Amp, that’s the one. Have you read his article on it? It’s very good, as I said.
March 17th, 2008 at 3:44 am
“I must have missed the article/blog entry about symmetry in games.”
It’s a conversation that sprung up in the “most balanced games” topic on the front page right now
March 17th, 2008 at 4:28 am
[…] Monkey: […]
March 17th, 2008 at 5:23 am
I’m all for asymmetrical games; not only is the puzzle of figuring out a character/etc. fun, it’s also fun to embrace a character/etc. as your own. Perhaps it’s comparable to a person’s love for one team in baseball.
World of Warcraft Battlegrounds and Arena are asymmetrical.
What about sport games where you have to choose a team?
Still, there aren’t enough asymmetrical competitive games around; keep up the good work!
March 17th, 2008 at 5:58 am
I fail to understand why people consider starcraft balanced. While some of their 6 matches are full of crap on how monotunus and unbalanced they are.
ZERG vs PROTOSS , i agree is a master pice of a match,
some others are decent
Now let’s go to ZERG VS ZERG.
OMG now massing mutas and spend the extra cyrstal in zerlings……..
That’s not balanced, zerg have no way to cunter Mutas, and thus in end in a mutas vs mutas, how’s that balanced?
Also you say you don’t remember anything good on the suggestion of balanced games.
Plz tell me what’s wrong in the wesnoth balance.
March 17th, 2008 at 10:26 am
“That’s not balanced, zerg have no way to cunter Mutas, and thus in end in a mutas vs mutas, how’s that balanced?”
Even if this were true, it would be a perfectly balanced situation. Mirror matches sometimes don’t have much variety, but they’re balanced just fine. Regardless, hydralisks are a good answer to mutalisks, as they are to just about anything.
March 17th, 2008 at 10:29 am
Ah, well, you asked for the most balanced asymmetric games, not necessarily for any good ones. I could name a few if we widen the question so… in particular Smash can then easily be included.
March 17th, 2008 at 11:51 am
Heh, Sirlin’s article on Guilty Gear XX got me interested in the game, and eventually hooked. Incidentally, I was just reading the frame data for Accent Core (it’s on dustloop.com), and, as Sirlin wrote above, the amount of numbers that go into these types of games is just mind-boggling. It’s a miracle that all that data manages to be very balanced against each other.
Good luck on your games, Sirlin. I’m looking forward to them.
~Z
March 17th, 2008 at 12:16 pm
Hydralisks are an excellent counter to Testament.
March 17th, 2008 at 8:49 pm
You couldn’t think of any good asymetric games, Sirlin? I mean…you were the one that brought up Team Fortress 2 in the original “balance” article. And Pokemon is quite asymetric…
March 17th, 2008 at 8:58 pm
And I didn’t mean to sound like an ass there.
March 17th, 2008 at 11:31 pm
Max believes TF2 is symmetric. The classes do not need to be in equal balance (it’s fine if one is worse and only counters another one or two classes) and players can switch classes during the match, which gives everyone equal access to all moves during a match. Not exactly like Ryu vs Zangief.
March 17th, 2008 at 11:52 pm
If TF2 is asymmetric, it’s to a much lesser degree than Street Fighter. It’s asymmetric because everyone does *not* have equal access to all moves at every moment… there’s a cost to using moves outside your current class (by changing classes), meaning that at any given moment, for example, you might have to try to fight a Sniper as a Heavy instead of changing to Spy to backstab him. You can’t just counter-pick your way out of every situation - a lot of the time, you have to live with what you’ve got. At the same time, though, being able to change classes at all during a single game makes it much more symmetrical than asymmetrical (unlike SF where character changes happen between games).
March 18th, 2008 at 1:35 am
To define symmetrical vs. unsymmetrical, go by what people are locked into before the match. For instance, in pro starcraft and competitive street fighter, each player must chose race/character before the game starts, thus locking them into a set of moves/abilities/units BEFORE the match has started. This makes it asymmetrical, because both players have different but fair abilities from the start of the game. Also, in the case of starcraft, the maps are a asymmetrical and players are randomly assigned one of 2-4 starting positions, so from the moment the game starts, you and your opponent are in strategically different, yet fair, positions.
Team Fortress has you chose AFTER the game starts, and lets you change your decision whenever you like, in response to your opponent’s decisions. The maps are symmetrical, as well. Therefore the teams are not in fundamentally different, yet fair, strategical positions before the game has even started. This means the game is Asymmetrical.
magic the gathering is asymmetrical in it’s constructed deck formats, because you are locked into a deck/sideboard before the tournament begins.
TL;DR
symmetrical = locked into different yet fair sets of strategical options before game starts
asymmetrical = not locked into strategically different positions at game start
M:tG and Starcraft = asymmetrical
TF2, Halo = symmetrical
March 18th, 2008 at 2:45 am
PoisonDagger: That’s not quite it really works. You’re never “forced” to fight a sniper as a heavy (a battle you would be certain to lose), because you’re on a team, and you can just ask one of your spies to take out the sniper before you proceed.
March 18th, 2008 at 3:05 pm
Didn’t some Splinter Cell do the guards vs spies multiplayer game?
March 19th, 2008 at 7:19 am
I don’t think it matters when your options are set. If the moment is NOW for me to use my options against you, and you to use yours against me, then at THAT MOMENT in time I have this to say: if our options are identical, this game is currently symmetric; if our options are different, then it is currently asymmetric.
A chess game begins almost symmetric (1st move actually determines our options to a larger degree than a non-player might believe) but can easily become asymmetric if, for example, I trade a bishop for your knight.
In fact, if you read Jeremy Silman (Reassess Your Chess, etc) he claims that all strategy in chess flows directly from “imblanace” which is the word he uses for asymmetry. That is, asymmetry is the most important feature of a chess game.
That’s a lesser kind of asymmetry than something like Protoss vs Zerg, but I think it’s crazy to say that a game like TF2 is asymmetric because you choose your character at a different point in the game than you do in SF.
A game is made up of moments, and if most of those moments are asymmetric, then so is the game.
TF2 might be boring, slow, repetitive, and easy (in other words state of the art), but it at LEAST has one nice feature, which is asymmetry.
March 19th, 2008 at 7:58 am
Ricefrog: That’s not the proper use of the term the way we’re defining it. A game can’t be symmetric “sometimes”, and not at other times. The term is being defined as a description of game design, so it’s a feature about how a game is made, not of how a game is played. Chess is symmetric, always, by definition. Street Fighter is asymmetric, by definition (even during mirror matches, cause that’s once again a feature of how the game is played, and is irrelevant when discussing design).
While you are correct that one can have an interesting discussion about how options vary over the course of a single game, it’s a completely different use of the word symmetry, and shouldn’t be confused with what Sirlin is talking about.
You’re right that just pointing out “when you choose your character” seems like a silly definition to use, and to some extent it is. However, it’s a simple thing that can pointed to that illustrates the concept. The real thing that makes TF2 symmetric is generally the team structure… that is a team of all heavies vs. a team of all snipers would be a stupid match, and noone plays that way, it’s completely degenerate. So what you almost always get is 1 heavy, 1 sniper, 1 demo… etc, etc, because all the classes have a good use. It doesn’t matter that 1 class counters another class, because teams usually represent all the possible counters at the same time, and noone plays 1v1 TF2, so all that diversity gets smoothed away. Add in the ability to change classes, and you aren’t even restricted to initial choice of team members, the way you choose a particular deck in MtG, and you lose even more of the chances for any asymmetry to actually arise during play.
March 19th, 2008 at 10:01 am
ricefrog yeah, I have to go with Clatyus. By your definition, just about every game is asymmetric, so it’s not that meaningful of a term anymore. Symmetric games have all sorts of wrinkly moments when advantages shift back and forth and temporary restrictions about. Even so, they still of a different character from asymmetric games.
Another point, I don’t know if we care so much whether a game is balanced accidentally or not. Even if various glitches happen to save a game and make it balanced, we could still look at it and figure out why it is balanced, and maybe how to apply some of that (even if was originally accidental stuff) to other games.
March 19th, 2008 at 10:38 am
Sirlin: Yeah, I don’t really disagree with that. It’s all interesting. I guess the issue that I was talking about in the other thread is more “where do you draw the line”. Going back to MvC2… obviously despite crazy tiers, there’s a very fun and competitive game there. However, half the characters don’t get played. So, if someone created a MvC2 re-release that was the exact same game with the bottom half of the characters removed, how do you classify that? Is it more or less balanced than the previous version? My opinion would be that it’s exactly the same game from the view of competitive players, since you only removed features that noone used. But, at the same time, I don’t like using that reasoning to say that MvC2 was “balanced”… since the other side of the coin is that the developers screwed up and should have created a game where the lost characters were useable, and removing them should have changed something.
I guess the question really comes down to… are we interested in the in-box version of a game, with all included features viewed equallly. Or, are we interested in the variants of a game that get played in tournaments to make up for shortcomings in the original version.
I’ve said before I use a rather harsh defintion, and stick with the former view of games. However, the main reason I do that is that basically any game is “balanced” if you allow people to just proclaim that their own house rules fixes any problem you come up with.
March 19th, 2008 at 8:19 pm
Clytus: I still suggest to measure ‘quality’ of balance by variety between viable ‘playstyles’ only. So no matter if you have 0 or 42 characters that are never used, the variety between the 8 usable characters counts. Still if you have 50 characters the probability is higher to have one of them beating every other than 8 being perfectly balanced among each other and still playing in completely different ways. Btw. in most cases following games in any series keep features ‘added by accident’ as official balancing elements if they were of any help.
Sirlin: About math I find it to be obvious that you still have to ‘guess’ a lot of things. First if you even have any formula you either have already locked most things or you have infinitely many solutions depending on how the variables are set. Then there is the point of human lack of determinism. You cannot make any character depend on 3-frame-windows to survive if any other has a lot more time to react. On the other hand you cannot consider any broken feature to be ‘impossible to use’ having only 3 frames to be launched, people might adjust and actually use it to break balance (they already did). But still math already helps here to find possible holes in your balance (calculating ‘frame advantage’ oder input-windows). It also helps to find other possible balance holes without even having to test the game, like seing any 100-ressource-unit beating one 1000-ressource-unit should make you think about if there’s any counter to this. So the point is while math can point out weaknesses without any testing it’s not magic. If you have to choose between one shot of 100 damage or 100 shots of 1 point damage math won’t help. You’ll have to pick one of them and then also factor in the probability of a player landing that one powerful hit versus the percentage of the 100 low-damage-shots hitting which again you can only guess (so it’s not only damage/time). Obviously if then any variable depends on 100 things plus player perception/reaction-time/whatever (which might even depend on animation) mathematical considerations will only be able to give you some rough hint on how to set something.
And I think symmetrical games are often just as hard to ‘balance’ as asymmetrical, you only fall softer if you fail. If you have some useless weapons in any shooter players avoid them and the game becomes rather one-sided and maybe gets ‘linear-slope’-effects but people still get to use those weak weapons if they have to. If you have any fighter where one character is a lot more powerful than all the others players will instantly drop it, even if that mirror-match could be interesting by itself (which is quite the same as what any FPS could degenerate into).
March 20th, 2008 at 1:59 am
I think this whole “math” thing depends strongly on what you consider math, and what your model is. From his descriptions, Sirlin seems to do a lot of analyzing what states the game can go to from what other states, and in particular what loops exist and how they behave. I’d call that math.
March 20th, 2008 at 4:28 am
Let’s be real, Sirlin, you love fighting games and balancing them is not only fun for you, it’s your dharma. You’re really good at it.
Do you really want to work on a symmetric fighting game?
March 20th, 2008 at 5:04 am
Sniffnoy, I have too much say about that. I’ll have to write a book about it. ;)
Santoki: I don’t know now much more clear I can make it than what I said in the original post: “And yet…that’s all I do!! All FIVE of the games I listed at the start of this article are asymmetric. And 100% of competitive games I’ve ever planned to do and still plan to do are also asymmetric.”
At least you had a nice use of the word dharma though. :)
March 20th, 2008 at 6:51 am
The definion of symmetric that I’m familiar with means that a two-player game (or a particular situation) is symmetric if players wouldn’t be able to tell the difference if they exchanged positions at the start of the game. So, if the players in chess are black and white, then it’s not a symmetric game. If we include a coin flip to see who plays white at the start the game, then it is. This definitition of symmetric is useful because it guarantees that the game is fair (that is, that neither player has an advantage.)
March 20th, 2008 at 7:11 am
NateTG: rather than a super nitpicky definition, I’d rather have a more useful one. For any practical purpose, Chess is symmetric. If we don’t classify it as that, then the terms themselves stop being very useful. Compared to Street Fighter, it’s pretty damned symmetric.
March 20th, 2008 at 1:54 pm
I’d just like to point out that Defense of the Ancients is a fantastic game that layers deeply asymmetrical player units over a roughly symmetrical NPC grid. It has many of the features and challenges of a fighting game vis a vis character balance, while adding the challenge/interest of tracking inter-hero interactions to their balance.
Sirlin if you’re not familiar with the mod I strongly suggest checking it out.
March 20th, 2008 at 5:00 pm
heh, sorry Sirlin. I rarely read what you write. You’re just so wordy!
March 20th, 2008 at 7:40 pm
Sniffnoy: Doesn’t seem as obvious as I thought. So after my description of why math fails to deliver you actual variables to work with (referencing comment #3 by Sirlin here) on a small scale you suggest to go up one level and hope they show up after you analyze the rough structure of the game? I know math is a lot more than putting numbers in a formula, the point I tried to show was why math fails to deliver equivalent and diverse variables even on a smaller scale and why mathematical considerations can still point out balance holes (even if they don’t present the variables you should use to fix them).
March 20th, 2008 at 8:42 pm
Sirlin:
I’m not sure that that notion of symmetric makes sense. Consider, for example, “Straw Man Chess”, where in addition to all of the usual moves, a player can make the “I win” move. Both sides have access to the same moves, but I can’t really see it as a symmetric game.
Considering that both players initially have character selection on the same clock, the versions of Street Fighter that allow mirror matches are symmetric while individual character match-ups like Ryu vs. Zangief are not.
March 21st, 2008 at 1:00 am
NateTG: I’ll have to disagree again. The terms really do make sense and are helpful. Straw Man Chess isn’t even a game because player 1 wins 100% of the time. Regular Chess is symmetric (or to be nitpicky, it’s soooooooooo close to symmetric, that it’s helpful to classify it as such). One could call Street Fighter symmetric saying everyone has equal access to the character select screen, but that’s more of a lawyer’s trick than anything meaningful. In practical terms, you are locked into a set of completely different moves from your opponent in Street Fighter. I’m trying to call that phenomenon something. The phenomenon does not happen in Chess, it does happen in Starcraft, and we’re calling it “asymmetric.”
Matt: Yeah…we’re familiar with Dota. The thread about Most Balanced Games goes on and on about it.
March 22nd, 2008 at 12:53 pm
Santoki said:
——————–
heh, sorry Sirlin. I rarely read what you write. You’re just so wordy!
——————–
Ha!
If Sirlin is wordy, Santoki, don’t ever visit <i>my</i> site! ;)
… not that you can, since it’s still in development, but I digress.)
- Bruce
March 24th, 2008 at 2:38 pm
MVC2 is an interesting case. I think removing the bottom half of the roster does make it more balanced because more of the possible moves are viable. MVC2 is only balanced because people know better than to pick 90% of the cast. In an ideal game the entire cast is viable, at least as counter-picks.
A good way to look at balance is assigning the sides randomly. (At least in a two player game) In Starcraft that random assignment works well, all permutations are fairly balanced. In MVC2 it would work horribly. That’s a very strict definition of balance.
You could broaden it to include counter-picks, but counter-picks only work when you know what your opponent is picking, it doesn’t factor in at all with blind selection. Counter-picking only makes characters viable because of the way tournaments are run. (With winner stays on picks)
March 25th, 2008 at 12:55 am
Sirlin Said:
I’ll have to disagree again. The terms really do make sense and are helpful. Straw Man Chess isn’t even a game because player 1 wins 100% of the time. Regular Chess is symmetric (or to be nitpicky, it’s soooooooooo close to symmetric, that it’s helpful to classify it as such). One could call Street Fighter symmetric saying everyone has equal access to the character select screen, but that’s more of a lawyer’s trick than anything meaningful. In practical terms, you are locked into a set of completely different moves from your opponent in Street Fighter. I’m trying to call that phenomenon something. The phenomenon does not happen in Chess, it does happen in Starcraft, and we’re calling it “asymmetric.”
I’m trying to understand what you mean when you write symmetric.
It seems like you’re making a distinction between games that can easily be broken into asymmetric subgames - that is, games with initial character or team selection - , and games that cannot - so, providing everything except for team selection is fixed, StarCraft is really a combination of 3 asymmetric and 3 mirror matches, and SSFIIT a combination of 16 mirror matches, and 120 asymmetric ones.
By comparison, although the effects of openings in high-level chess last the entire game, it’s not necessarily obvious what the initial moves were when someone looks at a particular position.
Do you consider Warcraft II (the choice of Humans or Orcs) to be symmetric?
March 25th, 2008 at 2:44 am
Nate: This is buried in the “most balanced games” comments somewhere, but basically we’ve agreed to ignore mirror matches. The point of looking at asymetric games when looking at balance is to see how two fundamentally different sets of moves interact. Just about every game we’ve called asymmetric allows mirror matches, but they’re boring in a strategic sense. You no longer have to worry about the balance, because, same as chess, both players have all the same pieces to work with, so it’s impossible for true imbalance to exist, just because the two players could do exactly the same thing. (Note: this also ignores advantages like who goes first in chess, and odd things like bugs that only the second player can perform in fighting games, but again, it’s just outside the scope of what we’re actually interested in talking about at the moment)
March 25th, 2008 at 2:53 am
It’s not that hard of a definition… in an asymmetric game, you pick a character or race or something to that effect before the match. In a symmetric game, you do no such thing.
March 25th, 2008 at 2:59 am
By the way Claytus, good point. I nominate Company of Heroes given your explanation, but I still wouldn’t consider Counter-Strike for this list for reasons I won’t get into.
March 25th, 2008 at 6:20 am
PoisonDagger: I think it’s only a hard definition because it’s a weak definition. You, Sirlin, most of the people on this thread, and I have an intuitive sense of what it means, but there may be some people who need a more absolute line between symmetric and asymmetric.
March 26th, 2008 at 5:51 am
Warcraft 2: Clearly asymmetric by Sirlin’s definition, as each player’s decision to pick “orc” or “human” affects what spells they have (orcs have bloodlust and IIRC raise dead, while humans have healing).
“It seems like you’re making a distinction between games that can easily be broken into asymmetric subgames - that is, games with initial character or team selection - , and games that cannot - so, providing everything except for team selection is fixed, StarCraft is really a combination of 3 asymmetric and 3 mirror matches, and SSFIIT a combination of 16 mirror matches, and 120 asymmetric ones.”
I assume by Sirlin’s definition of symmetry, any game in which asymmetric matchups exist is asymmetric. Otherwise any game that allows mirror matchups at all would be symmetric and Sirlin’s clearly looking for a useful definition
March 27th, 2008 at 3:29 am
You make a good point about the term being useless by my definition, because all games would be asymmetric. Claytus also makes an interesting point about the diversity of TF2 being “smoothed away” because of team play. I hadn’t really thought of the smoothing effect of team play before, but it immediately struck a chord with me — I think he is right.
But I’m still a little wary of the idea that the symmetric/asymmetric term need to be applied to a game as a whole, as some giant binary flag on the entire game.
I’m not a professional game designer so I have no say about how useful a tool this giant binary flag is to designers. But I do feel that the actual nature of games is more moment-by-moment.
In fact a game really is a whole giant universe of possible scenarios — basically a “problem space” — and each of those scenarios can be examined for asymmetry independently of all others.
So even when taking the macro view, don’t ask “is the game symmetric or asymmetric?” But rather, “How often is this game symmetric?” Because it really stems from the question “How asymmetric is this sort of encounter within the giant universe that is my game?”
Because TF2, for all of its faults, is NOT symmetric. It’s a very good point that the team play smooths out the asymmetry at a macro level, but when you zoom into the moment-by-moment tactical encounters, the actual EXPERIENCE of playing, there are a lot of good asymmetric matchups in that game.
So while my term may be broken in one way, I think the way you guys are using it might be broken in another.
March 27th, 2008 at 4:40 am
Well, remember the defintion we’re using here isn’t what designers use while creating a new game. The point of this exercise isn’t to do more designing, but to classify games that have already been made. We’re declaring it to be a binary flag so that we can throw all the “asymmetric” games in a bucket for now, and then probably look at them more carefully later. I don’t even know what that later step is really for… you’d probably have to ask Sirlin, maybe he wants to write a book about them or something.
As I said before, all your points are valid and interesting… they’re just a different defintion of the term, it’s not that either definition is “broken”… it’s just slightly ambiguous usage of a single word, since there aren’t any English words that precisely indicate the concepts we’re discussing.
March 27th, 2008 at 4:44 am
I agree that there are plenty of asymmetrical yet interesting matchups in TF2, and that’s why Sirlin could learn at least something from it… but he’s looking for games where you choose a character/race/whatever and are stuck with that choice for the whole match. In TF2, you aren’t stuck with your choice for the whole match.
The moment-to-moment asymmetry idea exists in every game… it’s kind of pointless to bring that up when we’re talking about something completely different.
March 30th, 2008 at 3:23 pm
This is getting very legalistic, unnecessarily so. Look, heat is a variable with infinite gradations. Something can be zero degrees or a hundred degrees, and something a hundred degrees is 100* hotter.
That doesn’t mean that I can’t divide a menu into “hot meals” or “cold meals”, even though the definition is nuanced and involves some intuition. I could also measure the exact temperature of each food item. This is missing the forest for the trees though.
When someone says “asymmetric game” assume they mean “possessing enough asymmetry as to be distinctly so”. Not to be a jerk, but this is basic linguistic communication. Now, you can have an interesting argument about stuff like “how asymmetric is chess, given white’s starting advantage”, but you do the whole debate a disservice by coming at it is a question of semantics and definitions.
Bleating about semantics is the death of all reason.
April 1st, 2008 at 3:53 am
As much as bleating about semantics may be the death of reason, there’s no communication if people are talking past each other. While Sirlin considers chess symmetric and Street Fighter to be asymetric, I consider chess asymmetric, and Street Fighter to be symmetric. Thus, when Sirlin writes, ’symmetric’ he must mean something different than what I understand the term to mean. (It seems very strange to me that Orcs vs Humans in WCII would be considered asymmetric when the difference between the two is probably smaller than the between black and white in chess which is considered symmetric.)
April 1st, 2008 at 4:19 am
I don’t understand what the big deal is with black vs. white in chess. Yes, the tempo loss of going second is a disadvantage, but it’s not really that significant. We’re talking small percentage differences in terms of win rates between the two sides… someone who’s better at the game can win on either side. It mostly matters during the opening anyway, and to memorize a sequence of opening positions, you need to know both black and white’s moves, anyway.
Regardless, my real point, is I think we should all talk about Go as the “base” symmetric game, rather than chess, which should put this all to rest. Go gives points to the white player in exchange for having to go second, so it removes the tempo disadvantage. Just another idea to drag us back out of the semantic mud.
April 1st, 2008 at 4:24 am
He means that you pick a character or race or something that you’re stuck with for the whole game.
April 1st, 2008 at 4:25 am
Oh, also, NateTG: We’re talking about ST games from the start of a match, not including the timed character select. So, in fact, ST would still be asymetric by your definition (except during a mirror match). It’s not like it’s unprecedented… many tournament require you to declare your character before you ever get near a controller. Either japanese style where you only get to use one character, or in a tournament where initial character selection is double blind.
April 1st, 2008 at 7:45 am
“(It seems very strange to me that Orcs vs Humans in WCII would be considered asymmetric when the difference between the two is probably smaller than the between black and white in chess which is considered symmetric.)”
Do you understand the difference between “both players start with identical moves that they retain throughout the game” and “player select a collection of moves that cannot be changed once the game starts”? This is basically what is meant by asymmetric vs. symmetric in this discussion. If you think that distinction should be called something else, what do you propose we call it?
Brian has the right of it.
April 1st, 2008 at 11:36 pm
I want to comment on two things.
First, Claytus, I think I may have misunderstood your point earlier.
I reread this part of your comment:
“The term is being defined as a description of game design, so it’s a feature about how a game is made, not of how a game is played. Chess is symmetric, always, by definition.”
So now I think I get it. You’re not talking about symmetry of the experience of play. You’re speaking here about the methodology of design. We’re not talking about “asymmetric games,” but rather “asymetrically created games.” If that’s so, then I suppose I went off on a tangent above.
Symmetrical games and symmetrical design are two totally different topics.
As a player I guess I’m just more interested in thinking about the nature of a game’s experience, rather than the mechanisms the designer used to create it.
Second, I want to comment again on the continuing usage of chess as the archetypical symmetric game.
I think chess is, as you say, symmetrically *designed*, and if that’s all we’re talking about that’s fine. But I think people overestimate the symmetry of the play experience.
As an example, take these descriptions of top-level openings. Notice that right away the descriptions get into things like White is doing such and such, while Black on the other hand works towards these ends:
http://en.wikipedia.org/wiki/French_defense
http://en.wikipedia.org/wiki/Sicilian_defense
The difference between black and white here is far more than a single move. Each decision shunts the game down a different path where the players have very different options and goals from each other.
In many ways, a game of chess doesn’t even begin until the memorized opening decisions are complete. You could think of the first 5-10 (or 30 in some cases) moves as the character selection screen, because the players are choosing between several well-understood positions whose characteristics they have both studied in great detail.
Now take Chad’s definition of symmetry and asymmetry: “both players start with identical moves that they retain throughout the game” and “player select a collection of moves that cannot be changed once the game starts”
Which one applies to chess? There is an illusion of the first one, but I claim the reality is much closer to the second.
April 2nd, 2008 at 2:36 am
It’s all too theoretical in my opinion. You can say that chess doesn’t “really start” until 30 moves in… but everyone knows a checkmate is technically possible from the 3rd turn on. That particular mate is pretty stupid, and once someone’s seen it once, the game will never unfold quite that way again, and yet it’s proof that the game can end before 30 turns have gone by, so you can’t say it didn’t start till that point. All the memorized openings are great, but they aren’t required… I’ve done lots of wacky stuff you won’t find in any opening book. Against someone of average skill, there’s about an even chance that either they’ll realize *why* the opening in the book is superior, and punish you, or they’ll just get confused, and make mistakes themselves.
There is not an “illusion” of the first concept you mentioned. Everyone has the same goals (checkmate, take pieces, control the center, etc…), and they have the same moves available to allow them to perform that task (pawn, knight, rook, etc…).
Strategic differences are a difference playstyle, and the results of superior vs. inferior play. However, they’re not inherent to the rules of the game. Black can literally mirror whites moves the entire game, and it’s completely valid, and it’s an identical strategy used by both players. Now, white will win (because of that tempo advantage that keeps getting mentioned), so black shouldn’t ever actually do this if he wants to win. However, it’s a completely valid way to play, and shows that, in fact, the two sides both have exactly the same options available.
April 2nd, 2008 at 3:10 am
The difference here is that a designer didn’t have to go through all the possible Chess openings and make sure all the matchups between them are fair. Both players have the exact same set of pieces arranged in the exact same way, and the advantage of first turn is insignificant (1/2 a pawn as someone mentioned), so the game is inherently fair. Just like Ryu vs Ryu, or Protoss vs Protoss. Sure, you can seriously dictate the flow of the game within the first several moves… but that happens in any game regardless, and isn’t helpful for studying how to make two vastly different movesets fair against each other.
April 3rd, 2008 at 12:42 pm
Whether chess is fair is an open question.
Even if you don’t like the example of ’straw man chess’, there’s nothing magical in a ’symmetric’ game that ensures that the game is going to be fair or balanced. In fact, for correct play, deterministic complete information games are unterminating, a guaranteed draw, or unfair.
April 3rd, 2008 at 12:54 pm
Well, there actually something magical in a symmetric game that ensures it’s fair. However, you are correct to point out that Chess does have a tiny element of asymetry in the tempo advantage white gets, so it’s not 100% guaranteed that chess will be fair itself. However, all evidence at this point seems to indicate that when chess is eventually solved, it will end in a guaranteed draw. The tempo advantage is incredibly minor, and statistically it doesn’t seem to result in white getting that many more wins than black. So, it’s basically a moot point.
It’s still too nitpicky anyway, I think you’d agree that chess is mostly symmetric, though not completely symmetric. And all mostly symmetric games, we’re calling symmetric, and all mostly asymetric games we’re calling asymetric. This really isn’t that hard.
April 3rd, 2008 at 12:59 pm
If you wanted to learn about how to make different sets of moves (chosen before the match starts) fair against each other, you would learn zero from chess. Or if you want to be nitpicky, then you’d learn very close to zero. You’d certainly learn infinitely more from looking at StarCraft, Street Fighter, MTG constructed decks, and even Puzzle Fighter.
I’m sad that most of this discussion is about which category to put things in, when it’s really pretty clear.
April 4th, 2008 at 8:09 am
Well, I can stop ruining the discussion, but I disagree on all counts.
Why can’t you examine endgame positions of bishop vs knight; rook vs bishop + 2 extra pawns; knight vs 3 extra pawns; etc, in order to glean some insight into asymmetric abilities that are balanced? Bishop vs Knight is a powerful example of variety + balance. Why is there zero to learn there? Just because 2.3 hours ago we each had two of each? That’s silly.
If you set out to design “asymmetric chess” where you could choose diff pieces than your opponent before starting, the above is exactly what you would study. You would learn more than zero from chess in order to design your asymmetric game.
You can study the asymmetric MOMENTS of games, and then apply them those lessons to design a game in either of these catagories, however you choose to define them.
But I seem to be getting stubbornly out on a limb alone here, which means I’m not adding much to the discussion. Didn’t set out to drive it into the dirt, I just honestly thought you guys were erroneously throwing out 90% of games.
April 4th, 2008 at 8:25 am
I think you’d learn very little from studying those positions in Chess. If it turns out that rook vs. bishop + 2 extra pawns favors rook, then so what? Good job for getting into that advantage situation. Either player can end up that way in this symmetric game, so it doesn’t matter if those particular arrangements are fair versus each other. They don’t HAVE to be fair.
What does matter is that white’s entire side is fair versus black’s entire side. But they have all the same pieces, so there isn’t much to study. Sure you could study why the set of pieces in Chess leads to a deep game, but that’s a different question. This is similar to how 4 Zerglings don’t have to be balanced versus 2 Zealots, and it would be pointless to study that. But Zerg does have to be balanced versus Protoss.
And if you want to study middle sections of a symmetric game that show moments of asymmetry so you can harness those into creating some different, actual asymmetric game…well maybe you cold do that. But how about this? Let’s start with the far more direct method of looking at asymmetric games to begin with. Dozens and dozens of post here and in the forums have managed to sidestep this, yet it’s the entire question.
Honestly, I don’t know if it matters anymore. It looks like StarCraft (and maybe other rts stuff, possibly Dota), Street Fighter (and a couple other fighting games), and MTG (and a couple other card games) are mostly it. Everything else seemed to be shot down or not given enough votes.
April 4th, 2008 at 8:41 am
Yeah, you could examine some contrived Chess endgames. Or you could examine games that are, through countless reams of variables, specifically designed to have a set of matchups that are all mostly fair. ST has 120 non-mirror matches if you don’t count the Old characters.
What really makes more sense to analyze?
April 8th, 2008 at 8:21 am
It should be noted, just to throw a wrench into things, that <i>Chess</i> isn’t perfectly symmetric because White gets to move before Black. This asymmetry is minor in <i>Chess</i> but considerable in <i>Go</i> (and accounted for through komi). (EDIT: I see Sirlin has already pointed this out.)
Other games featuring asymmetry:
- Arguably <i>Settles of Catan</i>. It is definitely asymmetric if you use the starter rules, and I’d argue that even if you use the full rules for set-up you’re still dealing with an asymmetric game.
- Reiner Knizia has developed many asymmetric games, including <i>Hera and Zeus</i> (a two-player game featuring gods with different sets of abilities on both sides). His design for the <i>Lord of the Rings</i> boardgame is non-competitive, but also asymmetric in a minor way (with each player possessed of a different unique ability). Since the game is scored, this is arguably important.
- However, once you add the <i>Sauron</i> expansion to the game — with one of the players taking on the competitive role of Sauron with a completely different set of rules governing their behavior — the asymmetry is clear and competitive.
- Any tabletop roleplaying game features asymmetric design. To take D&D as the most obvious example, you not only have the need to make sure that every character class can contribute to the game (so that one player doesn’t feel left out), you also have the asymmetrical balance necessary whenever the DM designs a challenge for the PCs to overcome.
(Thinking of the relationship as “DM vs. players” is frequently frowned upon because an adversarial DM — with their limitless abilities — can wipe out the players if they choose to. But that misses the point: The DM is, in a very real sense, constantly performing the tasks of a game designer. They are constantly being faced with an asymmetrical design challenge.)
April 8th, 2008 at 8:48 am
Chess being slightly asymmetric is noted about 50 times in this thread:
http://www.sirlin.net/archive/the-most-balanced-games/#comments
And it’s still not a very useful distinction. It’s so close to symmetric that it’s helpful to put it in the symmetric category. Reiner Knizia’s games are a good mention though. I really should have thought of several of those. I personally find his Lord of the Rings game to be pretty terrible and lost all interest after playing once, but yes he does try for asymmetry. I think he’s not even in the ballpark of asymmetry that is balanced under heavy competitive play though. Correct me if I’m wrong.
Actually, I guess I’m mixing threads. Even if his games aren’t the kind that hold up to hardcore tournaments, they still are an example of a designer trying for asymmetry and an example of fans really liking that. My other comments would fit better in the “Most Balanced Games” thread.
April 8th, 2008 at 11:55 pm
OK I get it now; you want a totally closed system that is not just asymmetric but also already fair. Moments of any game can be asymmetric, but they offer no guarantee of being fair so they’re no good.
You need something that starts off asymmetric because only at the start do we know the game is still fair.
I guess you guys are right — what you’re really looking for are games with a character selection screen that is not team-based and does not allow you to revisit your character selection until the unit of play is over.
I haven’t got any awesome examples, which is why I was looking at mining the moments from other types of games… but I can see how that’s not helpful if you need fairness.
The only decent example I can think of now is car combat, such as the Twisted Metal series.
I suppose you could broaden that and say racing games with sufficiently varied character/vehicle selections, like maybe MarioKart, although I hardly know that game.
Also some games allow you to construct your team in advance, rather than picking from a preset list of static choices, for example NES Ice Hockey, or Warhammer tabletop gaming.
Tribes 2 duel mod would also qualify, although again you’re selecting numerous things from a menu of choices rather than 1 from a set of pre-defined package deals.
April 9th, 2008 at 12:02 am
ricefrog, being able to assemble a team is just the same as choosing a single character. I believe that Sirlin considers Magic the Gathering to be one of the games on this list.
April 9th, 2008 at 7:20 am
Yay, 2 points for ricefrog in general.
Mario Kart probably qualifies, though maybe the differences in characters aren’t *that* huge, but yeah it qualifies. I don’t know if it’s balanced in high-level play or not though. Same goes for Mario Tennis, but is that game really a paragon of balance? No one plays it hardcore enough to know.
Choosing your team in NES Ice Hockey is pretty limited but yeah it’s the right idea. A more sophisticated similar example is choosing your team of 3 characters + 3 assist-types in MvC2, or your deck of 3 characters + 3 items in Kongregate.com’s Kongai. ;) And yeah an even more sophisticated example is choosing your 60 card constructed MTG deck.
April 9th, 2008 at 9:34 am
I think that one legitimate gray area in the “asymmetric vs symmetric” debate would be Counter-Strike. I’m curious as to where you’d put it (in terms of weapon choices, anyway).
Asymmetric: You can only hold one primary and one secondary weapon, so those are your main choices at the beginning of each round. Once gameplay starts, you’re stuck with your choice for the rest of the round.
Symmetric: You can buy many weapons and drop them on the ground for later use (although I’ve only seen one very specific instance of this being done, and you won’t see players doing this regularly since it’s so damn expensive to do). You can also drop your weapon for one that any dead player has dropped.
I see this as being fairly ambiguous, but if I had to choose, I’d say that it tips just slightly over to the asymmetric side - at least enough to put it in that category. The reason is that if the weapons/weapon combinations hard countered each other RPS style, that would degenerate the gameplay quite a lot - just like in a fighting game. If a weapon/weapon combination is easily countered, it wouldn’t ever be purchased.
However, CS’s weapons don’t have interesting variety in and of themselves in high level play (the game is interesting for other reasons), so I wouldn’t actually nominate it for Sirlin’s list.
April 10th, 2008 at 5:55 am
Well, CS could work if it was deathmatch free-for-all, but as Sirlin was pointing out earlier, in a team setting each available option doesn’t have to be fair on its own, it just has to be useful in some larger overall team configuration. This is why TF2 and Tribes (which have way more variety than CS) also fail.
I was trying to think of some deathmatch FFA examples that qualify, and I had a hard time, but I talked to a couple of friends and poked around a bit and I came up with the following three examples. I’ve actually never even played any of these games, but they SEEM like they might be interesting, and maybe others here can comment:
SubSpace (http://en.wikipedia.org/wiki/SubSpace_%28computer_game%29)
Soldat (http://en.wikipedia.org/wiki/Soldat)
Star Control (http://en.wikipedia.org/wiki/Star_Control)
These games are not all dedicated FFA; two of them look like they offer a lot of team modes as well, and soldat even offers you a chance to switch your initial choice by swapping weapons with a slain enemy. But the design goal of variety+balance appears to be at play in all of them.
April 10th, 2008 at 6:38 am
Well with my hypothetical CS weapon balance situation, weapons would be more like cards in a Magic deck… there would have to be a good variety of “decks” that are all clearly different yet equally viable.
Regardless, my question was more about the game’s classification, not about if its weapons are balanced - there’s only a few weapons on each side (including sidearms) that are actually used in high level play, so I already know the weapons aren’t balanced.
April 10th, 2008 at 9:15 am
Max (Halo 2’s multiplayer lead) said that whenever he did anything the least bit asymmetric, he made it so each team plays a different side next round (or after x rounds, whatever). This meant he didn’t have to REALLY make it exactly fair. He was instantly willing to call everything in Halo symmetric for this reason, and he included Counter-Strike in all that because we’re assuming you play X rounds as one side and X as the other. I haven’t actually been involved in any tournaments for CS though. Is that how they are really run?
ricefrog, interesting. I’m (barely) aware of all those games but I haven’t actually played any of them.
April 10th, 2008 at 9:59 am
Yeah, that’s how CS is played. Because of that, I’ve never considered CS asymmetric when it comes to teams (it’s not a choice - you play both sides).
But what about when it comes to weapon choices? Would you consider that aspect asymmetric?
April 11th, 2008 at 1:00 am
I wonder where people put Stratego on the list. (In Stratego, each player has to arrange the starting positions of his pieces - which limits possibilities as the game goes on.)
April 11th, 2008 at 1:17 am
PoisonDagger: The weapon choices aren’t *that* varied and if high level play doesn’t even see a lot of them used, then it doesn’t seem like a very good example, even if we did call it asymmetric.
NateTG: symmetric. What if lots of pieces in Stratego are weak? Doesn’t matter, both players have equal access to them, so nothing actually has to be balanced. Sure there is the idea that each piece can be used because of the rank (and wrapping around of lowest to highest), but the designer doesn’t even need to do any balancing beyond that. If some certain starting configuration is not good, then players should learn not to use it, as they learn to avoid any other bad strategy in any other game. It’s kind of an interesting example though because a super-nitpicky person could try to call it asymmetric, even though that wouldn’t help me at all.
April 11th, 2008 at 2:52 am
As far as I can tell, symmetric games still need to be balanced. If every situation has an obvious correct choice, the game is (strategically) boring like Tic-Tac-Toe. For example, if Stratego didn’t have have mines (or the spy), the game would, basically, be a race between the field Marshals to work through the opponent’s board looking for the flag (or each other).
April 11th, 2008 at 3:11 am
Yes, but that’s a question of adding depth to a symmetric game, not a question of balancing different races/characters versus each other.
If we were making a symmetric game and wanted to know how to add depth, maybe Stratego would teach us things about adding mines and spies. But to balance Puzzle Fighter characters versus each other, or SF characters, or Starcraft races, or Mario karts, I don’t think Stratego offers any insight.
April 11th, 2008 at 4:35 am
It doesn’t matter whether the sides have the same moves or not. If one player has an obviously correct strategy, then the opponent does as well, and the game is unbalanced. In SSFIIT there are unfair, but relatively balanced match-ups where characters like Boxer dominate using mix-ups, and unbalanced match-ups where characters like E Honda dominate with a zero risk strategy.
Of course, my notion of balance is that in a balanced match-up there is no dominating tactic.
April 11th, 2008 at 4:54 am
Oh, subspace, soldat and Star Control. I can speak on each of these actually.
Subspace is very server based and heavily modded. To say one plays ’subspace’ doesn’t mean much, as each server has it’s own setups, radically different from the next. I play (occasionally) on the Trenchwars server, which features great variety in their ship designs. Unlike most other subspace servers, Trenchwars keeps certain abilities exclusive to certain ships. One shit killing Warbirds, base breaking Javlins with rebounding trick shots, the machinegun spider, the terrir, which is is a fast, featureful vehicle that can be teleported to but lacks strong offense but is crucial to holding base — and the shark, which is both the mine layer and anti mine vehicle. Theres also the lumbering, super splash damage missile launching Levithan and the cloaking weasel — but they don’t get used in tournament play. Oh yeah and this shotgun thing thats kinda scrubbish.
While theres a lot of great design elements (and many problems, but the game works well competitively), it falls into TF2 balance. Indvidual matchups do not need to be balanced and that a team has access to all the ships. Interesting and fun, but doesn’t qualify.
Soldat has differences in guns — and you can only have one at a time (until you die). It’s assymetrical, but only slightly. Each death you can simply pick another weapon — most of which aren’t so radically different from the other weapons.
Star Control is just gdlk. We’re talking a Guilty Gear level of variety here. I really believe this game was way before it’s time. One ship can teleport at will and move uneffected by gravity and firing a homing laser. One is slow moving, firing high power sniper shots and can burn ot’s own crew for energy. One has a rotating turret and deploys space marines which break into ships and kill their crew. Another fires huge buzz saws into space the act as area of denial… and theirs still a ton more ships. To balance stronger and weaker ships, each recieve a point value (which was designed into the game — not as an added on rule), and one must construct a force at or under that point cost. It works a lot like magic in fleet construction. One must be sure to build a solid fleet with few weaknesses. Individual ship matchups are very STish — even counter matches are winnable with only a few blowouts. I think you might find it worthwhile to look into it more.
April 11th, 2008 at 5:03 am
For clarifications sake I was speaking of SC2, though I don’t know how different SC1 is. Anyways, I think this is one of the earliest massively assyemtrical game that I know of, having preceeded Street Fighter II. I really really really think it would be worth your time to look into it more.
Theres a version released for windows now, called The Ur-Quan Masters. http://sc2.sourceforge.net/
Again, I think this will be very very worth your time. I’m ashamed I didn’t bring it up first, but I completely forgot it existed.
April 11th, 2008 at 5:06 am
KayiN, great info, thanks.
NateTG: We’re talking about different things. It’s confusing because there is no standard terminology. For now, let’s call “balance” the thing where asymmetric sets of things chosen before the game starts have to be fair against each other. Let’s call “depth” the idea that there are interesting choices to make throughout the game. So directly following from the definitions, it doesn’t even make sense to ask if Stratego is balanced (because it’s symmetric) but mines and spies do add depth.
If you don’t like those terms, we could call them Q and Z or whatever else instead of “balance” and “depth.” In any case, Stratego has none of what I meant to be asking for. Yes it has the thing you are talking about though. If we cared about “depth” as I defined it, then we could start talking about Chess, Quake, Stratego, and any number of other (mostly) symmetric games, but I meant to be asking about the “balance” between asymmetric sets.
It seems like a lot of people are feeling left out for some reason, but this is like me asking for a good way to make blueberry ice cream and then people who like chocolate ice cream feel left out. They shouldn’t because later we can talk about chocolate ice cream.
April 11th, 2008 at 6:40 am
Yeah, Subspace is awesome… I started to get into that game back in the day just as all my friends were quitting, so my knowledge is somewhat limited, but the stories they told me were just great. I had also completely forgotten about it until just now.
Regarding Star Control 2, also a great game… I had the original version on PC when I was a kid, and would play against the computer all day. This was all pre-internet dark ages when multiplayer wasn’t a possiblity, but it was still a great game, and I’d second everything KayinN said. Also, I think I did read up on Star Control 1 once, just for the heck of it… I don’t remember any specifics, but it struck me as soooo different that it’s not entirely clear why they felt the need to make SC2 a sequel, rather than just a seperate game.
April 11th, 2008 at 6:59 am
That last paragraph you wrote was something I wanted to say during this whole conversation, but couldn’t quite find the right explanation for. That analogy is awesome. And delicious…
Anyway, my whole question about Counter-Strike may be irrelevant… but what about Call of Duty 2’s Search and Destroy mode (aka CS WWII)? The game’s high level setup is remarkably similar to CS, except, IMO, better. Players choose their weapon class between rounds without any money restrictions. Instead, different main weapons come with different amounts of smoke grenades. Each side gets a mostly similar array of weapons, but there’s enough variety and balance among the weapons to provide enough overt and subtle choices (for a shooter that has additional complexities in the maps themselves, anyway). You mainly have sniper rifles, bolt action rifles, semi-auto rifles, assault rifles, submachine guns, and the (infamous, at least to scrubs) shotgun. Before the round starts, your team will want to figure out an effective set of weapons to use, taking the opponents’ previous positions into account as well as their own players’ effective playstyles. For example, for going longrange you have a few choices - you can go sniper if you like the extra zoom, you can go single-shot rifle if you like the extra awareness/mobility, or you can go semi-auto rifle if you like to expose yourself for less time at the cost of accuracy (you can afford to peek for less time when you don’t have to wait so long between shots). When using a shortrange weapon, you get more smoke grenades to help you move up close to longer range players, plus you move faster.
IMO, CoD2 had a lot of things going for it that made it potentially better than CS (my favorite part was the focus on single shot weapons, which rewards good prediction and quick peeking more than accurately tracking a moving opponent with the mouse, and single-button grenades and single-button, insta-kill melee attacks were awesome too). Unfortunately, the whole WWII setting turned many new players off, especially since CoD4 is out with its fancy “Modern Warefare” and “automatic weapons” and “grenades that can be cooked to guarantee kills instead of simply flushing opponents out of hiding”.
Anyway, CoD2 might be worth looking at if you consider it asymmetric (remember, you can swap your weapon with one from a dead player - that throws a wrench into the asymmetry definition!). I’d argue that it’s asymmetric enough to consider it so.
April 11th, 2008 at 7:00 am
My first paragraph there was about Sirlin’s post, not Claytus’s… although the word “delicious” is probably a good tipoff.
April 11th, 2008 at 11:37 am
Sirlin: So the sort of balance that you’re looking for might be achievable by adding a head-to-head handicapping table?
April 12th, 2008 at 1:12 am
NateTG, well sort of, but I wouldn’t actually do that. If you have such a table, that means you even KNOW in the first place exactly how good character A is against character B. Knowing that at all is at least half the problem. For example, ok here’s Starcraft 2. What handicap do you want to give Zerg vs Protoss? You better know a hell of a lot about the game to attempt that.
Second, if you somehow did know how much handicap you wanted to assign a certain match, then actually assigning a handicap is a cheap way out. I have never even considered using that solution in anything I’ve done so far. That’s the lazy thing to do in a fighting game, and it’s artificial so we should find better solutions. If character A loses to B, then find a way to give A something that helps against B that is mostly irrelevant in other matches.
So yeah, I’m not in favor of handicap tables. But I guess it is technically correct that I’m looking for the kind of balance that might be achievable by adding handicap tables…I just wouldn’t actually add them, ha.
April 12th, 2008 at 1:26 am
The problem with handicap tables is that it doesn’t necessarily make the game more fun. If character A has an easily abusable move that only works against character B, then how does a handicap help? If you make B more powerful than A, then for B the game is about trying to avoid that abusable move, and trying to simply get a few hits against A to win the round. The game for A is about using that abusable move as effectively as possible and trying to avoid those few hits that will knock him out.
Even if in the long run the matchup is balanced, the gameplay is pretty degenerate and uninteresting.
April 12th, 2008 at 2:52 am
Sirlin:
I’m not really a huge fan of per-match handicapping either, but your answer clarifies what you mean by balance…
Getting a bit mathy, assuming, for the moment, that such a table could be generated in practice, it basically represents n choose 2 equalities (one for each character match-up), so if it’s possible to find n choose 2 variables that can be tweaked for approximately linear and independent response in the game, it should be possible to implement the equivalent of such a table by modifying those variables - rather than by using an explicit handicap table.
PoisonDagger: I won’t pretend to be a competition-level player of anything, but, as long as damage handicapping can make the game fair, it’s possible that a game like that can be interesting. AFAIK people don’t complain that Boxer vs. Dhalsim is boring, and it sounds a lot like what you describe - ’sim trying to control the match for the win, and Boxer trying to get in to land a handful of crucial blows. (And, I suspect, that the Dictator’s instant kill stuff is similar in some match-ups.)
April 12th, 2008 at 3:07 am
Both players still have to think and do a variety of things, though, even though it’s not tipped in Boxer’s favor. My “character A” was more along the lines of Old Sagat, with a mindless, easily-repeatable, and difficult to counter move.
April 12th, 2008 at 3:57 am
NateTG: Interesting concept, but I don’t think single variable changes are enough to fix an imbalanced game. It kind of works with matches in ST, because the game as a whole is pretty well balanced. There are just a few bad characters, and just a few bad matchups to fix, and some of them weren’t *really* broken. For example, a very small number of T.Hawk players won tons of tournaments, because he had an insanely good throw mixup in a corner, it was just ridiculously hard to pull off the proper series of joystick inputs, so very few players could use it.
Think of something like MvC2… Magneto can pull off an infinite on any other character. You can’t easily solve his infinite by changing a single move, because he actually has multiple move sequences that cause the infinite. You can’t raise his opponent’s damage to one-hit him in exchange for his infinite being sort of a “kill you if I manage to hit with a particular move” since that’s even more degenerate. Magneto has to pull off a lengthy and rather difficult infinite combo, but he dies to a single easy-to-use attack isn’t fair, and it would be completely broken for new players who don’t know the infinite combo in the first place. There’s just a huge number of things that makes the “good” character better than the “bad” characters (multi-directional air dashes, infinite combos, Sent’s unblockable, AHVB recovery speed). Fixing one of Cable’s “broken” moves takes him from 10 different ways to rape half the cast, down to 9 different ways to rape half the cast… but he’s still gonna win.
April 18th, 2008 at 10:05 am
O. Sagat vs someone bad is basically that bad character has to get in(which basically means guessing what sagat is going to do a few times), then after that if they guess right again they can start to do damage, but Sagat is doing damage for all of those wrong guessest. By making it so you do a lot more damage than sagat keeps the game play the same but fixes the risk vs reward. Making it easier to get in on sagat changes the game play and also fixes the risk vs reward. The thing is, giving a character 3x the damage changes every matchup, while giving them a way to get past tiger shots wouldn’t have as much as an effect. As was said, making the damage change only for that matchup is a poor way out, but it still is a valid solution if balance is your only concern.
Back to symmetric vs asymmetric for a second. Firstly, you can’t just go and issue a new definition for a term and expect everyone to understand what you mean. Saying you want an asymmetric game, but what you really want is a game to learn how to balance two different things isn’t going to get you what you’re looking for. It just happens that every game you want is asymmetric.
There is still a lot you can learn about balance from unbalanced games. First of all, how will you learn about balanced from a balanced game? You would have to know why the game is balanced to learn anything about it. Likewise, if you see how unbalanced games are unbalanced, you can learn from mistakes. Also you can just as well learn from asymmetric parts of symmetric games. If you know why(/why isn’t) 1 knight vs 1 biship + 2 pawns is balanced then you can learn from it. Even if you are unable to learn from unbalanced games, if it happens to be balanced then you should still be able to learn from it. Granted it might not be as good use of time as studying something else, but disregarding it completely is ignorant.
Ignoring symmetric games altogether is also a bad idea if mirror matches are available. If you were making a board game rts with the races of stratego, risk, and catan, you would need to have made bombs and spies or stratego vs stratego would be boring. Granted, having one or two boring mirror matches in a game with over 10 selections isn’t going to cause a huge fuss.
All in all, while you are still likely to learn more from a balanced asymmetric game than one that isn’t, you shouldn’t completely discount games that don’t fit that description.
April 18th, 2008 at 2:56 pm
Zaelar, seriously, you have some strange straw-man arguments. I asked for asymmetric games that are balanced. It’s a fine question without you saying I should also look at other types of games. No one anywhere said other types of games are bad. No one said you can’t learn from wildly unbalanced games. No one said really anything you’re arguing against. It’s a simple question. What are some balanced asymmetric games? Asking that does not preclude learning about other topics.
How can I make good blueberry ice cream? Oh you can’t go asking that without looking at chocolate ice cream. You can’t even ask that without looking at poorly made ice cream. You can’t expect people to know what you mean by blueberry ice cream. Please.
What i mean is pretty simple and people like Claytus and PoisonDagger understand it readily. You start with different sets of moves/units/cards/whatever (for example Ryu vs. Zangief) and somehow the different sets of moves/units/cards/whatever end up at a close power-level to each other. I really can expect people to understand that.
Knowing to put bombs and spies in your Stratego example would make the game deeper yes, and not more “balanced” in the way we mean it here, as you surely understand. So my question would not lead to understanding how to make that match deeper? Correct. It also doesn’t lead to understanding how to make it accessible or have good graphics, or a million other things. It’s as if you expect my one question must hold up the weight of all the world and answer all things. It must only answer what we are looking at right now, this one concept. Other questions can get at other concepts. And yet again I’ll say how sad it is that so much discussion is spent on things like this–the nature of question-asking–rather than actually finding some good example games. Thanks to those who did. I still should check out StarControl maybe.
April 18th, 2008 at 11:59 pm
I didn’t read “The Most Balanced Games” article before that post. Now after reading it I see it was just me making a bad interpretation of the second half of a conversation. Basically I was saying instead of concentrating on blueberrys for flavor that you could take a look at chocolate ice cream since even though you’d mostly get texture from it you can still get some help with your flavor if you taste it the right way.
Just skimming through the other article I came to a post where you said looking at an fps game for balance doesn’t work since even though you can pick up different weapons you start the same. This is what I’m trying to say is wrong since you can look at how those weapons are balanced. You aren’t looking at the whole game, just a particular aspect of it that you might learn from. Now as you said you are only interested in balance that has been proven through high level play, all you would need to do to make sure the weapons are balanced is see if high level players are doing well with different weapons or if they try to get the same one every time. Now it isn’t likely worth it to do this if you’ve never played an fps before, but if you’re familiar with them and someone points you to a game with good gun balance then you might be able to find something.
The (a)symmetrical part was about language, not about your clearness at this point. My point is that there needs to be a clear definition or it isn’t working as a word. These problems start when you have multiple people making up their own meanings of unclear definitions. Calling chess an asymmetric game because of the different openings is no different then saying that it is symmetric just because it doesn’t start out differently and the decisions are made later. It’s lawerly either way, and this only happens because the definition was unclear to begin with. Considering none of us are judges in this matter it won’t get resolved without an agreement from everyone as to what meaning to use or to use different words altogether, which is already happening. Now we might be fine with the definition of asymmetric being those games that you were looking for, but what happens when you meet some other groups that came up with their own definitions that are different? You get fierces(ones with frame disadvantage!) linking into hadokens.
April 19th, 2008 at 12:29 am
I have gotten these two threads (this one and the “most balanced games”) mixed up in my head at this point. So sorry if I expected you to know about the other thread in this one.
To your other points. Chess is clearly symmetric for these purposes. The two sides are the same so there is nothing to study for “balance” only depth, which is a separate topic. And an FPS where you can pick up different weapons doesn’t help either. The weapons do not need to be “balanced” in that game. It would be perfectly fine to have 4 dominant weapons that everyone uses, one that only counters 2 of those an otherwise sucks, one that only counters one and otherwise sucks, etc. You’re not locked into those weapons anyway, so they don’t need to be a the same level of fairness versus each other that characters need to be in a fighting game.
Anyway, I think the terms were defined over and over and over again in that other thread. Only TF2 took a few posts to work out where it fits. Not one single other game was confusing to classify (to me or to Claytus or to PoisonDagger). Just use common sense in what would make these terms useful. Calling chess asymmetric wouldn’t be useful. What happens when I meet up with other people who use different terms? I don’t care–at all. The terms aren’t the point. The GAMES that are balanced and asymmetric are the point.
Once you get past the 90% of all this where people can’t seem to read the definitions that are repeatedly stated in the other thread, it looks like there just aren’t many games that fit this. Any given game is either not diverse enough to really help much (Nintendo’s Ice Hockey) or diverse but not tested with real competitive play so who knows if it’s fair at all (Nintendo’s Mario Tennis), or finally it is truly asymmetric and it has been tested but people in the thread disagree on whether it’s fair/balanced or not (perhaps Dota is an example of that). These games are mysteriously hard to find.
April 19th, 2008 at 12:33 am
Zaelar, the thing about that FPS example is that in theory, you can have weapons R P and S, where R is strong against S, S is strong against P, and P is strong against R, and the game wouldn’t be unfair to any player. Both players are capable of using all the weapons, so both players can always counterpick each other during the game.
If that happened in a fighting game, that would be terrible. The character choice happens before any actual gameplay (this characteristic pretty much defines what Sirlin refers to as an asymmetric game), so if one player is counterpicked, he has a significant disadvantage for the entire game. The most balanced asymmetric games don’t have many of these situations, and are great, fun tests of skill despite the wide variety of abilities to choose from before the gameplay. Unbalanced asymmetric games aren’t good skill tests because of all the counter-picking outside of the gameplay, or they’re boring because they have few viable options to choose from.
Even if Quake’s weapon matchups are all fair, the fact remains that it’s not an asymmetric game and as such it makes little sense to study it to find out how to balance an asymmetric game.
April 19th, 2008 at 3:21 am
My fps example is just to give the idea that an asymmetric part of a symmetric game can give insight to asymmetric balance. Given the weapons have a fighting game level of balance, IE not a flat out rps win/lose scenario but various advantages between them, you could essentially have the fps play out just like a fighter, you pick your weapon and thats it for one game. Now I don’t know of any that are exactly like that, but the same balance could still be there.
There are several openings available for both sides in chess that get used at high level play, so why can’t the balance between them be analyzed like a character choice? Each has it’s strengths and weaknesses. They don’t need to be balanced the same way as characters do, but some of the ways they are balanced could work for both. It’s analyzing the balance between the choices that give depth to the gameplay in hopes to find something that applies to pre-gameplay.
April 19th, 2008 at 3:45 am
Zaeilar, this Chess question is asked-and-answered in posts #63 and #64 of this thread.
Chess has the same pieces on each side. If you want to know how on-purpose local imbalances add up to a global balance, Chess is not going to help. We already know without even looking that Chess has global balance (the two sides are the same!) and that it has local imbalances when you have X,Y,Z pieces left and I have A,B,C pieces left. Do those local imbalances add up to global balance at the start? Yes. Why? Because both sides are the same!
Zealots and Zerglings have local imbalance too. Same resource cost yet one beats the other. How does this lead to global balance overall between the two races? Good question! You have to actually play the game to find out because it’s not automatically answered by “both races have exactly the same pieces,” as it is in Chess.