The new forums will be named Coin Return (based on the most recent vote)! You can check on the status and timeline of the transition to the new forums here.
The Guiding Principles and New Rules document is now in effect.
[PATV] Wednesday, July 18, 2012 - Extra Credits Season 4, Ep. 22: Perfect Imbalance
Dog
Registered User, Administrator, Vanilla Staff admin
Registered User, Administrator, Vanilla Staff admin
This week, we discuss the benefits of subtle imbalances in games (and welcome our newest team member)!<br /> Come discuss this topic in the <a href="http://extra-credits.net/episodes/perfect-imbalance/#discuss" target="_blank">forums</a>!
Dog on
0
Posts
In your champion A, B, C example, would this imbalanced game suddenly become balanced if it turns out champion A can consistently beat champion C? Are there ways to avoid this, or can every game become rock-paper-scissors?
I suspect there is some amazing formula to predict the minimum number of champions required to "balance" a game, based on the number and scope of their stats.
The Raffinity Season I mentioned was when the balance went out of whack. Raffinity was a deck revolving around the use of artifacts to get really fast kills (like turn three or four with a perfect hand), which most decks couldn't stabilize against to get their own strategies, so the only other deck you saw at the top was a green deck full of artifact destruction. I laboured on with a White Weenie variant with limited success, but all the tournament reports had at most three different deck types being played. Wizards recognised how stagnant the metagame had gotten, so had to get the banhammer out.
We also need to separate what is possible in the game with what the players are capable of. This separation is most limited in the case of master players.
I still like to believe that formula is out there. I'm sure game theory books are full of such things. I haven't got far beyond the "How many edges, between X points" level. :-/
If the game is imbalanced, all that happens is that instead of having Lots of good builds, you limit the number to a few, and reduce the gameplay to Exactly what you accuse balanced games to be - hordes of players with the exact same builds on their toons, and no way for a new player to find a playstyle for themselves - I've been excluded from wow raids for exactly this reason - choose one of these three standard build or gtfo.
The only way the kind of imbalance you describe works is if the game keeps changing; as in magic where new cards keep changing the power levels and meaning that good players need to explore the new imbalance all the time.
Additionally, you picked plain old bad examples - examples that don't even illustrate your point, showing your lack of understanding of the issue at hand! Starcraft, for instance, didn't become the way it was because of balance, and indeed, the very idea that it did is just plain old silly. What Starcraft fell prey to was something entirely different. In games like Starcraft, there's a section prior to encountering your opponent where you are (essentially) playing a single player game. As the game is non-interactive at this point, and the game has no randomness in this segment, the game is perfectly deterministic up to a certain point. Up to that point, there are only a small number of possible optimal build orders. There may be more than one (indeed, if you look at, say, SC2, you'll see a number of interesting variations on starts, particularly for the zerg thanks to their odd builder dynamics) but each will be for different purposes and in reality, you are more or less making a (somewhat) blind decision, which is also hidden from your opponent. Only once you see what your opponent is up to does the game actually leave this mode as you finally begin to interact and make important decisions - do you send a worker to chase off their worker, or let them wander around your base and see what you're up to? When do you make sure they're not building a bunker or a pylon or something just outside of your visual range? When do you send out a worker to scout? How much are you going for economy vs military? These are all real decisions to make, but until your opponent interacts with you (or vice versa) they're non-interactive decisions and therefore can be easily optimized.
In other words, your actual complaint has absolutely nothing whatsoever to do with balance. THe issue at hand is bad design - namely, that the "real game" of starcraft doesn't actually start until several minutes into the game, as up to that point you could just put in a program that said what you did to set up. They don't do it this way, though maybe they should let you preprogram in a set of build orders to get the "not really playing" part out of the way as it is a waste of several minutes at the start of every game.
This is the same for chess as well - the issue isn't balance, its determinism, which is something entirely seperate. Anything which is determinisic allows for very deep levels of analysis.
I will also note that your choice of LoL to show what is RIGHT is actually hilariously bad. LoL is NOT any better about this than Starcraft is; indeed, it is rather worse. While you are "really playing" from quite early in the game (usually about 1:30 in, though possibly sooner if someone tries to gank) this isn't actually that much earlier than you start "really playing" in Starcraft. Worse. the build orders actually are much, much MORE deterministic in LoL than they are in Starcraft. You have to deviate from a preset BO quite early in SC oftentimes, to either react to your foe or to take advantage of an opportunity, whereas in LoL, there's typically a handful of correct sets of items for a character, and all you really do is plug in which set you're using in any particular game. Yeah, there's some minor variation, but oftentimes that doesn't show up until 15+ minutes into the game, long past the point where Starcraft has become a deeply interactive game. Additionally, there are far fewer decisions to be made in terms of actual strategy. League of Legends isn't really a strategic game, its mostly a tactical game.
Unfortunately your discussion of League of Legends also shows why a lot of game designers need to understand competitive play a lot better. You talk about how a certain champion is above average, but has a counter, and ergo you end up with a cycle of metagame play where you have the overpowered champion get beaten by some weaker champion, so a lot of people go and switch to some other overpowered champion that beats that weaker champion, but the original champion is better than that new champion, and so life goes on. Unfortunately the problem with this is that this sort of thing isn't actually balance, or even "perfect imbalance". What it is is just plain old imbalance.
If you were to read Sirlin's articles on game design, you'd see the whole RPS thing here, or Yomi. But as Sirlin points out, if you've lost the game before you've started because of the character you picked, that's just not good design. Once you're in the game, you should lose or win based on your merits, and while some amount of imbalance is inevitable, it is in no way desirable. What you want is for different characters to have different strengths and weaknesses. In the case of a team game, there is a bit more flexibility here, as really its more about your role on a team than your ability to be able to beat any other character by yourself, but there still needs to be some dynamism in there to make it so your character is different from others and has distinctive strengths and weaknesses. While this is "imbalance" in the sense that the characters don't all have the same strength and weaknesses, this is actually more properly termed asymmetry.
To put it simply: If you pick a character, and then spend the next 30-45 minutes losing a game because your opponents picked a counter to it, that's actually very bad. That's AWFUL. Its not fun at all. The shorter the interval between these decisions, the less bad this sort of imbalance is, but what it really does is make it so some characters are just much less used than others because they are, in fact, weak, and in a game like League of Legends, 90% of the time the actual proper answer to a strong champion isn't one of the weaker champions with a specific strength against that champion, but a STRONG champion with a specific strength against that champion. Its okay for them to have an advantage for picking the right champion, but it shouldn't be a huge one - 60-40 is okay, 75-25 is terrible. And your goal isn't for it to be 60-40 (though you should try to make sure no character has nothing but advantageous matchups) but rather to be as close to even as possible, and if 60-40 is the best it can be, so be it.
Once you're actually in the game, yeah, its good for risks and rewards to be different - for instance, you could 6-pool in starcraft 2, or you could try to expand rapidly, or you could do a roach rush, or you could build up for economy in a somewhat more aggressive or less aggressive way. Its okay that these strategies have different success rates against different strategies because you make that choice in-game - you aren't going to lose before the game starts, but while actually playing it.
To expand, it's my belief that any game that has player-alternated games, "I go, you go, etc", has an inherent unbalance in it that designers must account for. Probably the most popular turn-based games, besides chess(I'm speaking globally; chess is a lot more popular in places like India or Russia than in places like the US, UK, etc.), are card games, as you touched upon. I admit to not being an expert on any card games, but I'd be curious to know the statistics for how turn-order effects your chances of winning, if it has any effect at all; and if there are any strategies that work best when going first or last. And that's into delving into the possibilities created by turn-based games with more than two people, like Monopoly or something.
Go gives a point advantage to the person who goes second.
In reality, imbalance is far more interesting than balance. It's true that balance makes for more "skill-based" matches- meaning no one has an inherent advantage, and the win will almost always go to the better player. However, it gets very boring and same-y very quickly. Soon, there is absolutely no way to challenge a top-tier player. He's just better, and that's all there is to it.
But if the game has calculated imbalances, that top-tier player may find himself faced with something he is unfamiliar with, something that he shouldn't be seeing. The example that comes to mind is Gears of War- I know, it's not the best.
In GoW 1&2, the Gnasher shotgun was largely the "best weapon" due to a couple things (including the games' netcodes), but the fact remains that it was extremely popular, and most top-tier players used it to great effect. Now comes along Gears 3, which introduced two completely new weapons: the Sawed-Off Shotgun, and the Retro Lancer. Both were designed to act as a counter to the Gnasher.
At first, there was a deafening outcry from the "hardcore" players at this "noobification" of the game- as both new weapons lent themselves to easier kills for new players, and made things difficult for Gnasher users... at first. Many months and a small patch later, players are finding it's actually fairly easy to counter these new weapons. The long-time players are still getting kills, but they had to change strategies before things got better. They had to learn. They had to adapt. They couldn't continue using the same stagnant strategies they had before, because those strategies didn't take new guns into account. What's more, these new weapons can be wielded just as easily and effectively as the old ones, so long as the user knows how. Now, being a top-tier player requires more skill than before because of the change-up, and if you're not careful, you can very easily fall prey to a counter-weapon/strategy.
I have a feeling this is what they were talking about. It's not that imbalances are "good" or "desirable" or "necessary" (since they would be very easy to screw up), but that they inherently make for much more interesting gameplay. Not necessarily "better quality" gameplay, just more interesting.
The Champions A, B, and C thing reminded me of Efron's dice, where you've got four dice, A, B, C, and D, with different face values such that A usually rolls higher than B, B usually rolls higher than C, C usually rolls higher than D, and D usually rolls higher than A. Each die has a 2/3 chance of beating the next die. It's pretty weird, even after you work through the possibilities by hand.
your = belonging to you
their = belonging to them
there = not here
they're = they are
your = belonging to you
their = belonging to them
there = not here
they're = they are
Creating imbalances MAY arguably make a better mid-tier of play, but it creates a far shallower top tier. You're confusing the benefits of creating new content with the benefits of imbalance. LoL/DotA 2/HoN are all hard to figure out mainly because new content and patches constantly change the meta, but these are games with VERY established tiers, EXACTLY what you want to avoid! Why give me 90 characters if only 15 are usable?!
The problem is that a lot of people will play Champion A, or F, in a balanced play setting too. By chance, some characters will be easier to pick up, scrub-stompers or just crucial to a particular team construct. Those get picked up, then counters figured out, etc. The problem with "perfect imbalance" is that Champion A will ALWAYS see more play than Champion B or Champion C, over time, because they're better by definition BECAUSE THEIR AVERAGE MATCHUPS ARE BETTER. That's bad. Worse, if Champion S comes out that plays anything like Champion A, that character better either be better than Champion A or different in some key way, which risks them being too good, or they're worse and then why would you play them.
Rock-paper-scissors gameplay is just as boring as anything else. It's the most frustrating thing in the world to lose because I picked the wrong deck and a random tournament structure eliminated me early, even though another configuration I'd have won.
This episode was quite interesting to me but in some ways I feel like it may have missed the mark some. The issue to me between the games talked about didn't seem to be between perfect balance vs slight imbalance but rather a matter of how solvable a game is.
The issues that Chess has isn't because it's a 'perfectly' balance game (it's not, white has a known advantage to the point that many chess theorists throughout history feel that black shouldn't even play for a win, but merely to try and draw, given perfect play. Of course there is always debate going back and forth about this). Rather, the issue is that Chess is a game of perfect information, a game with no variation, a game with no luck or chance and a turn-based game meaning a player's twitch reflexes don't factor in. Ultimately this means that Chess is a solvable game. In fact its being solved more and more by computers each day (though if we will ever have computers powerful enough to solve chess and if so, when, is a matter of some debate as well. Checkers has already been fully solved, but, it's a much simpler game). Because of the way is chess is set up and because people have been teasing out its intricacies for hundreds of years is why a new player has no chance against a strong player. That is why there are books of opening moves and end games to read and memorize. Not because of its balance. If white started with 4 bishops and no knights and black started with 4 knights and no bishops the game would be just as solved (maybe even more so, if it turns out this unbalanced the game to the point where one side could easily win with a simple strategy).
The reason why League of Legends or Magic the Gathering have a dynamic, ever changing playing field not because they are slightly imbalanced, its because they are constantly adding new champions or new cards which is constantly changing how the game is played. If Chess had you start with your back row randomized it would be a far harder game to solve, which means there would have to be more thinking on your feet, and thus a more dynamic game. (This is exactly what happens in Fischer Random Chess, a chess variant proposed by Bobby Fischer due to worries over the popularity of chess disappearing due to all games resulting in draws as people get closer and closer to perfect play). If every month how a knight moved changed, it would be a similar situation. But this is exactly what is happening in League of Legends and MTG. Next week LoL will be getting 3 of its old champions revamped and getting a new one added. MTG is constantly adding new sets of cards and cycling out old ones. Some imbalance is fine in this kind of set ups because people are constantly having to learn what is strong and weak as it is constantly changing. But if MTG stopped releasing new cards? Eventually the game would move towards being solved (as much as it can be solved). I mean, even with the current Type 2 set up where new cards are rotated in and old cards rotated out there still always forms a tournament core of netdecks that pretty much everyone uses at the high competitive level of play (sometimes counter decks can be built that are purpose built to exploit some or all of the current netdecks). Only a tiny pool of all possible decks are competitive decks.
You can not solve a game that is always changing. Chess hasn't changed, LoL and MTG has. You may solve one version of the game but then a new one is hot at its heels. That is the real difference to me. Its not a matter of how well or poorly balanced the game is (not directly at least, poorly balanced games can often be easy to solve). New content is what keeps it fresh.
1) You didn't mean what you said, you mean what I say you mean, and you're therefore wrong.
I'm going to go ahead and assume the words coming out of their mouth are what he meant. Sticking to those rules, they're not breaking any logical consistency that I can see.
2) X is not balanced.
Chess, is balanced, when considering the full game, not the board on the table. Two cloned players walk into a room to play chess. Who will be more likely to win? As the colour is not predetermined, at this point both have exactly equal odds. As the determination is incorporated into the game, it is balanced.
Basically the only mistake I see Extra Credits have made is picking games as examples that are any more complex than rock/paper/scissors, because there's any room for interpretation.
The idea that there are only so many champions that are truly viable I believe is a bit ridiculous. After all, Evelyn, a champion considered to currently be the very worst of all champions is a character that a 1900 elo player plays exclusively. Somehow he has figured out that this character, is really still viable if player a specific way with a specific build. I have personally been on the receiving end of a damn good Eve player, and let me tell ya, it wasn't fun.
Even character to character comparison, most champions aren't outclassed completely by there newer counterparts. For instance take Garen and Darius. There move set is very similar, and many people have gone to say Darius is by far the better champion, with his incredibly powerful ultimate ability. However, Garen is still an incredibly viable, and fearsome champion. His kit has always been a strong one, and with ghost on him, there is really no stopping him or getting away.
What most champion match ups come to really two things, the individual skill, and how well you counter build against them. For instance, malphite vs alistar, in that kind of match up its really a toss up who wins, based entirely on there builds and masteries and run page set up. Malphite may do more damage, but Alistar can be one of the greatest tanks in the game.
My end point is that I think some of you are confusing the versilitiy of champions as the reason why the game is always changing and not easily pend down. When honestly champions are not the biggest point . Instead it really comes down to it comes down to item build and individual skill levels. League is much like starcraft in that it requires massive macro capabilities. And even with a cloned player, taking one champion against another is honestly going to put it up in the air if they are of the same type.
3DS: 0473-8507-2652
Switch: SW-5185-4991-5118
PSN: AbEntropy
Chess's "Canonically strategies" are not because the game is "Perfectly balanced" it's just been played forever and close to "Solved" Also Chess isn't "Perfectly balanced" White obviously has an advantage.
Metagames were constantly evolving in Brood War. They probably still are. What you're really seeming to praise is games having different units and powers, but that's not the same thing as "Balance"
100 episodes!
I've grown so accustomed to Allison's art that hiring on a new artist full time is a bit of a shock, but a welcome one. Grats to Leelee and all that.
Here's to you guys on your next 100 insightful and informative shows. Thanks for all that you do EC. This is one of the best web shows out there, mostly because it's not a rant show or a bash of anything but an exploration of what the medium can become (at least that's how I see it).
How much will I have to give to Kickstarter to let us all go back to calling it "TRENCHED!" instead of Iron Brigade?
Sincerely, -ZZoMBiE13
Imbalance is never a virtue by its very nature: it provides absolutely no advantage over imbalance except making it much easier to discover what the best strategies are.
http://www.sirlin.net/blog/2012/7/18/a-discussion-of-balance.html