Results 1 to 8 of 8

Thread: flaws in teh numeric approach to balance

  1. #1

    Default flaws in teh numeric approach to balance

    I just realzied some things:

    inrecent and not-so-recent interviews, blizz rep's have been referring to %wins in race-by-race matchups as a messurment of how we're doing in terms of race-by-race balance.

    i realized something: these stats are very much affected by the workings of the matchmaking system;meaning, if i find the ZvT matchup or whatever, really hard, and lose plenty in it, that means i will be pitted against oponents that are also performing worse, which will for amost everyone end up balancing out win/loss ration across racial matchups

    another way to look at it would b this: what if i ahve one matchup that i consider quite easy; that will increase my win % by a bunch, making me face better players in ALL the matchups, presumably resulting in lower win percentage in the other matchups; that is to say, below the very highest levels of play, our general knowledge, preferences and playstyles have a huge say in our success in various racial matchups; to determine what has to do with balance and what doesnt, will be hard.

    .... Above problems can probably be controlled for by factoring in skill as people, inclduing blizz reps have mentioned; whos favoured etc, in calculating the win/loss ratios across matchups.

    however, the system has no control over the races we choose to play: for exmaple, i've grown bored of zerg.. if enough players do the same, im sure this will affect the accuracy / usefulness of the ladder stats blizz are leaning on to justify changes.

    in teh end, to figure out which race is actually easier to do better with overall, will be a subjective judgment call, because most players are skewed towards playing one race or another, and noone would be able to prove that players of a certain race were on average better / worse than players of another race (given the hypothetical that they took equal effort in getting better at playing the other races..)
    Last edited by Todie; 05-17-2010 at 07:41 AM.
    I am an enthusiast of good strategy games, sc2Esports and rollplay, although i dont really play anything atm.
    I work an internship at a government agency this fall, and have a good time at it.
    I'm being more social, active and honest lately. in all forums.

    Hi.

  2. #2
    DesertRose's Avatar Junior Member
    Join Date
    Apr 2010
    Posts
    99

    Default Re: flaws in teh numeric approach to balance

    Blizzard considers only the %wins in gold and platium league, and I guess Blizzard only count the games of those who stayed in gold/platium after the first revaluation.
    While your examples may be true, if the test data is wide-ranging enough those will be balance themself out.

    But I agree, balancing just by empiric data is a bad idea, but I doubt Blizzard does that, they simply don't explain their balance changes fully because most player would be unable to understand them anyways.
    --<-{(@

  3. #3

    Default Re: flaws in teh numeric approach to balance

    There is something in your flaw i don't understand. You say if you are bad at a specific match up you will lose more then stabilize with less skilled people and it will balance the stats out. But you only chose one side of the match up you are going to play, you don't know if your going to end up against p t or z . So the match up you are the worst will be balanced with your other match up.

    The thing to understand here is that the placement and match up system don't give a F*** about your race and you opponent race but blizzard stats does. The system balance the things out making a mid-value of your match up and blizz takes the difference.

    If there is something i don't understand please tell me.

  4. #4
    TheEconomist's Avatar Lord of Economics
    Join Date
    May 2009
    Posts
    6,895

    Default Re: flaws in teh numeric approach to balance

    I'd say they hold the win ratio of progamers as being more important. They've been working on their methods for years. There's nothing you're going to think of in such a short time that they haven't thought of and considered for quite some time.

    Just sit back and let Blizzard work their magic.

  5. #5
    Junior Member
    Join Date
    May 2010
    Posts
    101

    Default Re: flaws in teh numeric approach to balance

    Blizzard in no way shape or form said that the percents from win/loss in any league or such was a sign of the game being balanced. They merely said that it was a sign they were on the right track, or a good sign. Blizzard did not employ a bunch of morons they are actually very intelligence people and balancing a game is a very complicated process.

    That being said they are not only balancing the game but they are changing stats to force us to use certain or more units and to make this and that more interesting which makes balancing the game that much challenging.
    Last edited by ChAoS; 05-17-2010 at 10:12 AM.
    ChAoS.800
    US server

  6. #6

    Default Re: flaws in teh numeric approach to balance

    Quote Originally Posted by Gt2slurp View Post
    There is something in your flaw i don't understand. You say if you are bad at a specific match up you will lose more then stabilize with less skilled people and it will balance the stats out. But you only chose one side of the match up you are going to play, you don't know if your going to end up against p t or z . So the match up you are the worst will be balanced with your other match up.

    The thing to understand here is that the placement and match up system don't give a F*** about your race and you opponent race but blizzard stats does. The system balance the things out making a mid-value of your match up and blizz takes the difference.

    If there is something i don't understand please tell me.
    While we can't control what race our opponent will choose, I think what Todie is saying is the balancing effect of this randomness is possibly masking the flaw in the final/total race win% statistic. Todie appears to be saying that a master race win% as opposed to separate win% for each matchup is/may be preventing Blizz from seeing the problem matchups clearly. For example, PvZ may have a 33% win% where PvT has a 64%, evening out to 48.5% overall, which implies balance.

    However, it must also be noted the win% statistics they have shown most likely are not the only ones they are using. We must therefore wait and see. Or we could request a broader array of statistics be released. Either way, this statisitic is generated much too simply to be viable in a true statistical analysis. What I mean is the data is very broad, creating a general statistic that doesn't describe enough for us to really create anything more than implications. Todie's concern is valid, but we must have patience and understanding that Blizz is doing this right and is not telling us everything. When have they told us everything about a game's development anyway?

    Edit: It is also worth noting that converging player skill and familiarity as we continue to play will reduce this discrepancy.
    Last edited by flak4321; 05-17-2010 at 10:46 AM. Reason: spelling and additional note
    I am a master tactician. It is my execution that keeps getting me killed.

  7. #7

    Default Re: flaws in teh numeric approach to balance

    Ok i can see the point now, thanks flak. But they surely got the statistic by match up, or they are moron, which i doubt.

  8. #8

    Default Re: flaws in teh numeric approach to balance

    And remember blizz did say that Bnet 2.0 BETA is inherently flawed for testing. We don't know how they match you up.
    Click or else your a egg killer. Do you want to be a baby killer O.o?

Similar Threads

  1. Balance the game ... game
    By Wankey in forum StarCraft Discussion
    Replies: 53
    Last Post: 05-05-2010, 01:55 AM
  2. Current balance discussion(V.7)
    By SlickR in forum StarCraft Discussion
    Replies: 44
    Last Post: 03-25-2010, 02:38 PM
  3. StarCraft 2 Balance Team is...
    By XSOLDIER in forum StarCraft Discussion
    Replies: 118
    Last Post: 01-14-2010, 10:48 AM
  4. Balance team says...
    By Eligor in forum StarCraft Discussion
    Replies: 85
    Last Post: 01-06-2010, 08:48 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •