Balanced MU
Balanced MU
The other day I had a discussion with @momuuu about when we can say a MU is balanced. First of all let's assume we have players of roughly equal skill. When is a MU balanced and when is it tilting towards a civ? Is 55-45 balanced or we want a smaller difference. Ideally a 50-50 is the perfect balance, but where should our hypothetical threshold lie?
Secondly, how many games do they have to play to conclude the status of a MU ? For a winrate margin of 5%, we can accept something less than 400 games, if we decrease that to 2.5% we need around over 1500 games.
Considering these 2 points, the balance threshold and the number of games needed, is it possible to ever achieve a trustworthy result for the status of a MU?
Note: Let's talk about the 91 non-mirror MUs as well.
Secondly, how many games do they have to play to conclude the status of a MU ? For a winrate margin of 5%, we can accept something less than 400 games, if we decrease that to 2.5% we need around over 1500 games.
Considering these 2 points, the balance threshold and the number of games needed, is it possible to ever achieve a trustworthy result for the status of a MU?
Note: Let's talk about the 91 non-mirror MUs as well.
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations
"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
http://www.tylervigen.com/spurious-correlations
"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
- [Armag] diarouga
- Ninja
- Posts: 12710
- Joined: Feb 26, 2015
- ESO: diarouga
- Location: France
Re: Balanced MU
BrookG wrote:The other day I had a discussion with @momuuu about when we can say a MU is balanced. First of all let's assume we have players of roughly equal skill. When is a MU balanced and when is it tilting towards a civ? Is 55-45 balanced or we want a smaller difference. Ideally a 50-50 is the perfect balance, but where should our hypothetical threshold lie?
Secondly, how many games do they have to play to conclude the status of a MU ? For a winrate margin of 5%, we can accept something less than 400 games, if we decrease that to 2.5% we need around over 1500 games.
Considering these 2 points, the balance threshold and the number of games needed, is it possible to ever achieve a trustworthy result for the status of a MU?
Note: Let's talk about the 91 non-mirror MUs as well.
A MU is balanced when both civs can win when both players play well.
You can't really test that because first it's impossible to find two equally skilled players. And also, there's a random factor: treasures, map spawn, positioning, counter strat. And if there wasn't this random factor regarding strats, the MU actually wouldn't be balanced, because if two players go for the same strat again and again, one strat will become better in the end.
Re: Balanced MU
Guys can you focus on the part that we assume to have them? You can't make any conclusions like that
Correlation doesn't mean causation.
http://www.tylervigen.com/spurious-correlations
"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
http://www.tylervigen.com/spurious-correlations
"mr.brookg go buy jeans and goto the club with somppuli" - Princeofkabul, July 2018
- [Armag] diarouga
- Ninja
- Posts: 12710
- Joined: Feb 26, 2015
- ESO: diarouga
- Location: France
Re: Balanced MU
BrookG wrote:Guys can you focus on the part that we assume to have them? You can't make any conclusions like that
Honestly, a match up, unless it is a mirror, can't be "objectively" balanced.
First, it depends on the map. For instance, Germany vs Russia is 60/40 on TP maps, and 40/60 on no TP maps. Thus it's an unbalanced MU (because the MU never gives both players equal chances), but if you play 1500 games on 1 TP map, and 1500 games on 1 no TP map, it's going to be 50%.
And it also depends on the meta. Sometimes new strats come up and it totally changes the way a MU is played. Thus to call a MU objectively balanced, you'd have to assume that you know how this MU should be played, which isn't true.
Thus, at best, you can judge a MU in the current meta.
Re: Balanced MU
60/40 is a god very synthetic assesment of balanced MU for this game.
Basically 20% variance allows for random elements to favor one or the other civ but still within what it can be changed by outplaying.
With that said, there is much more to it. A MU can be balanced at start but can go very wrong depending on age1 outcome or even just a tiny mistake from one player. Jap MUs are usually balanced on paper but if Jap is given just about too much time then the assessment changes greatly in favor of Jap.
Clearly the more similar are the civs the more balanced the MU is and keeps being throughout the entirety of the game. Because of that the most balanced MUs are those within the same expansion.
Basically 20% variance allows for random elements to favor one or the other civ but still within what it can be changed by outplaying.
With that said, there is much more to it. A MU can be balanced at start but can go very wrong depending on age1 outcome or even just a tiny mistake from one player. Jap MUs are usually balanced on paper but if Jap is given just about too much time then the assessment changes greatly in favor of Jap.
Clearly the more similar are the civs the more balanced the MU is and keeps being throughout the entirety of the game. Because of that the most balanced MUs are those within the same expansion.
Re: Balanced MU
Tfw when you try to make a thread anout statistics and then diarouga and garja dont get your point at all but start arguing that you are wrong anyways.
Re: Balanced MU
I admit I didn't read the OP carefully. Then again I kinda answered to that anyway. 60-40 is already fine to me, so 55-45 is as well.
As for the sample size, it either has to be gigantic if random games or even just very small if qualified games (games were two good players actually used proper strategies and didn't make evident mistakes).
As for the sample size, it either has to be gigantic if random games or even just very small if qualified games (games were two good players actually used proper strategies and didn't make evident mistakes).
-
- Pro Player
- Posts: 10278
- Joined: Jun 6, 2015
- Location: Paris
- GameRanger ID: 5529322
Re: Balanced MU
momuuu wrote:Tfw when you try to make a thread anout statistics and then diarouga and garja dont get your point at all but start arguing that you are wrong anyways.
Tfw people answer in a constructive manner to the thread but Jerom still complains
LoOk_tOm wrote:I have something in particular against Kaisar (GERMANY NOOB mercenary LAMME FOREVER) And the other people (noobs) like suck kaiser ... just this ..
Re: Balanced MU
Garja wrote:I admit I didn't read the OP carefully. Then again I kinda answered to that anyway. 60-40 is already fine to me, so 55-45 is as well.
As for the sample size, it either has to be gigantic if random games or even just very small if qualified games (games were two good players actually used proper strategies and didn't make evident mistakes).
What you are trying to conclude is that its basically impossible to measure balance with the resources we have. 60-40 doesnt sound very balanced to me to be honest. I know that they use a 5% margin in sc2 which seems somewhat fair.
Re: Balanced MU
It's not like you can actually quantify the balance for any mu anyway. Saying x mu is 55-45 is at the very best nothing more than a somewhat educated guess and usually not even that
Re: Balanced MU
momuuu wrote:What you are trying to conclude is that its basically impossible to measure balance with the resources we have. 60-40 doesnt sound very balanced to me to be honest. I know that they use a 5% margin in sc2 which seems somewhat fair.
Yes, the assessment of balance in AOE3, at the moment at least, is mostly based on a qualitative approach. This is why I tend to consider my personal experience over any tourney game or w/e. In any case the sample size provided by tourneys is not enough to be statistically relevant.
60-40 to me is a good range for balance in this game. In AOE3 civ design is a strong feature and it often sacrificies balance. Civs are quite different from each other and each civ is given some peculiar strenght that doesn't cope well with the pure concept of balance. Also the level of play is so that the magnitude of mistakes easily alters the chance of winning in the order of 20%. Or at least, that is in many MUs and in most cases.
Over time a consistent 60-40 will clearly show that one civ is favorited. But that's where tourney rules come into play. Also we are talking about a specifciMU but the balance process considers all MUs at once. So it is entirely possible and yet still fair, for each civ to have couple MUs that they strongly win or lose.
Re: Balanced MU
And a civ that is 60-40 in all match ups?
Re: Balanced MU
A civ like that is clearly not balanced on paper.
However it can still be considered acceptable if the civ is harder to play than the others. Winning chances (or civ potential) should be inversely proportional to its difficulty.
However it can still be considered acceptable if the civ is harder to play than the others. Winning chances (or civ potential) should be inversely proportional to its difficulty.
Re: Balanced MU
Garja wrote:A civ like that is clearly not balanced on paper.
However it can still be considered acceptable if the civ is harder to play than the others. Winning chances (or civ potential) should be inversely proportional to its difficulty.
So, within what margin of average winrate would you consider it 'fine'. I know there's a lot of practise but that is not the matter at hand garja.
Re: Balanced MU
Well, I don't see much point in assessing that, as it would be a completely unrealistic scenario anyway.
Ideally a civ that has consistent winrate throughout all the MUs should have something very close to 50-50 in order to be called balanced by definition. Again, that is by definition and as an average winrate. This means that it can also be 60-40 if they mix up and compensate to an overall 50-50. In fact that is kinda the realistic achievable balance for AOE3, without completely mutilating the game of design features.
Ideally a civ that has consistent winrate throughout all the MUs should have something very close to 50-50 in order to be called balanced by definition. Again, that is by definition and as an average winrate. This means that it can also be 60-40 if they mix up and compensate to an overall 50-50. In fact that is kinda the realistic achievable balance for AOE3, without completely mutilating the game of design features.
Re: Balanced MU
Why cant you just respond to simple questions.
If a civ has a 55% winrate across the board, is it overpowered or not?
If a civ has a 55% winrate across the board, is it overpowered or not?
Re: Balanced MU
I already answered. Ideally the average win rate should be 50-50 to be balanced. In practice it can be vary a lot from that.
Also no, it wouldn't be OP just not ideally balanced.
Anyway to sum it up, considered practical purposes:
- average of all MUs should be roughtly 50-50;
- variance can be up to 25% in one direction.
I am talking of a single civ here, for simplicity. But the same reasoning can be extended to all MUs in general.
Also no, it wouldn't be OP just not ideally balanced.
Anyway to sum it up, considered practical purposes:
- average of all MUs should be roughtly 50-50;
- variance can be up to 25% in one direction.
I am talking of a single civ here, for simplicity. But the same reasoning can be extended to all MUs in general.
Re: Balanced MU
Garja wrote:I already answered. Ideally the average win rate should be 50-50 to be balanced. In practice it can be vary a lot from that.
Also no, it wouldn't be OP just not ideally balanced.
Anyway to sum it up, considered practical purposes:
- average of all MUs sohuld be roughtly 50-50;
- variance can be up to 25% in one direction.
So you think the average of all mus being 75-25 is okay?
Re: Balanced MU
No?! I didn't say that.
The average should be 50-50. Some MUs can vary up to 25% in one direction. They clearly have to compensate for the average to be 50-50.
The average should be 50-50. Some MUs can vary up to 25% in one direction. They clearly have to compensate for the average to be 50-50.
Re: Balanced MU
Gendarme wrote:We should start discussing balance when people get good at the game. You all suck ass.
You're not wrong
Re: Balanced MU
Gendarme wrote:We should start discussing balance when people get good at the game. You all suck ass.
And some of us know what you suck
Who is online
Users browsing this forum: No registered users and 6 guests