|
On August 03 2025 02:46 Poopi wrote:Show nested quote +On August 03 2025 02:35 PremoBeats wrote:On August 03 2025 02:33 Poopi wrote:On August 03 2025 02:13 Balnazza wrote:On August 03 2025 01:41 MJG wrote:I see people went down the route of asserting that BL/Infestor was fine because the winrate statistics looked okay.  A demoralising outcome, but entirely expected. Game-design and balance are not the same thing. I think Medivacs are fucking stupid, doesn't make them imbalanced per se... BL/infestor was completely broken level of imbalance. Yet, the other races probably had ways to counter it to arrive at approximately 50/50 win rates... Gosh, this is so frustrating. Not really, we just had to wait for HotS to patch the game out. Blizzard was relatively active back then
So they did it so fast that the imbalance didn't have a big effect on win rates? Or what is your take?
If so, then what's the big deal?
On August 03 2025 02:49 Charoisaur wrote:Show nested quote +On August 02 2025 21:44 PremoBeats wrote:On August 02 2025 21:20 WombaT wrote:
I think one can make the argument Reynor had the greater year, in the way that winning a golf major or a tennis grand slam are the biggest prizes in not just money, but prestige, but he definitely had the worse season overall.
Let’s assume Serral is a golf or a tennis player and think of what that looks like over the span from 2018 outwards
1. He’d have been world number 1 for almost all of that period.
2. He’d have won the most regular tournaments going (or equal, can’t 100% remember.
3. He’d have many statistical records, and ‘best regular seasons’ ever recorded.
4. He’d also have had dominant years, or years he did win majors.
If a player had that profile, I think the discussion wouldn’t be about if they were the era’s best, it would shift into ‘they’re so good they should have won more slams/majors’ territory. Of course it’s a little different because those games have a century+ of history in terms of GOAT chat and legacy, but it’s something said about a Rory McIlroy (although he’s not quite a Serral) But G8 didn't have any legacy in SC2. Prize money was there.. prestige in the sense of legacy not so much, I'd say. On August 02 2025 21:30 Charoisaur wrote:On August 02 2025 14:24 PremoBeats wrote:. WombaT wrote:Where are you getting the idea that Protoss was particularly bad in 2018 from?
Probably because ejozl used prize money as a metric and didn't realize how it isn't a very good data set to check for balance or determine the GOAT. Taking prize money as single indicator to determine balance is just as senseless as taking winrate as single indicator Not nearly as senseless as prize money, as my example from a couple of pages ago has shown. But what exactly is wrong with taking map wins rates from around 1000k games/year from the top tournaments, on which the 3 races duked it out? And how would you make it better? Or which better metric do you have at hand to scan for balance? WombaT already explained a few of the reasons why pure winrate isn't a good metric. To the question what would be a good metric... well I don't think there is any one that the community would agree on, that's why we always have those heated debates. Some people look at tournaments won, some at ro32 representation, some at GM representation, some at winrate, some at prize money... usually aligning with what would make their race look underpowered and the others overpowered. Personally, I think the best we have is representation at later rounds of tournaments. It's not perfect either because it's possible the players of one race are just better than of the other races, but with multiple players that start overperforming after a certain patch, I think it becomes rather unlikely that it's only because of skill. Additionally, that measure only considers the very highest level of skill, which means the practical human ceiling for each race. I really don't think a match between Mixu and JuggernautJason is relevant for the question whether Serral or Maru have an easier time winning tournaments.
A good metric is not necessarily one that everyone agrees on, but one that - to the closest proximity - gives numbers to the phenomenon we want to analyze.
And as you perfectly laid out, Ro16 or Ro8 analyses are heavily skewed by 1 and even more so 2 or 3 good players as the sample size is too small.
I did not get though, why you think map win rates are not worthwhile according to WombaT's input. He mostly forwarded his own way of assessing balance, as far as I remember.
|
France12886 Posts
The imbalance had an impact, but I am not sure what your point is. Do you truly think BL/infestor was balanced? 
Let’s imagine a game then, with two races, A and B, a player of race A with skill at 100, two players of race B with skills of 100 and 130.
In this game, A is overpowered as long as your opponent isn’t more than 20 points better than you in terms of skill. So you win vs player below 120 skill points.
Player from race A plays the equally skilled player ten times and win ten times. He plays the superior skilled player ten times and loses ten times. Overall win rate of race A versus B is 50%. Does that mean the game is balanced?
So if you think BL/infestor was balanced due to map win rates or whatever, consider another approach that does not lead to such an absurd conclusion and come back with it
|
In regards to your example: you think that the Terran players were all so much more skilled than the Zerg players (including Serral and Reynor back then), that they made the BL/Inf-imbalance solely void because of their skill not because Terran was given tools to deal with the matchup?
That is your explanation?
Or that there were a lot more low level Zerg that made it possible for the Terrans to equal out the win rate in the end?
To word it out: no, BL/Inf was not balanced in late game. But the most logical explanation is that the Terrans had tools to deal with it at other stages of the game. Otherwise, when it was removed, we would have seen that Zergs get absolutely obliterated, as the Terrans were supposedly so much better and had this massive hurdle put of their way.
|
On August 03 2025 03:05 PremoBeats wrote:Show nested quote +On August 03 2025 02:46 Poopi wrote:On August 03 2025 02:35 PremoBeats wrote:On August 03 2025 02:33 Poopi wrote:On August 03 2025 02:13 Balnazza wrote:On August 03 2025 01:41 MJG wrote:I see people went down the route of asserting that BL/Infestor was fine because the winrate statistics looked okay. A demoralising outcome, but entirely expected. Game-design and balance are not the same thing. I think Medivacs are fucking stupid, doesn't make them imbalanced per se... BL/infestor was completely broken level of imbalance. Yet, the other races probably had ways to counter it to arrive at approximately 50/50 win rates... Gosh, this is so frustrating. Not really, we just had to wait for HotS to patch the game out. Blizzard was relatively active back then So they did it so fast that the imbalance didn't have a big effect on win rates? Or what is your take? If so, then what's the big deal? Show nested quote +On August 03 2025 02:49 Charoisaur wrote:On August 02 2025 21:44 PremoBeats wrote:On August 02 2025 21:20 WombaT wrote:
I think one can make the argument Reynor had the greater year, in the way that winning a golf major or a tennis grand slam are the biggest prizes in not just money, but prestige, but he definitely had the worse season overall.
Let’s assume Serral is a golf or a tennis player and think of what that looks like over the span from 2018 outwards
1. He’d have been world number 1 for almost all of that period.
2. He’d have won the most regular tournaments going (or equal, can’t 100% remember.
3. He’d have many statistical records, and ‘best regular seasons’ ever recorded.
4. He’d also have had dominant years, or years he did win majors.
If a player had that profile, I think the discussion wouldn’t be about if they were the era’s best, it would shift into ‘they’re so good they should have won more slams/majors’ territory. Of course it’s a little different because those games have a century+ of history in terms of GOAT chat and legacy, but it’s something said about a Rory McIlroy (although he’s not quite a Serral) But G8 didn't have any legacy in SC2. Prize money was there.. prestige in the sense of legacy not so much, I'd say. On August 02 2025 21:30 Charoisaur wrote:On August 02 2025 14:24 PremoBeats wrote:. WombaT wrote:Where are you getting the idea that Protoss was particularly bad in 2018 from?
Probably because ejozl used prize money as a metric and didn't realize how it isn't a very good data set to check for balance or determine the GOAT. Taking prize money as single indicator to determine balance is just as senseless as taking winrate as single indicator Not nearly as senseless as prize money, as my example from a couple of pages ago has shown. But what exactly is wrong with taking map wins rates from around 1000k games/year from the top tournaments, on which the 3 races duked it out? And how would you make it better? Or which better metric do you have at hand to scan for balance? WombaT already explained a few of the reasons why pure winrate isn't a good metric. To the question what would be a good metric... well I don't think there is any one that the community would agree on, that's why we always have those heated debates. Some people look at tournaments won, some at ro32 representation, some at GM representation, some at winrate, some at prize money... usually aligning with what would make their race look underpowered and the others overpowered. Personally, I think the best we have is representation at later rounds of tournaments. It's not perfect either because it's possible the players of one race are just better than of the other races, but with multiple players that start overperforming after a certain patch, I think it becomes rather unlikely that it's only because of skill. Additionally, that measure only considers the very highest level of skill, which means the practical human ceiling for each race. I really don't think a match between Mixu and JuggernautJason is relevant for the question whether Serral or Maru have an easier time winning tournaments. A good metric is not necessarily one that everyone agrees on, but one that - to the closest proximity - gives numbers to the phenomenon we want to analyze. And as you perfectly laid out, Ro16 or Ro8 analyses are heavily skewed by 1 and even more so 2 or 3 good players as the sample size is too small. I did not get though, why you think map win rates are not worthwhile according to WombaT's input. He mostly forwarded his own way of assessing balance, as far as I remember.
Quoted from WombaT's post
It a game has an imbalanced period, in theory it’ll elevate worse players up to a level they generally don’t compete at normally. So win rates are useful, but they’ll almost naturally stabilise to some degree if we’re talking minor imbalance versus the game being broken.
That was also my point when in a previous discussion I linked the Code S season with 4 out of 32 players being terran but winrates being 50/50. Because only the tip of the top terrans were left in a wildly imbalanced meta
|
You'd have to have at least two tournaments in a row to exclude pure chance and to determine representation bias.
If it would actually be the case that this occured several times on the same patch, then yes, one would need to dig deeper and add another analysis method for that time frame.. but how often did such a thing happen?
But point taken... I will think about including an offset for representation imbalances when they occur for x times in a certain amount of time, despite "normal" looking win rates.
|
Northern Ireland25342 Posts
On August 03 2025 02:49 Charoisaur wrote:Show nested quote +On August 02 2025 21:44 PremoBeats wrote:On August 02 2025 21:20 WombaT wrote:
I think one can make the argument Reynor had the greater year, in the way that winning a golf major or a tennis grand slam are the biggest prizes in not just money, but prestige, but he definitely had the worse season overall.
Let’s assume Serral is a golf or a tennis player and think of what that looks like over the span from 2018 outwards
1. He’d have been world number 1 for almost all of that period.
2. He’d have won the most regular tournaments going (or equal, can’t 100% remember.
3. He’d have many statistical records, and ‘best regular seasons’ ever recorded.
4. He’d also have had dominant years, or years he did win majors.
If a player had that profile, I think the discussion wouldn’t be about if they were the era’s best, it would shift into ‘they’re so good they should have won more slams/majors’ territory. Of course it’s a little different because those games have a century+ of history in terms of GOAT chat and legacy, but it’s something said about a Rory McIlroy (although he’s not quite a Serral) But G8 didn't have any legacy in SC2. Prize money was there.. prestige in the sense of legacy not so much, I'd say. On August 02 2025 21:30 Charoisaur wrote:On August 02 2025 14:24 PremoBeats wrote:. WombaT wrote:Where are you getting the idea that Protoss was particularly bad in 2018 from?
Probably because ejozl used prize money as a metric and didn't realize how it isn't a very good data set to check for balance or determine the GOAT. Taking prize money as single indicator to determine balance is just as senseless as taking winrate as single indicator Not nearly as senseless as prize money, as my example from a couple of pages ago has shown. But what exactly is wrong with taking map wins rates from around 1000k games/year from the top tournaments, on which the 3 races duked it out? And how would you make it better? Or which better metric do you have at hand to scan for balance? WombaT already explained a few of the reasons why pure winrate isn't a good metric. To the question what would be a good metric... well I don't think there is any one that the community would agree on, that's why we always have those heated debates. Some people look at tournaments won, some at ro32 representation, some at GM representation, some at winrate, some at prize money... usually aligning with what would make their race look underpowered and the others overpowered. Personally, I think the best we have is representation at later rounds of tournaments. It's not perfect either because it's possible the players of one race are just better than of the other races, but with multiple players that start overperforming after a certain patch, I think it becomes rather unlikely that it's only because of skill.Additionally, that measure only considers the very highest level of skill, which means the practical human ceiling for each race. I really don't think a match between Mixu and JuggernautJason is relevant for the question whether Serral or Maru have an easier time winning tournaments.
Agreed, I think ideally you somehow crunch them all into some metric, although fuck knows how that would work haha.
But there are flaws in focusing too much on all of those specific metrics, especially as, as you say, people tend to cherry pick to suit their assumptions.
We’re also at a stage where the field is not nearly as deep too, so a handful of players can skew things to a degree they couldn’t maybe a decade ago. Folks can argue balance or otherwise, I but that aside don’t think it’s coincidental that Toss have more regularly been at the business end of tournies with both herO and Classic in good form. Versus when it was Trap or Zest showing up with his new build, or basically just herO, you’re not relying heavily on one guy.
In a hypothetical world where Classic wasn’t motivated enough to get back into top shape, and herO had the same tournament he actually had at EWC, some Toss fans on Reddit and here would have been whining incessantly, even if the game state was actually the same.
GM representation absolutely is skewed. But it doesn’t really answer the question if Toss is unbalanced at the top level, it answers a different one, namely that it’s easier to get to a very high, but not necessarily elite level with Toss. I’ve personally never disputed this since WoL, I know some do but I find it really hard to make that argument.
There’s compounding effects too, potentially. Maybe Joe Terran and Jane Zerg are as skilled as Dave GM Toss, but maybe they got sick of playing vP all the time and just quit playing. So even if Blizz changed tack and decided to balance for ladder players (or indeed, patch much at all), maybe the GM representation still remains similar.
Overall tournament representation can be quite good to look at, although it has flaws as well. We’ll have players qualified from regionals in early rounds that likely wouldn’t be there if it was only open international qualifiers. You’ll also see fluctuations almost purely based on bracket luck, one can even get a sense of this skimming various Premiers. For an extreme hypothetical WombaT’s Single Elimination Tournament let’s say Serral the ZvP god faces herO round 1 and Classic round 2. Protoss end up having a very bad tournament indeed, when maybe they do absolutely fine with a different bracket.
On the bolded, 100% agree. It’s really what makes the BL/Infestor era stand out. There are players posting results (at various tiers), that they weren’t really doing before, or indeed after this phase, and really obviously so.
If we look at some of the other eras being mentioned in relation to this thread, you really don’t see that to near the same degree. It’s a handful of players really at the very bleeding edge maybe doing slightly better.
Although, as I stressed prior, I’m not saying some of these years we saw perfect balance, but I don’t think it was anything like as extreme as BL/Infestor.
For one other thing, BL/Infestor, you saw the same comp and styles to get there being an issue in both Zerg non-mirror matchups. Plenty of other problem metas were really only skewed by one matchup having problems, I’d say arguably the majority. In the others were a faction was perhaps too strong in both matchups, it’s often not the same problem, at least not easily visibly so, so how one fixes it becomes a good deal trickier
|
Northern Ireland25342 Posts
On August 03 2025 04:12 PremoBeats wrote:
In regards to your example: you think that the Terran players were all so much more skilled than the Zerg players (including Serral and Reynor back then), that they made the BL/Inf-imbalance solely void because of their skill not because Terran was given tools to deal with the matchup?
That is your explanation?
Or that there were a lot more low level Zerg that made it possible for the Terrans to equal out the win rate in the end?
To word it out: no, BL/Inf was not balanced in late game. But the most logical explanation is that the Terrans had tools to deal with it at other stages of the game. Otherwise, when it was removed, we would have seen that Zergs get absolutely obliterated, as the Terrans were supposedly so much better and had this massive hurdle put of their way.
I dunno if you were following the game in WoL during peak BL/Infestor, it was pretty bad.
Thing is, it wasn’t the only viable way to play, it was just easier and more reliable a lot of the time.
And, unlike some other periods people complain about, you really did see players getting atypically good results, ones they didn’t really match either before or after this period.
Nerfs come in and the top Zergs could go back to playing ling/bling/muta or whatever, which they could have played anyway, and indeed did.
Another thing with BL/Infestor is it was a great ‘equaliser’ for the worse players, but not a toolkit the best players can really be much ‘better’ with past a point.
Say you buffed, I dunno, Battlecruisers or Carriers, that’s probably going to help out lower tier players more than the very top ones (to a point), whereas if say, Stalkers or marines got even some very minor buff, it would be guys like herO or Clem with the ability to take advantage.
|
On August 03 2025 02:13 Balnazza wrote:Show nested quote +On August 03 2025 01:41 MJG wrote:I see people went down the route of asserting that BL/Infestor was fine because the winrate statistics looked okay. A demoralising outcome, but entirely expected. Game-design and balance are not the same thing. I think Medivacs are fucking stupid, doesn't make them imbalanced per se...
Thank god someone else has this opinion! You always hear about how Warp Gate and Recall makes Protoss impossible to balance around but the cheap flying transport unit with virtually infinite heals and instant pickup ability gets a pass for some reason.
I'll be the first to admit I don't know anything about how to balance a game, so take this with a grain of salt - I think abilities that prevent mistakes from being punished should be approached with as much caution as free units. Z over extends late game into a fortress? Snipes take out most of the ultras on retreat. Terran over extend during their 6 minute stim pressure into a full surround? Every unit gets away scot-free every time.
Like idk but I think mistakes should be punishable for all races.
/end rant
|
Northern Ireland25342 Posts
On August 03 2025 09:22 sc2turtlepants wrote:Show nested quote +On August 03 2025 02:13 Balnazza wrote:On August 03 2025 01:41 MJG wrote:I see people went down the route of asserting that BL/Infestor was fine because the winrate statistics looked okay. A demoralising outcome, but entirely expected. Game-design and balance are not the same thing. I think Medivacs are fucking stupid, doesn't make them imbalanced per se... Thank god someone else has this opinion! You always hear about how Warp Gate and Recall makes Protoss impossible to balance around but the cheap flying transport unit with virtually infinite heals and instant pickup ability gets a pass for some reason. I'll be the first to admit I don't know shit about how to balance a game, so take this with a grain of salt - I think abilities that prevent mistakes from being punished should be approached with as much caution as free units. Z over extends late game into a fortress? Snipes take out most of the ultras on retreat. Terran over extend during their 6 minute stim pressure into a full surround? Every unit gets away scot-free every time. Like idk but I think mistakes should be punishable for all races. /end rant
‘But it’s unfair!’
People don’t consider their own tools and how annoying they are.
At my peak, relative to my humble level I was a PvT monster. I was also rather good at TvP as a result. You see it from both sides. I had about the same MMR/league as a P and T, although partly because I was absolutely awful at PvZ with my main, but OK at TvZ.
If you sneak a double drop in somewhere, no amount of warp ins will save you, because small numbers versus small numbers, with decent micro, that drop will nuke everything. Which I imagine is why recall even exists.
If you’re decent at micro, a few medivacs of bio can just kite and kill almost infinite defensive zealots. And there’s nothing the Protoss player can do better to switch that equation.
On the flipside a Toss in a good game state can basically just A-move to a win, absolutely does happen.
For me a big problem in perception is Terran players never consider what they can do that a Toss can’t, and it’s forever been the case. Toss is easier in many ways to micro but you can’t extract nearly the same value from it with good micro, whereas Terran yeah it’s hard but you can. If your micro is good enough you can almost not think about the game state or transitions and play MMM very well and do ok.
Terran players seemingly want to pick the ‘hard’ race and complain it’s hard, while disregarding the upside.
I played shitloads of WC3 so my micro is pretty good, back in the days before F2P and just getting bored of the Toss meta I was able to just switch to Terran outright and, a few initial stomps I did OK and basically kept my account at a similar range.
The different factions bring their own challenges. I’m a super risk averse player so all the Toss cheeses/timings while being strong, I don’t really go for, which means I miss out on a lot of the faction’s strengths.. But I’m not great at reading games, so Zerg which in terms of my mentality and approach overall should be my best, is not and is my worst. Terran at my level I don’t really have to engage the brain, just survive early game and try to win a mechanical slugfest that I’m ok in
|
On August 03 2025 06:37 WombaT wrote:Show nested quote +On August 03 2025 04:12 PremoBeats wrote:
In regards to your example: you think that the Terran players were all so much more skilled than the Zerg players (including Serral and Reynor back then), that they made the BL/Inf-imbalance solely void because of their skill not because Terran was given tools to deal with the matchup?
That is your explanation?
Or that there were a lot more low level Zerg that made it possible for the Terrans to equal out the win rate in the end?
To word it out: no, BL/Inf was not balanced in late game. But the most logical explanation is that the Terrans had tools to deal with it at other stages of the game. Otherwise, when it was removed, we would have seen that Zergs get absolutely obliterated, as the Terrans were supposedly so much better and had this massive hurdle put of their way. I dunno if you were following the game in WoL during peak BL/Infestor, it was pretty bad. Thing is, it wasn’t the only viable way to play, it was just easier and more reliable a lot of the time. And, unlike some other periods people complain about, you really did see players getting atypically good results, ones they didn’t really match either before or after this period. Nerfs come in and the top Zergs could go back to playing ling/bling/muta or whatever, which they could have played anyway, and indeed did. Another thing with BL/Infestor is it was a great ‘equaliser’ for the worse players, but not a toolkit the best players can really be much ‘better’ with past a point. Say you buffed, I dunno, Battlecruisers or Carriers, that’s probably going to help out lower tier players more than the very top ones (to a point), whereas if say, Stalkers or marines got even some very minor buff, it would be guys like herO or Clem with the ability to take advantage.
I never played the game competitively, so my vision on it might be skewed in comparison to people who actually did.
But speaking of WoL: I found some heavy asymmetries in win rates - not for Zerg, but for Terran, especially versus Toss. Given that I am not finished with the year, it might still equal out or get worse, but a thought crossed my mind, when seeing Terrans being a lot more represented on top of that:
You and Charoisaur made the point that imbalances in the game will eventually lead to a roughly 50:50ish map win rate if given enough time, as the disfavored race will lose players in qualifiers but the elite will figure out a way to deal with it. Thus, the win rate signals a rather ok 50:50, but player representation is in the gutters. Is that description accurate?
So what about an expected offset-correction?
1. We'd expect 33% representation, leaving out random players as they were and are an absolute minority.
2. Chance and randomness as well as 1 or 2 strong players having a high impact on representation numbers need to be considered, so I'd say +/- 5 percentage points are fine, which means a 15% deviation marks the correction.
3. Everything above/below that threshold will be penalized/buffed with x percentage points in the multiplier per y percentage point deviation from the expected outcome threshold.
4. I'd count per year, because otherwise I would have to correct for numbers of tournaments too. It will go faster, if I take the whole year as a basis and shouldn't make much of a difference.
I'd still have to figure out what to do about MaxPax, because he is missing in all these offline events, which has got nothing to do with balance... but perhaps I can extrapolate from the online-events.
What's your opinion on that?
|
Northern Ireland25342 Posts
I’m genuinely not sure it’s calculable in the abstract.
I mean just crudely speaking, Toss is easier to a point, but has a lower mechanical ceiling.
I think there’s two components to conceptions of balance, chance to win is part of it, but how to do it, how difficult it is or how limited your options are is part of it, what I’d call the ‘balance of fairness’.
I’ve advocated for atypical performance as a potential metric, but it does somewhat presume the typical baseline performance is in balanced periods, which may not actually be the case. Additionally I think it becomes harder to ascertain in more recent times, there’s just a less deep pool of equivalently good players.
There’s also intangibles in various potential data sets. If we take something like weeklies, some top players don’t partake. Some do, but it’s just a shot for some pocket money and a way to keep in shape. For some, it’s kinda the peak level they can compete at, and I I imagine in combination that is going to skew what win rates we can glean from them.
For bigger tournaments, most have regional qualification, so tournaments don’t tend to get the best possible players for each faction at any given time.
If we try to factor in an ELO system like Aligulac, some players are over or underrated depending on activity in things like weeklies, or regional discrepancies. Something even the Aligulac crew concede is a problem they can’t really fix.
I think your method would work if SC2 was fully international in qualification, and only had Premier tournaments, more of them and maybe a WC tier event.
|
|
|
|
|
|
EPT
