|
On September 22 2020 18:57 Shuffleblade wrote:Show nested quote +On September 22 2020 18:13 mounteast0 wrote:
Mate, firstly, you are not talking about balance, but "game state" / "entertainment value". Balance is having 50% win rate for equal skilled player, regardless of whether the game is fun, enjoyable to watch etc. If ZVT end with 50% win rate with 100% of the games being drone rush, it is balance. Of course it is not going to be enjoyable for the player or the viewer.
Well firstly, the "50% winrate for equal skilled players" is pretty much impossible to measure because what really is equally skilled players. MMR is not an accurate measure of skill, especially not when we take into account that you can have one really strong matchup. The argument if zerg is op at the higest level or if many zerg players in general are just higher level that players of other races is an argument you simply cant disprove. Even if Zerg has a higher winrate overall it can be due to skill disparity, in theory. Secondly in that case what is even the point in discussing balance. By your definiton balance simply doesn't matter. If ZvT has 50% winrate 100% due to drone rush then that matchup will soon have a winrate way below 50% because terran players will learn to defend drone rush. When looking at the state of the game right now it is actually important to look at the state of the game longterm, it is fine if the winrates are bad right now if we can reasonably assume they will stabilize down the road. Equally if the winrates are perfectly balanced right now because of drone rush we should definitely discuss right now what we can do to fix the problem that is game state/entertainement but also the coming balance problem. Show nested quote +On September 22 2020 18:13 mounteast0 wrote: The elephant of the room where almost no one talk about is the required sample size (power of calculation if you like). We do not have an agreement on how many (or what) games/ series / tournaments are required to determine if there is imbalance or not, nor any scientific / statistic analysis to determine the (minimum) number required. We have different people using different method to justify whether there is, or there is not imbalance, but ultimately, they reverse engineer the analysis after they formed the conclusion.
Yeah, sounds really important to find some kind of universal measurement of useless data.
Two things here 1. In relation to the drone rush example if everybody started doing it the overwall winrate would go up or down depending on it's success rate and you would be able to see the imbalance of a broken strat (after a period of time) 2. it's impossible to pull 2 players of the same skill as current ELO is based on the current version of SC2 which may or may not be imbalanced already
As mentioned earlier i think some people that state you have to look at balance long term are correct and after thinking about it i actually had a semi decent idea to consider the "balance" of a race.
You take a set of tournaments (you would have to define which, say premier for now) and pull the winners over a period say of 6 months (meaning 2 balance patches a year) and then you tot up the race of each winner however you only count a point for each individual player in that race. so for example over a course of 6 months you have Maru GSL x2 Dark x1 GSL Serral X2 WCS Reynor X1 WCS Scarlett x2 WCS Astrea X1 WCS
You would pull this data and give each race a point for each individual player that won a tournament. If a certain race was winning more events across the board that would suggest imbalance. In this example Protoss would have 1 point. Terran 1 point and Zerg 4 points. Zerg would need looking at as their overall win ratio is +3 over the rest of the group. (I would say +1 is insignificant.)
I would expect each race to have around the same points if balance was equal and each race had an equally skilled pool of players. This means tournament winnings would then be varied by either a small pool of similarly skilled players. My theory is that if a singular race has an advantage and with equally skilled players the amount of winners each race should produce is equal. If however one race has more winners (more players winning with that race) then there would be imbalance as it would not be just skill at that point but some outside factor helping those players win.
This method also allows me to account for superstars. You could ignore a +1 result for an individual race as it would allow you to account for a single dominating player who is just better than everybody else (think Maru/Serrals insane periods). Either that or a statistical outlier (and in science you always have to have a ratio that is considered significant/insignificant and significance is based on being significantly higher above your expected results).
edit: formatting
|
Czech Republic12116 Posts
On September 22 2020 18:20 mounteast0 wrote:Maru who got 1 RO8 and RO16 finish are ranking higher than TY who got 1 win and 1 RO4 finish, or Innovation who got 1 RO4 and 1 RO8 finish. This is totally accurate Maru also won the 2020 GSL ST1. You're totally inaccurate.
|
On September 22 2020 19:49 La1 wrote:
Two things here 1. In relation to the drone rush example if everybody started doing it the overwall winrate would go up or down depending on it's success rate and you would be able to see the imbalance of a broken strat (after a period of time) 2. it's impossible to pull 2 players of the same skill as current ELO is based on the current version of SC2 which may or may not be imbalanced already
As mentioned earlier i think some people that state you have to look at balance long term are correct and after thinking about it i actually had a semi decent idea to consider the "balance" of a race.
You take a set of tournaments (you would have to define which, say premier for now) and pull the winners over a period say of 6 months (meaning 2 balance patches a year) and then you tot up the race of each winner however you only count a point for each individual player in that race. so for example over a course of 6 months you have Maru GSL x2 Dark x1 GSL Serral X2 WCS Reynor X1 WCS Scarlett x2 WCS Astrea X1 WCS
You would pull this data and give each race a point for each individual player that won a tournament. If a certain race was winning more events across the board that would suggest imbalance. In this example Protoss would have 1 point. Terran 1 point and Zerg 4 points. Zerg would need looking at as their overall win ratio is +3 over the rest of the group. (I would say +1 is insignificant.)
I would expect each race to have around the same points if balance was equal and each race had an equally skilled pool of players. This means tournament winnings would then be varied by either a small pool of similarly skilled players. My theory is that if a singular race has an advantage and with equally skilled players the amount of winners each race should produce is equal. If however one race has more winners (more players winning with that race) then there would be imbalance as it would not be just skill at that point but some outside factor helping those players win.
This method also allows me to account for superstars. You could ignore a +1 result for an individual race as it would allow you to account for a single dominating player who is just better than everybody else (think Maru/Serrals insane periods). Either that or a statistical outlier (and in science you always have to have a ratio that is considered significant/insignificant and significance is based on being significantly higher above your expected results).
edit: formatting I must have totally misunderstood your previous post. I don't really know what to say because from some perspectives you seem to take the scientific quantitative approach while at the same time looking at the least amount of data we have, actual champions.
While I myself at times likes discussing race disparity in champions it is mostly from the angle of how I believe different formats of tournaments favor different races.
If I would seriously try to make some kind of sense of the actual balance in the game at the highest level I think I would look at RO16 or RO8 results.
In my opinion it more interesting which races actually went to the finals than who actually won. When you look at the amount of factors that decide who wins a championship there is way too much clutter to believe that actually represents any kind of foundation to discuss balance. Lets say the finals is TvZ, the zerg played a BO5 that went all the way against another terran in the semis while the terran hasn't played a TvZ with high stakes since the previous GSL. The amount of prep advantage the terran has is insane, sure the zerg can do mindgames based on that but still. Going into the match the terran has an advantage, how is it with practise partners? There is so much that decides the outcome of a finals outside of skill and race
Is the reason the terran won the championship against that zerg really balance in any way? Even if you look at the results of 50 finals the number is way too small to believe that its quantity is so large as to make up for the other outside factors besides race.
Generally today SC2 is too small, the amount of pros are too few, the amount of high stakes games are too low that it seems to me not even possible to ever have enough quantity of data to draw any meaningful conclusion on balance at the top level based on purely numerical results.
|
Czech Republic12116 Posts
On September 22 2020 21:17 Shuffleblade wrote:Show nested quote +On September 22 2020 19:49 La1 wrote:
Two things here 1. In relation to the drone rush example if everybody started doing it the overwall winrate would go up or down depending on it's success rate and you would be able to see the imbalance of a broken strat (after a period of time) 2. it's impossible to pull 2 players of the same skill as current ELO is based on the current version of SC2 which may or may not be imbalanced already
As mentioned earlier i think some people that state you have to look at balance long term are correct and after thinking about it i actually had a semi decent idea to consider the "balance" of a race.
You take a set of tournaments (you would have to define which, say premier for now) and pull the winners over a period say of 6 months (meaning 2 balance patches a year) and then you tot up the race of each winner however you only count a point for each individual player in that race. so for example over a course of 6 months you have Maru GSL x2 Dark x1 GSL Serral X2 WCS Reynor X1 WCS Scarlett x2 WCS Astrea X1 WCS
You would pull this data and give each race a point for each individual player that won a tournament. If a certain race was winning more events across the board that would suggest imbalance. In this example Protoss would have 1 point. Terran 1 point and Zerg 4 points. Zerg would need looking at as their overall win ratio is +3 over the rest of the group. (I would say +1 is insignificant.)
I would expect each race to have around the same points if balance was equal and each race had an equally skilled pool of players. This means tournament winnings would then be varied by either a small pool of similarly skilled players. My theory is that if a singular race has an advantage and with equally skilled players the amount of winners each race should produce is equal. If however one race has more winners (more players winning with that race) then there would be imbalance as it would not be just skill at that point but some outside factor helping those players win.
This method also allows me to account for superstars. You could ignore a +1 result for an individual race as it would allow you to account for a single dominating player who is just better than everybody else (think Maru/Serrals insane periods). Either that or a statistical outlier (and in science you always have to have a ratio that is considered significant/insignificant and significance is based on being significantly higher above your expected results).
edit: formatting I must have totally misunderstood your previous post. I don't really know what to say because from some perspectives you seem to take the scientific quantitative approach while at the same time looking at the least amount of data we have, actual champions. While I myself at times likes discussing race disparity in champions it is mostly from the angle of how I believe different formats of tournaments favor different races. If I would seriously try to make some kind of sense of the actual balance in the game at the highest level I think I would look at RO16 or RO8 results. In my opinion it more interesting which races actually went to the finals than who actually won. When you look at the amount of factors that decide who wins a championship there is way too much clutter to believe that actually represents any kind of foundation to discuss balance. Lets say the finals is TvZ, the zerg played a BO5 that went all the way against another terran in the semis while the terran hasn't played a TvZ with high stakes since the previous GSL. The amount of prep advantage the terran has is insane, sure the zerg can do mindgames based on that but still. Going into the match the terran has an advantage, how is it with practise partners? There is so much that decides the outcome of a finals outside of skill and race Is the reason the terran won the championship against that zerg really balance in any way? Even if you look at the results of 50 finals the number is way too small to believe that its quantity is so large as to make up for the other outside factors besides race. Generally today SC2 is too small, the amount of pros are too few, the amount of high stakes games are too low that it seems to me not even possible to ever have enough quantity of data to draw any meaningful conclusion on balance at the top level based on purely numerical results. Well, what if numbers are fine but we still have more zerg wins than terrans and protoss combined? What then? This isn't about game balancing en masse, because the advantage which zerg provides can be used by few not many. (e.g. absence of ZvZ finals)
|
So essentially, T<P<Z<T
Perfect rotational balance?
I kind of wonder if anyone has TvP, PvZ or ZvT as their best matchups...
|
On September 22 2020 22:21 Slydie wrote: So essentially, T<P<Z<T
Perfect rotational balance?
I kind of wonder if anyone has TvP, PvZ or ZvT as their best matchups...
Isn't that what happened in BW at the top? Haven't watched in ages
|
On September 22 2020 22:12 deacon.frost wrote: Well, what if numbers are fine but we still have more zerg wins than terrans and protoss combined? What then? This isn't about game balancing en masse, because the advantage which zerg provides can be used by few not many. (e.g. absence of ZvZ finals) You mean we have more zerg champions than terran and protoss champions combined?
To be honest I am still not convinced we have a problem, you mention that zerg has an advantage, can you prove it?
Do we know that there are more zerg champions than terran and protoss players because zerg has an advantage and not because of various other factors?
If you want to talk about numbers and statistics, this low amount is statistically insignificant. Scientifically we have no basis for drawing any conclusions. Especially when we take into account how many factors that effects the outcome of a finals, the results are not found in a vacuum were only skill and race decide the outcome.
With that said my personal opinion is that balance overtime will come to favour the race that gains the most advantage from macro (zerg). The more a game or patch ripens the the more the early games gets figured out and the more important macro becomes. I also believe, which I have written about many times that zerg is the race generally favored by weekend tournaments due to how terrans and protoss builds rely on doing damage. When an opponent knows your prefered and well practised builds the likelihood that you can succeed in doing economic damage shrinks.
If you look at the stats we really don't see that zerg is overall favored to win championships, if you look for example at prep tournaments in lotv (GSL and SSL) it looks like this: P: 12 finals, 3 championships T: 12 finals, 9 championships Z: 9 finals, 5 championships
As I have already written I don't think you can draw any conclusions from this small sample size but you can at least say that these numbers surely doesn't support the argument that zerg ha any advantage. Not in GSL style tournaments at least, so here comes my usual question: When we are discussing balance at the pro level, are we talking about pro level balance in weekend tournaments? Prep style tournaments? Or the very top of GM ladder?
Because it seems to me that the results supports (even if the sample size is low) that these three different formats favour different races.
|
Czech Republic12116 Posts
On September 22 2020 22:55 Shuffleblade wrote:Show nested quote +On September 22 2020 22:12 deacon.frost wrote: Well, what if numbers are fine but we still have more zerg wins than terrans and protoss combined? What then? This isn't about game balancing en masse, because the advantage which zerg provides can be used by few not many. (e.g. absence of ZvZ finals) You mean we have more zerg champions than terran and protoss champions combined? To be honest I am still not convinced we have a problem, you mention that zerg has an advantage, can you prove it? Do we know that there are more zerg champions than terran and protoss players because zerg has an advantage and not because of various other factors? If you want to talk about numbers and statistics, this low amount is statistically insignificant. Scientifically we have no basis for drawing any conclusions. Especially when we take into account how many factors that effects the outcome of a finals, the results are not found in a vacuum were only skill and race decide the outcome. With that said my personal opinion is that balance overtime will come to favour the race that gains the most advantage from macro (zerg). The more a game or patch ripens the the more the early games gets figured out and the more important macro becomes. I also believe, which I have written about many times that zerg is the race generally favored by weekend tournaments due to how terrans and protoss builds rely on doing damage. When an opponent knows your prefered and well practised builds the likelihood that you can succeed in doing economic damage shrinks. If you look at the stats we really don't see that zerg is overall favored to win championships, if you look for example at prep tournaments in lotv (GSL and SSL) it looks like this: P: 12 finals, 3 championships T: 12 finals, 9 championships Z: 9 finals, 5 championships As I have already written I don't think you can draw any conclusions from this small sample size but you can at least say that these numbers surely doesn't support the argument that zerg ha any advantage. Not in GSL style tournaments at least, so here comes my usual question: When we are discussing balance at the pro level, are we talking about pro level balance in weekend tournaments? Prep style tournaments? Or the very top of GM ladder? Because it seems to me that the results supports (even if the sample size is low) that these three different formats favour different races. Premiere tournaments 2018 - 2020 2020: 9 Z Winners, 1 Z WC title, 3 runner ups 2 T, 0, 3 1 P, 0, 6 combined numbers: P - T - Z = 7 - 5 - 12, 1 Z WC title 9 titles > 3, 12 = 12; Some argue to not count NA, which would actually make the case worse
2019: Z: 11, 2, 10 T: 3, 0, 3 P: 3, 0, 4 combined numbers: P - T - Z = 7 - 6 - 21 11 titles > 6 combined, 21 total > 13 combined.
2018: Z: 9, 2, 3 T: 4, 0, 2 P: 2, 0, 10 combined numbers: P - T - Z = 12 - 6- 12 9 titles > 6 combined, but at least 12 < 18! Funny thing, all 4 titles were Maru.
I usually call this out from the Rogues title in 2017 as this was his ascend to the patchzerg title awarded by TL. (which would add 1 Z, 1 P champs, 1 Z WC title and 2 Z runner ups)
We have 2 and 3/4 years with the current statistics: 5 Z WC titles, 0 for other races Titles: P - T - Z = 6 - 9 - 29 29 > combined 15, zerg won in this time span 65 % of all tournaments COmbined: P - T - Z = 26 - 17 - 45 45 > 43. Out of 88 possible players we have 51 % zergs and 49 % are other races combined.
Out of the 9 epic Terran titles 5 were done by Maru (55 %) Out of the 29 Zerg titles 11 were done by Serral(38 %), 5 Reynor, 4 Rogue & 3 Dark. All 6 titles of Protoss are equally distributed between Stats, Classic and Neeb. (which means that with the Classic departure Protosses lost 33 % of their championship potential )
Either after the Rogues WC title @2017 Blizzcon zergs suddenly became godlike or other races died out. Stats is still around, Maru is still around. Yet Zerg wins 65 % of the tournaments and is majority in the finals. Please, explain this in any other way than game state. Because before the Patchzerg the game wasn't as heavily zerg dominated as is now.
edit> counted premiere tournaments and considering we're discussing victors not the game balance per se I didn't check RO8s. Also if you discount NA then Protosses are left in an abysmal state which we knew already when Classic retired. Edit 2> obviously I can make errors
edit3> Some fun facts 1) If you want to win titles, be zerg. 2) If you want to win titles as Terran, be Maru. 3) if you want to win 2nd places, be Protoss, the Kong race 4) terrans either win or not try to play at all 46 % of all 2nd places belongs to Protoss, 36 % to Zerg, 18 % Terran. (Protoss was closer to rounding up)
edit 4: 2017: 17 premiere tournaments: 4 z, 1 WC Z, 10 runner ups 6 t, 1 WC T, 3 7 p, 0, 4 P - T - Z = 7 - 6 - 4 (no race has more than half victories, or half) combined PTZ = 11 - 9 - 14 (no race is dominant either)
|
France12466 Posts
What I really wonder though, is if the theory/narrative that Blizzard keeps zerg strong on purpose because they noticed foreigners were more likely to win ZvZ vs koreans than PvP or TvT (except Neeb 2017?) so there are more foreigners doing well / more viewership because of that, is actually true, or just a coincidence. Maybe it's because zerg is indeed more difficult to balance than the other races? 2019 was really awful in terms of balance but thankfully 2020 looks a bit better, although it still seems safe to bet on a zerg winning a big event every time to come out with >50% right predictions.
Regarding aligulac being accurate or not, it kinda is in the same region, but there are of course multiple issues: some players don't play "online" tournaments (as in online even before covid) as much as others, for example Maru, and the mirror elo affects the other ELOs as well as the global ELO of a player, which had more of an effect in PvP for example (with most protoss players having less PvP ELO than zergs and terrans have ZvZ / TvT elo). And the elephant in the room is also that there are a lot less good players in Korea right now, therefore almost all regular Code S players manage to qualify season after season, and there aren't for example enough zergs in Korea to flood Code S with it, regardless of any perceived or real imbalance. Plus the nature of the game makes it so that top players find ways to beat inferior players most of the time, except if the imbalance is extreme enough in the match-up or in a strat (akin to what Mizen said, unless you buff zerg ridiculously, there is very little hope that you'd see Impact or Armani win a GSL, even if you were to make them play only ZvP/ZvT to avoid them losing to superior ZvZ players...)
The good thing in my mind about GSL though, is that its format kinda balance things out so that balance isn't as big of a problem as in weekend style tournament, so the viewer experience isn't as bad as for example the ro8/ro4 ZvZ / couple ZvP days in foreign WCS back in 2018/2019.
|
Czech Republic12116 Posts
What i find really weird is that the more you look at the whole tournaments the less the zerg dominance appears. But once you focus on titles it's huge. Other races cannot combine enough titles to even compare themselves to Zergs combined!!! COMBINED. And this didn't happen before 2017 and you cannot say it's Serral because it's not. Even if you remove Serral's titles(11) you get 18 Zerg titles, 4 WC titles > 15 combined titles and 0 combined WC titles. It's insane.
|
France12466 Posts
On September 23 2020 00:00 deacon.frost wrote: What i find really weird is that the more you look at the whole tournaments the less the zerg dominance appears. But once you focus on titles it's huge. Other races cannot combine enough titles to even compare themselves to Zergs combined!!! COMBINED. And this didn't happen before 2017 and you cannot say it's Serral because it's not. Even if you remove Serral's titles(11) you get 18 Zerg titles, 4 WC titles > 15 combined titles and 0 combined WC titles. It's insane. That is kinda a good thing, in the sense that in order to exploit the perceived or real imbalance, you need to be sufficiently at the top. That prevents the patchzerg effect that saw mediocre / unknown players in 2012 to beat favorites (johnnyrecco, miniraser, etc.). On a recent note, it seems very difficult for both protoss and terran players to beat good zergs opponents in a row. Like, Trap almost beat Reynor but he didn't have to beat a good zerg for the whole tournament until the finals. What would have happened if he meet some on his way? I bet the finals would have looked more one sided. Same for Clem, he manages to beat Reynor in DH EU, but can't beat him the next day (although that can happen too to zergs for example Serral beating Reynor in winners but losing to him in grand finals in the same vein as Clem... so it's hard to distinguish from the obvious: it's hard to beat top players in a row, regardless of the race). He manages to beat Serral in the main event, but then can't beat Reynor in the following match. And doesn't manage to beat Serral again in the ESL Pro Cup #37 finals.
|
Czech Republic12116 Posts
On September 23 2020 00:08 Poopi wrote:Show nested quote +On September 23 2020 00:00 deacon.frost wrote: What i find really weird is that the more you look at the whole tournaments the less the zerg dominance appears. But once you focus on titles it's huge. Other races cannot combine enough titles to even compare themselves to Zergs combined!!! COMBINED. And this didn't happen before 2017 and you cannot say it's Serral because it's not. Even if you remove Serral's titles(11) you get 18 Zerg titles, 4 WC titles > 15 combined titles and 0 combined WC titles. It's insane. That is kinda a good thing, in the sense that in order to exploit the perceived or real imbalance, you need to be sufficiently at the top. That prevents the patchzerg effect that saw mediocre / unknown players in 2012 to beat favorites (johnnyrecco, miniraser, etc.). On a recent note, it seems very difficult for both protoss and terran players to beat good zergs opponents in a row. Like, Trap almost beat Reynor but he didn't have to beat a good zerg for the whole tournament until the finals. What would have happened if he meet some on his way? I bet the finals would have looked more one sided. Same for Clem, he manages to beat Reynor in DH EU, but can't beat him the next day (although that can happen too to zergs for example Serral beating Reynor in winners but losing to him in grand finals in the same vein as Clem... so it's hard to distinguish from the obvious: it's hard to beat top players in a row, regardless of the race). He manages to beat Serral in the main event, but then can't beat Reynor in the following match. And doesn't manage to beat Serral again in the ESL Pro Cup #37 finals. Seems so. Also it's quite disturbing that there are so few Terran names, if any of them ends it's a big disaster.
|
On September 22 2020 23:36 deacon.frost wrote:Show nested quote +On September 22 2020 22:55 Shuffleblade wrote:On September 22 2020 22:12 deacon.frost wrote: Well, what if numbers are fine but we still have more zerg wins than terrans and protoss combined? What then? This isn't about game balancing en masse, because the advantage which zerg provides can be used by few not many. (e.g. absence of ZvZ finals) You mean we have more zerg champions than terran and protoss champions combined? To be honest I am still not convinced we have a problem, you mention that zerg has an advantage, can you prove it? Do we know that there are more zerg champions than terran and protoss players because zerg has an advantage and not because of various other factors? If you want to talk about numbers and statistics, this low amount is statistically insignificant. Scientifically we have no basis for drawing any conclusions. Especially when we take into account how many factors that effects the outcome of a finals, the results are not found in a vacuum were only skill and race decide the outcome. With that said my personal opinion is that balance overtime will come to favour the race that gains the most advantage from macro (zerg). The more a game or patch ripens the the more the early games gets figured out and the more important macro becomes. I also believe, which I have written about many times that zerg is the race generally favored by weekend tournaments due to how terrans and protoss builds rely on doing damage. When an opponent knows your prefered and well practised builds the likelihood that you can succeed in doing economic damage shrinks. If you look at the stats we really don't see that zerg is overall favored to win championships, if you look for example at prep tournaments in lotv (GSL and SSL) it looks like this: P: 12 finals, 3 championships T: 12 finals, 9 championships Z: 9 finals, 5 championships As I have already written I don't think you can draw any conclusions from this small sample size but you can at least say that these numbers surely doesn't support the argument that zerg ha any advantage. Not in GSL style tournaments at least, so here comes my usual question: When we are discussing balance at the pro level, are we talking about pro level balance in weekend tournaments? Prep style tournaments? Or the very top of GM ladder? Because it seems to me that the results supports (even if the sample size is low) that these three different formats favour different races. Premiere tournaments 2018 - 2020 2020: 9 Z Winners, 1 Z WC title, 3 runner ups 2 T, 0, 3 1 P, 0, 6 combined numbers: P - T - Z = 7 - 5 - 12, 1 Z WC title 9 titles > 3, 12 = 12; Some argue to not count NA, which would actually make the case worse 2019: Z: 11, 2, 10 T: 3, 0, 3 P: 3, 0, 4 combined numbers: P - T - Z = 7 - 6 - 21 11 titles > 6 combined, 21 total > 13 combined. 2018: Z: 9, 2, 3 T: 4, 0, 2 P: 2, 0, 10 combined numbers: P - T - Z = 12 - 6- 12 9 titles > 6 combined, but at least 12 < 18! Funny thing, all 4 titles were Maru. I usually call this out from the Rogues title in 2017 as this was his ascend to the patchzerg title awarded by TL. (which would add 1 Z, 1 P champs, 1 Z WC title and 2 Z runner ups) We have 2 and 3/4 years with the current statistics: 5 Z WC titles, 0 for other races Titles: P - T - Z = 6 - 9 - 29 29 > combined 15, zerg won in this time span 65 % of all tournaments COmbined: P - T - Z = 26 - 17 - 45 45 > 43. Out of 88 possible players we have 51 % zergs and 49 % are other races combined. Out of the 9 epic Terran titles 5 were done by Maru (55 %) Out of the 29 Zerg titles 11 were done by Serral(38 %), 5 Reynor, 4 Rogue & 3 Dark. All 6 titles of Protoss are equally distributed between Stats, Classic and Neeb. (which means that with the Classic departure Protosses lost 33 % of their championship potential ) Either after the Rogues WC title @2017 Blizzcon zergs suddenly became godlike or other races died out. Stats is still around, Maru is still around. Yet Zerg wins 65 % of the tournaments and is majority in the finals. Please, explain this in any other way than game state. Because before the Patchzerg the game wasn't as heavily zerg dominated as is now. edit> counted premiere tournaments and considering we're discussing victors not the game balance per se I didn't check RO8s. Also if you discount NA then Protosses are left in an abysmal state which we knew already when Classic retired. Edit 2> obviously I can make errors edit3> Some fun facts 1) If you want to win titles, be zerg. 2) If you want to win titles as Terran, be Maru. 3) if you want to win 2nd places, be Protoss, the Kong race 4) terrans either win or not try to play at all 46 % of all 2nd places belongs to Protoss, 36 % to Zerg, 18 % Terran. (Protoss was closer to rounding up) edit 4: 2017: 17 premiere tournaments: 4 z, 1 WC Z, 10 runner ups 6 t, 1 WC T, 3 7 p, 0, 4 P - T - Z = 7 - 6 - 4 (no race has more than half victories, or half) combined PTZ = 11 - 9 - 14 (no race is dominant either) Whoa, nice work putting together all those numbers! O_O
Interesting indeed, I will still say however that it proves nothing but it does show an indication of something.
As I wrote earlier in my rant about how weekender tournament style favors zerg, I also think that it does so even more the longer the tournament goes. Unless the T/P got lucky and ran a whole bracket without facing zerg. You show that zerg seems to win a lot more weekender tournaments that other races, I showed that zergs does not tend to win more prep style tournaments. Which indicates that the "problem" might be in the format of the tournaments and not that one race is generally overpowered if you look at the game from a ladder perspective.
MVP and Maru has shown us that there are players that are just that much better than other professional players that it says nothing about game balance when they win, Imagine for a moment that Reynor and Serral is actually much better than everyone else, so much better that their wins are an anomaly. They just win because they are the best and it has nothing to do with balance, if we play with the thought of that being true than your numbers doesn't look as bad anymore. T: 4 Z: 13 P: 6
The numbers are still more than the other races combined for zerg, but with this small sample size its not big enough to rule out statistical variance.
As I wrote before though, my personal opinion is still that zerg is stronger than the other races due to how their macro mechanics work and that they are generally favored in weekender tournaments but I don't think we have proof of that.
|
Czech Republic12116 Posts
On September 23 2020 00:14 Shuffleblade wrote:Show nested quote +On September 22 2020 23:36 deacon.frost wrote:On September 22 2020 22:55 Shuffleblade wrote:On September 22 2020 22:12 deacon.frost wrote: Well, what if numbers are fine but we still have more zerg wins than terrans and protoss combined? What then? This isn't about game balancing en masse, because the advantage which zerg provides can be used by few not many. (e.g. absence of ZvZ finals) You mean we have more zerg champions than terran and protoss champions combined? To be honest I am still not convinced we have a problem, you mention that zerg has an advantage, can you prove it? Do we know that there are more zerg champions than terran and protoss players because zerg has an advantage and not because of various other factors? If you want to talk about numbers and statistics, this low amount is statistically insignificant. Scientifically we have no basis for drawing any conclusions. Especially when we take into account how many factors that effects the outcome of a finals, the results are not found in a vacuum were only skill and race decide the outcome. With that said my personal opinion is that balance overtime will come to favour the race that gains the most advantage from macro (zerg). The more a game or patch ripens the the more the early games gets figured out and the more important macro becomes. I also believe, which I have written about many times that zerg is the race generally favored by weekend tournaments due to how terrans and protoss builds rely on doing damage. When an opponent knows your prefered and well practised builds the likelihood that you can succeed in doing economic damage shrinks. If you look at the stats we really don't see that zerg is overall favored to win championships, if you look for example at prep tournaments in lotv (GSL and SSL) it looks like this: P: 12 finals, 3 championships T: 12 finals, 9 championships Z: 9 finals, 5 championships As I have already written I don't think you can draw any conclusions from this small sample size but you can at least say that these numbers surely doesn't support the argument that zerg ha any advantage. Not in GSL style tournaments at least, so here comes my usual question: When we are discussing balance at the pro level, are we talking about pro level balance in weekend tournaments? Prep style tournaments? Or the very top of GM ladder? Because it seems to me that the results supports (even if the sample size is low) that these three different formats favour different races. Premiere tournaments 2018 - 2020 2020: 9 Z Winners, 1 Z WC title, 3 runner ups 2 T, 0, 3 1 P, 0, 6 combined numbers: P - T - Z = 7 - 5 - 12, 1 Z WC title 9 titles > 3, 12 = 12; Some argue to not count NA, which would actually make the case worse 2019: Z: 11, 2, 10 T: 3, 0, 3 P: 3, 0, 4 combined numbers: P - T - Z = 7 - 6 - 21 11 titles > 6 combined, 21 total > 13 combined. 2018: Z: 9, 2, 3 T: 4, 0, 2 P: 2, 0, 10 combined numbers: P - T - Z = 12 - 6- 12 9 titles > 6 combined, but at least 12 < 18! Funny thing, all 4 titles were Maru. I usually call this out from the Rogues title in 2017 as this was his ascend to the patchzerg title awarded by TL. (which would add 1 Z, 1 P champs, 1 Z WC title and 2 Z runner ups) We have 2 and 3/4 years with the current statistics: 5 Z WC titles, 0 for other races Titles: P - T - Z = 6 - 9 - 29 29 > combined 15, zerg won in this time span 65 % of all tournaments COmbined: P - T - Z = 26 - 17 - 45 45 > 43. Out of 88 possible players we have 51 % zergs and 49 % are other races combined. Out of the 9 epic Terran titles 5 were done by Maru (55 %) Out of the 29 Zerg titles 11 were done by Serral(38 %), 5 Reynor, 4 Rogue & 3 Dark. All 6 titles of Protoss are equally distributed between Stats, Classic and Neeb. (which means that with the Classic departure Protosses lost 33 % of their championship potential ) Either after the Rogues WC title @2017 Blizzcon zergs suddenly became godlike or other races died out. Stats is still around, Maru is still around. Yet Zerg wins 65 % of the tournaments and is majority in the finals. Please, explain this in any other way than game state. Because before the Patchzerg the game wasn't as heavily zerg dominated as is now. edit> counted premiere tournaments and considering we're discussing victors not the game balance per se I didn't check RO8s. Also if you discount NA then Protosses are left in an abysmal state which we knew already when Classic retired. Edit 2> obviously I can make errors edit3> Some fun facts 1) If you want to win titles, be zerg. 2) If you want to win titles as Terran, be Maru. 3) if you want to win 2nd places, be Protoss, the Kong race 4) terrans either win or not try to play at all 46 % of all 2nd places belongs to Protoss, 36 % to Zerg, 18 % Terran. (Protoss was closer to rounding up) edit 4: 2017: 17 premiere tournaments: 4 z, 1 WC Z, 10 runner ups 6 t, 1 WC T, 3 7 p, 0, 4 P - T - Z = 7 - 6 - 4 (no race has more than half victories, or half) combined PTZ = 11 - 9 - 14 (no race is dominant either) Whoa, nice work putting together all those numbers! O_O Interesting indeed, I will still say however that it proves nothing but it does show an indication of something. As I wrote earlier in my rant about how weekender tournament style favors zerg, I also think that it does so even more the longer the tournament goes. Unless the T/P got lucky and ran a whole bracket without facing zerg. You show that zerg seems to win a lot more weekender tournaments that other races, I showed that zergs does not tend to win more prep style tournaments. Which indicates that the "problem" might be in the format of the tournaments and not that one race is generally overpowered if you look at the game from a ladder perspective. MVP and Maru has shown us that there are players that are just that much better than other professional players that it says nothing about game balance when they win, Imagine for a moment that Reynor and Serral is actually much better than everyone else, so much better that their wins are an anomaly. They just win because they are the best and it has nothing to do with balance, if we play with the thought of that being true than your numbers doesn't look as bad anymore. T: 4 Z: 13 P: 6 The numbers are still more than the other races combined for zerg, but with this small sample size its not big enough to rule out statistical variance. As I wrote before though, my personal opinion is still that zerg is stronger than the other races due to how their macro mechanics work and that they are generally favored in weekender tournaments but I don't think we have proof of that. I'm not saying it proves anything, but it shows a big problem. We have few good Terrans, we don't have any Protoss champions, but we have plenty of Zerg Champions while the general balance stays more or less fine. I have no idea why it's happening and I blame Canada Zerg, because it\s the common denominator. This may turn into the snowball effect - hey, zergs are winning, let's switch to zerg. This effect ruined many servers in MMOs where people started to switch to the more dominant side leaving the people behind in even worse state than before. IMO Blizzard won't solve this because this would require a big analysis and nothing points they are willing to do that. (also I believe that if Trap wouldn't played the Kong race he would have won over Reynor, but sadly Protoss can't win titles )
|
On September 22 2020 21:17 Shuffleblade wrote:
I must have totally misunderstood your previous post. I don't really know what to say because from some perspectives you seem to take the scientific quantitative approach while at the same time looking at the least amount of data we have, actual champions.
While I myself at times likes discussing race disparity in champions it is mostly from the angle of how I believe different formats of tournaments favor different races.
If I would seriously try to make some kind of sense of the actual balance in the game at the highest level I think I would look at RO16 or RO8 results.
In my opinion it more interesting which races actually went to the finals than who actually won. When you look at the amount of factors that decide who wins a championship there is way too much clutter to believe that actually represents any kind of foundation to discuss balance. Lets say the finals is TvZ, the zerg played a BO5 that went all the way against another terran in the semis while the terran hasn't played a TvZ with high stakes since the previous GSL. The amount of prep advantage the terran has is insane, sure the zerg can do mindgames based on that but still. Going into the match the terran has an advantage, how is it with practise partners? There is so much that decides the outcome of a finals outside of skill and race
You have, you are not looking at the big picture. Your saying i have taken a small amount of data when actually the champion covers ALL data in the tournament as he has to win matches across the whole event and all the micro-advantages cascade over the course of the tournament. The players that win tournaments have to have 1 of 3 things 1. An insane amount of luck (lucky run etc) 2. they are much better than everybody else 3. they have an outside advantage (race etc..)
If they win by luck or are much better i have accounted for both of these in my previous post (by significance and only counting individual players). By actually including a wide breath of players (for example top 16 or top 8) you are increasing the volatility from both items 1 and 2 (luck and better players) especially as there is a depth issue when it comes to really good players.
Let me put it another way. If player A matches vs equal skill opponents for a tournament for the whole bracket (5 matches RO32,RO16 etc..) however his race is worse (say by 5%) then in each match he players he has a 45% chance to win.
Say he wins his first 2 matches, eventually that 55/45 split will catch up and the more matches he plays and the longer the tournament goes on he will eventually lose the 55/45 split so he might get to RO8 or RO4. It would be very rare for him to win.
In terms of the dataset you could include the finalists but anything more increases variance.
On September 22 2020 21:17 Shuffleblade wrote:
Is the reason the terran won the championship against that zerg really balance in any way? Even if you look at the results of 50 finals the number is way too small to believe that its quantity is so large as to make up for the other outside factors besides race.
Generally today SC2 is too small, the amount of pros are too few, the amount of high stakes games are too low that it seems to me not even possible to ever have enough quantity of data to draw any meaningful conclusion on balance at the top level based on purely numerical results.
Again. I am not looking at the final game. the winner of an event may play in 5 or 6 matches which act as filters. I think you could quite easily judge balance on amount of tournament wins by a race/player denominator. It would not give you an accurate number on exactly how balanced the game is but a general idea of the direction. You would only need 3-6 months of data from the previous patch to work it out.
edit: To all the people talking about the numbers and stats above from 2017 please stop. The game is far different from 2017 to now (Observers got made slow ). You should probably look at tournament wins from say the last patch or the last 2 patches.
the last major balance change was 09Jun2020 Since then we have had 4 premier tournaments 4 Zerg wins with 3 different Z players Using my theory this would suggest Z is slightly overpowered vs both terran and protoss who are equal. If you expanded on this and looked at finalists it would read:
4 Zerg wins (via 3 players) 3 Protoss 2nd's (via 2 players) 1 Terran 2nd this would suggest protoss is slightly better than terran but Z is still OP.
edit more data for people who want further variability - going back 3 patches (last patch 4.11.3 dated 17Dec2019)
9 Z wins (over 5 players) 2 T wins (over 2 players) 1 P win 2nd's 3 Z 2nds (over 2 players (1 in the 1st place list) 6 P 2nds (over 4 players) (0 in main list) 3 T 2nds (over 3 players (0 in main list)
Total scores - Z- 6, T-5 P-5 - Zerg a tiny bit ahead but not significant - P/T equal.
I would actually give 2nd place a 0.5 rating instead of 1 point as i think 1st place is more valuable. this would make the scores Z - 5.5 , T-3.5, P - 3 - Which shows P/T slightly equal but a slight imbalance towards Z
|
On September 23 2020 00:49 La1 wrote: You have, you are not looking at the big picture. Your saying i have taken a small amount of data when actually the champion covers ALL data in the tournament as he has to win matches across the whole event and all the micro-advantages cascade over the course of the tournament. The players that win tournaments have to have 1 of 3 things 1. An insane amount of luck (lucky run etc) 2. they are much better than everybody else 3. they have an outside advantage (race etc..)
If they win by luck or are much better i have accounted for both of these in my previous post (by significance and only counting individual players). By actually including a wide breath of players (for example top 16 or top 8) you are increasing the volatility from both items 1 and 2 (luck and better players) especially as there is a depth issue when it comes to really good players.
Let me put it another way. If player A matches vs equal skill opponents for a tournament for the whole bracket (5 matches RO32,RO16 etc..) however his race is worse (say by 5%) then in each match he players he has a 45% chance to win.
Say he wins his first 2 matches, eventually that 55/45 split will catch up and the more matches he plays and the longer the tournament goes on he will eventually lose the 55/45 split so he might get to RO8 or RO4. It would be very rare for him to win.
In terms of the dataset you could include the finalists but anything more increases variance. You are the one not looking at the big picture, measuring the champions cover all the data in the tournament? That is just flat out not true, you can tell any story you want with data, if you just angle it your way. You really seem to like to angle things.
The champions fails to cover bracket, opponents, race matchups and race diversity.
If five tournaments in a row a zerg won yet every tournament they were the only zerg to even move on to the RO16 does that mean zerg is OP? I don't think so, I disagree. You talk about win percentage due to the race being worse as if that is the only factor that effect who wins. Sorry but everyone is not equally skilled and even if they were equally skilled there are a lot of variables outside of race balance and skill that determines how a series goes. Yes with big enough of a sample size such variables can be reduced enough so it doesn't matter but you would need a much larger sample size than what you are using, at least ten times as large. It would be great if we had massive data but we simply don't.
You say we only need to measure the champions race to know balance, 2018 Maru won every code S. He won four premier tournaments that year, one of those GSLs he was the only terran in the RO16. There were only two other terran finalists in premier tournaments, Inno once and TY once. Was terran overpowered? Probably not right, probably can't just look at tournament results.
You also interestingly mention that looking at semifinalists would increase the variance too much because there is a depth issue when it comes to really good players. That is interesting, since you only like to look at the top 2 in a given tournament are you saying that the top 8 players are best of the best and anyone below that is so bad as to include them would increase variance? That just simply doesn't make any sense at all, because if there isn't even 16 players good enough to compare between than why are you even talking about balance. If you are saying there are only 8 players good enough to compare between at that point its not about balance but about form. Maru is god like so he cant be counted as terran, Serral is the same. Alright now lets compare how the games go between the remaining six players so you can look at that data and determine the racial balance. It makes zero sense, if there are that few top level players in the first place its not about balance but about form, matchup and all the others things that you like to pretend doesn't exist.
On September 23 2020 00:49 La1 wrote:Again. I am not looking at the final game. the winner of an event may play in 5 or 6 matches which act as filters. I think you could quite easily judge balance on amount of tournament wins by a race/player denominator. It would not give you an accurate number on exactly how balanced the game is but a general idea of the direction. You would only need 3-6 months of data from the previous patch to work it out. edit: To all the people talking about the numbers and stats above from 2017 please stop. The game is far different from 2017 to now (Observers got made slow ). You should probably look at tournament wins from say the last patch or the last 2 patches. the last major balance change was 09Jun2020 Since then we have had 4 premier tournaments 4 Zerg wins with 3 different Z players Using my theory this would suggest Z is slightly overpowered vs both terran and protoss who are equal. If you expanded on this and looked at finalists it would read: 4 Zerg wins (via 3 players) 3 Protoss 2nd's (via 2 players) 1 Terran 2nd this would suggest protoss is slightly better than terran but Z is still OP. edit more data for people who want further variability - going back 3 patches (last patch 4.11.3 dated 17Dec2019) 9 Z wins (over 5 players) 2 T wins (over 2 players) 1 P win 2nd's 3 Z 2nds (over 2 players (1 in the 1st place list) 6 P 2nds (over 4 players) (0 in main list) 3 T 2nds (over 3 players (0 in main list) Total scores - Z- 6, T-5 P-5 - Zerg a tiny bit ahead but not significant - P/T equal. I would actually give 2nd place a 0.5 rating instead of 1 point as i think 1st place is more valuable. this would make the scores Z - 5.5 , T-3.5, P - 3 - Which shows P/T slightly equal but a slight imbalance towards Z Man semifinalist, even RO16 players go through multiple series/games to get there that can be used as "filters". The champion says 0 about game balance, the data simply doesn't support what you are saying.
Lets look at that 2018 Maru and Serral year, Maru won 4 premier tournaments, Serral won 7.
T: Maru Z: Serral, Scarlett, Rogue P: Classic, Stats
If you include second places
T: TY, Innovation Z: Dark, Reynor P: sOs, Has, Mana, Zest, Showtime
Using your 0.5 for the silver medalist 2018 was the year of protoss! Sounds about right doesn't it, you sure got it down.
|
On September 21 2020 16:39 Morbidius wrote: We all know that Maru,Inno,Trap, Stats, Rogue and Dark aren't patch abusers. We've seen them win in multiple different metas, and Trap has bloomed where all flowers wilted. The question is who will get the patchzerg title when all is said and done? Someone is winning their fair share of the pie, but someone has to be a race abuser when a race wins 80% of tournaments.
Who is Life and who is Sniper?
Rogue isnt patch abuser
LULLLLLLLLLLLLLLLLLLLLLLLL
Look at the two eras Rogue won
|
On September 23 2020 03:15 Shuffleblade wrote:
Man semifinalist, even RO16 players go through multiple series/games to get there that can be used as "filters". The champion says 0 about game balance, the data simply doesn't support what you are saying.
Lets look at that 2018 Maru and Serral year, Maru won 4 premier tournaments, Serral won 7.
T: Maru Z: Serral, Scarlett, Rogue P: Classic, Stats
If you include second places
T: TY, Innovation Z: Dark, Reynor P: sOs, Has, Mana, Zest, Showtime
Using your 0.5 for the silver medalist 2018 was the year of protoss! Sounds about right doesn't it, you sure got it down.
Considering this was over at least 5 balance patch updates that i could count (which creates veriance) your probably right. I would suggest that any data pulled is pulled over a certain amount of patches rather than "a year".
If you remember protoss had a lot of really strong mid game pushes with zealots/ immortals and prism drops in 2018. 2018 was the year of serral and maru (clearly the 2 best players) but this would suggest protoss was the strongest race even though tournaments were dominated by 2 very good players.
I can't remember when protoss got massively nerfed (zealots lost dmg on charge, warp prism got butchered) but they were causing zerg loads of problems before those changes got implemented and the change in results from 2018 to 2020 shows this.
you seem to think i have it out for a single race when actually i don't. I also wasn't trying to insult you by saying you need to look at the bigger picture so sorry if you felt that I do also agree with you that potentially just looking at the champion is shortsighted but i would not move below top 4 (or even top 2) as otherwise i think player skill becomes a factor. +1 to you sir!
in both your other examples, the top z player winning 5 tournaments in a row and the maru example my analysis would hold up In the top z player example he would get 1 point. Everybody from other races would finish 2nd so say 5 different terrans finished second. it would be clear theres an imbalance even though he is clearly just the best player. (this is where i agree with you that the data should be broader) in the Maru example of 2018 it was clear hes just a god and Terran were not overpowered.
it seems we are making progress though! In the interest of finding a balance measurement this thread has been interesting
|
Mexico2169 Posts
To be fair, if the races were perfectly balanced and skill equal, then deciding the winner would be like flipping a coin. In that case you could still have times weere you would get tails 5 or 10 or 30 or 300 times in a row. It's just that if you do that infinetly the spread would be 50/50 (or 1/3 in the case of starcraft). Streaks happen in random sequences and in every game of chance. 1/3 of chances doesn't mean you will get protoss then terran then zerg all the time. You could get one 10 times in a row. But eventually they should converge. So zerg winning so much, as much as I hate to say it, could still not be proof that they are unbalanced.
However I think they are unbalanced, by desing.
This takes me back to an ages old argument, that I feel has proven true.
Zerg has the highest skill cap.
Terran may have the highest micro potential, but as we all know macro>micro, and zerg as a race has the highest skill cap. Not only that, Zerg by design is the most versatile, and has the most ammount of viable unit comps, alongside units that shouldn't belong to zerg (cost and supply effective units like baneling, lurkers, swarmhost and vipers go against the design philosophy of zerg causing balance problems).
This has lead to zerg dominating by versatility and macro mechanics alone. And when Blizzard shifted the game in legacy of the void to put even more importance in macro and number of bases, this benefited zerg the most.
Zerg will always be the strongest race (design wise) unless some big design changes come or they heavily change some units. They can nerf them, they can band-aid, and we can get close to equal winrates like we are now, but there will always be problems until a big change comes. Their design is just superior and some units work too well with it.
|
On September 23 2020 03:47 La1 wrote:Show nested quote +On September 23 2020 03:15 Shuffleblade wrote:
Man semifinalist, even RO16 players go through multiple series/games to get there that can be used as "filters". The champion says 0 about game balance, the data simply doesn't support what you are saying.
Lets look at that 2018 Maru and Serral year, Maru won 4 premier tournaments, Serral won 7.
T: Maru Z: Serral, Scarlett, Rogue P: Classic, Stats
If you include second places
T: TY, Innovation Z: Dark, Reynor P: sOs, Has, Mana, Zest, Showtime
Using your 0.5 for the silver medalist 2018 was the year of protoss! Sounds about right doesn't it, you sure got it down. Considering this was over at least 5 balance patch updates that i could count (which creates veriance) your probably right. I would suggest that any data pulled is pulled over a certain amount of patches rather than "a year". If you remember protoss had a lot of really strong mid game pushes with zealots/ immortals and prism drops in 2018. 2018 was the year of serral and maru (clearly the 2 best players) but this would suggest protoss was the strongest race even though tournaments were dominated by 2 very good players. I can't remember when protoss got massively nerfed (zealots lost dmg on charge, warp prism got butchered) but they were causing zerg loads of problems before those changes got implemented and the change in results from 2018 to 2020 shows this. you seem to think i have it out for a single race when actually i don't. I also wasn't trying to insult you by saying you need to look at the bigger picture so sorry if you felt that I do also agree with you that potentially just looking at the champion is shortsighted but i would not move below top 4 (or even top 2) as otherwise i think player skill becomes a factor. +1 to you sir! in both your other examples, the top z player winning 5 tournaments in a row and the maru example my analysis would hold up In the top z player example he would get 1 point. Everybody from other races would finish 2nd so say 5 different terrans finished second. it would be clear theres an imbalance even though he is clearly just the best player. (this is where i agree with you that the data should be broader) in the Maru example of 2018 it was clear hes just a god and Terran were not overpowered. it seems we are making progress though! In the interest of finding a balance measurement this thread has been interesting Dude you so nice :p
You really seem to think patches really effect racial balance, I believe that patches that actually change the balance are rare. I think it more effects meta and gamestate but overall balance mostly remains the same, the most recent patch effecting balance in my opinion would be from august 2019.
Unless you can point to a specific change that you believe is so pivotal that it actually effects the winrate in a certain matchup distinguishing between patches isn't really necessary. Periods of grave imbalance where a particular strat is OP is rare nowadays.
I think there is one major hole in your argument I want to point out though. aside from what Phantom wrote about probability.
You argue two different perspectives which are kind of contradictory. 1) It is enough to look at the champion because the race of the champion must have gotten that far because of racial imbalance. A lot of matches pile upon eachother to add a "filter" which means looking at the champions also include looking at the matches leading to the championship. You wrote:
On September 23 2020 00:49 La1 wrote: me put it another way. If player A matches vs equal skill opponents for a tournament for the whole bracket (5 matches RO32,RO16 etc..) however his race is worse (say by 5%) then in each match he players he has a 45% chance to win.
Say he wins his first 2 matches, eventually that 55/45 split will catch up and the more matches he plays and the longer the tournament goes on he will eventually lose the 55/45 split so he might get to RO8 or RO4. It would be very rare for him to win.
Here to support your argument that looking at champions is enough you argue that all the games leading up to the finals are equal 50/50 due to equal skill in all matchups and what decides it all is race balance. That is the only way for you to make sense of why looking only at the champion is doing anything more than simply looking at the result of 1 series (the finals).
But you also argue this: 2) You will not look at the results below the finals because the skill disparity between players is too large, even compared to semifinalists.
So you justify that the champion has to go through a lot of even matches to even get to the finals in the first place and thats why looking at the winner makes sense but then you say everyone below the finals are not good enough at the game to be included in your statistics.
I am sure you understand what I am saying here, you pick whichever perspective you want to make your argument make sense but when you look closer it doesn't makes sense. If everyone below the finals are so much worse they will effect the stats if you include them then the finalist going to the finals should be the most likely outcome in the first place right. No matter the balance, you cant eat your cookie and save it at the same time. Either looking at champions doesn't mean much because it is only the result of one series where the path to the finals was a formality or the players in the RO16 are also skilled an worth to keep in the data.
My perspective is, if you want to look at balance, instead of looking at whom out of the few best players in the world wins this time (which is a very small data sample and very dependable on factors besides balance) I think it is more reliable to look at the results of who goes into RO16 and RO8.
Lets put it like this, if few of one race advances into the RO16 two seasons in a row but instead there are many from one particular race, different mid tierplayers moves on instead. Imagine seeing Zoun, Creator and Patience in the RO16 but no terrans? Does that say anything about balance, I would argue that it says more about it that who eventually ends up winning the championship at least. When a lot of underdogs from a particular race starts doing better at the same time that is something that I think is easier to spot and indicates a possible imbalance.
|
|
|
|