|
On August 07 2010 06:55 kajeus wrote:Show nested quote +On August 07 2010 06:52 Logo wrote:On August 07 2010 06:46 kajeus wrote:On August 07 2010 06:29 Logo wrote:On August 07 2010 06:17 kajeus wrote:On August 07 2010 05:22 kGold wrote:
The only defense could be that while there are fewer Zerg players than the other races on Korea, there are more BETTER Zerg players per capita of Zerg players in Korea vs. the other races.
I can believe it because Zerg is the hardest race to play so the beginner-intermediate players will show more of a tendency to stay away from Zerg. Ok, here's the data that is directly relevant to YOUR hypothesis.. The problem with statistical analysis like this is it's not actually based on any sort of anything. If you aren't going to use some actual methods of statistical analysis with math to back it up, just don't even bother. For example you don't even account for the fact that Zerg is a smaller sample size so there's a higher chance that the result will deviate from the expected value. For all we know all three values are within an acceptable deviation from the expected value. Do you want me to do a t-test or something? I mean, I could... or you could. If you accept everything up to the point of demanding one t-test on the difference between 4223 (expected diamond zerg population) and 4059 (skill-adjusted diamond zerg population), then I'm doing ok. I'd like to do a simple binomial test on this difference to see the probability of that difference having popped up by chance. But I gotta read back over an old paper and remember how to do that. And I don't have time for that right now. D: So... I'll try to later. Or you could do it! ((By the way, on your point about ZvT, Dustin Browder has said that high-level ZvT is 50/50.)) Everything is 50/50 with good matchmaking regardless of balance =/. Also browder said that back in early/mid July. And no that's not really the only problem I have with the analysis. I mean the numbers aren't even valuable numbers to datamine. Simply put the data needed to statistically determine any sort of racial imbalance is not available to us as players. Your time would be better spent statistically analyzing tournament results. Why aren't the numbers valuable numbers to datamine? It's census data! And on the 50/50 point, two things: 1) We've established the zergs are well-represented at the top, yet STILL ZvT is 50/50, even though good zergs are playing good terrans; and 2) please give the Blizz staff a little more credit... They're not so stupid as to see 50/50 ZvT diamond results and ignore the possibility that good zergs are playing bad terrans and getting 50/50. And there have been no major nerfs to zerg or buffs to terran since his comment -- he made it AFTER the "ZvT is unfair" cries became so loud. Tournament results seem so much worse. They're extremely anecdotal (there aren't many) and there is no matchmaking component at all. It's not a system that churns out data like the ladder is. Actually, I'd say perceived tournament outcomes CAUSED this perception that TvZ is so imbalanced. People watch qxc turtle Sheth through three waves of ultras and call imba. Later, IdrA destroying everyone in KotB is completely ignored when Silver beats IdrA 2-0. So people are dumb about tournament results, imo. That's the opposite of science.
This is just so wrong it's hard for me to start =/.
Tournaments are the #1 source of data to identify balance issues. While it's a smaller sample size we know that a) we're dealing with the highest caliber of players, b) all of said players are playing to the best of their ability because $ is on the line. Plus it's our biggest source of data of specific match-ups.
Second people selectively viewing tournament results has ABSOLUTELY NOTHING TO DO WITH THE VALUE OF LOOKING AT TOURNAMENT RESULTS. That's just so absurd I can't believe you said it. Just because some bozos decide to only focus on high profile zerg losses doesn't mean that taking an objective look at all tournament results is worthless. I mean is that really the argument you are putting forward?
Lastly your comments on 50/50 make absolutely no sense to me. Do you believe Blizzard is looking at every single game played and deciding that the players are of equal skill? How in the world could they possibly tell from the data that we have that there is a big balance issue?
A lot of the affects of a matchup imbalance are going to be hidden by differences in player skill and the fact that only roughly 1/3rd of the games are played with that matchup. As an example of how much this affects things lets say you play 100 games vs a hypothetical evenly matched blob of players. If you have 50/50 for 2 matchups and 40/60 for one matchup you'd still be expected to win
46 of your 100 games. The fact that it's 1 of 3 possible matchups means the actual affect, even if the imbalance is pretty significant, appears as a very tiny thing. When you factor in how ELO fudges this and drives it more towards 50% it's obvious that you can't just look at the data we have.
Besides the quote starts with, "I have the same feel, but the numbers don't support it..." and he goes on to say, "We are studying the issue and trying to figure out if we should do something..." That seems to imply that they're not sure of the data and need further analysis.
Couldn't agree more, people are only seeing what they want to see. Z is very well represented in tournaments, both in participation numbers and in first or second place finishers. When a Z wins a tournament, people seem not to notice. IdrA sweeping the KotB, etc etc. Hell, last night's GosuCoaching #7 had a ZvZ finals (Sheth vs Slush). I guess when a Z wins it's not noteworthy at all, but when a T wins it's because of imba.
[Citation Needed] Saying it doesn't make it true. If you want to say tournaments well represent each race then do it out and show that they do. If you want to say there is an imbalance in tournament wins by match-up then likewise show it.
|
On August 07 2010 08:34 Logo wrote:Show nested quote +On August 07 2010 06:55 kajeus wrote:On August 07 2010 06:52 Logo wrote:On August 07 2010 06:46 kajeus wrote:On August 07 2010 06:29 Logo wrote:On August 07 2010 06:17 kajeus wrote:On August 07 2010 05:22 kGold wrote:
The only defense could be that while there are fewer Zerg players than the other races on Korea, there are more BETTER Zerg players per capita of Zerg players in Korea vs. the other races.
I can believe it because Zerg is the hardest race to play so the beginner-intermediate players will show more of a tendency to stay away from Zerg. Ok, here's the data that is directly relevant to YOUR hypothesis.. The problem with statistical analysis like this is it's not actually based on any sort of anything. If you aren't going to use some actual methods of statistical analysis with math to back it up, just don't even bother. For example you don't even account for the fact that Zerg is a smaller sample size so there's a higher chance that the result will deviate from the expected value. For all we know all three values are within an acceptable deviation from the expected value. Do you want me to do a t-test or something? I mean, I could... or you could. If you accept everything up to the point of demanding one t-test on the difference between 4223 (expected diamond zerg population) and 4059 (skill-adjusted diamond zerg population), then I'm doing ok. I'd like to do a simple binomial test on this difference to see the probability of that difference having popped up by chance. But I gotta read back over an old paper and remember how to do that. And I don't have time for that right now. D: So... I'll try to later. Or you could do it! ((By the way, on your point about ZvT, Dustin Browder has said that high-level ZvT is 50/50.)) Everything is 50/50 with good matchmaking regardless of balance =/. Also browder said that back in early/mid July. And no that's not really the only problem I have with the analysis. I mean the numbers aren't even valuable numbers to datamine. Simply put the data needed to statistically determine any sort of racial imbalance is not available to us as players. Your time would be better spent statistically analyzing tournament results. Why aren't the numbers valuable numbers to datamine? It's census data! And on the 50/50 point, two things: 1) We've established the zergs are well-represented at the top, yet STILL ZvT is 50/50, even though good zergs are playing good terrans; and 2) please give the Blizz staff a little more credit... They're not so stupid as to see 50/50 ZvT diamond results and ignore the possibility that good zergs are playing bad terrans and getting 50/50. And there have been no major nerfs to zerg or buffs to terran since his comment -- he made it AFTER the "ZvT is unfair" cries became so loud. Tournament results seem so much worse. They're extremely anecdotal (there aren't many) and there is no matchmaking component at all. It's not a system that churns out data like the ladder is. Actually, I'd say perceived tournament outcomes CAUSED this perception that TvZ is so imbalanced. People watch qxc turtle Sheth through three waves of ultras and call imba. Later, IdrA destroying everyone in KotB is completely ignored when Silver beats IdrA 2-0. So people are dumb about tournament results, imo. That's the opposite of science. This is just so wrong it's hard for me to start =/.
Given that I'm essentially playing the role of a statistics textbook in this thread, that's an interesting reaction.
Tournaments are the #1 source of data to identify balance issues. While it's a smaller sample size we know that a) we're dealing with the highest caliber of players, b) all of said players are playing to the best of their ability because $ is on the line. Plus it's our biggest source of data of specific match-ups.
Second people selectively viewing tournament results has ABSOLUTELY NOTHING TO DO WITH THE VALUE OF LOOKING AT TOURNAMENT RESULTS. That's just so absurd I can't believe you said it. Just because some bozos decide to only focus on high profile zerg losses doesn't mean that taking an objective look at all tournament results is worthless. I mean is that really the argument you are putting forward?
I'm not quite sure what you're so upset about. The fact of tournament results is that there are very few games played. This results in extremely small data pools, which is always bad. You can't conclude much scientifically from 10 tournament TvZ games dominated by 2 or 4 players.
Bnet matchmaking is determined by performance. Tournament rosters are determined by invitation or popularity... not an objective system. This makes the "match-ups" rather anecdotal.
But I'm not outright opposed to looking at tournament results. If you can prove, with numbers, that zerg is underrepresented or underperforming in recent tournaments, I'll listen.
Lastly your comments on 50/50 make absolutely no sense to me. Do you believe Blizzard is looking at every single game played and deciding that the players are of equal skill? How in the world could they possibly tell from the data that we have that there is a big balance issue?
The Bnet matchmaking system determines ratings on the basis of past performance. If TvZ is 50/50 controlling for rating (at high levels, when lots and lots of games have been played), then TvZ is balanced -- unless zergs are underrepresented in diamond. They're not, relative to the total percentage of players who play zerg.
A lot of the affects of a matchup imbalance are going to be hidden by differences in player skill and the fact that only roughly 1/3rd of the games are played with that matchup. As an example of how much this affects things lets say you play 100 games vs a hypothetical evenly matched blob of players. If you have 50/50 for 2 matchups and 40/60 for one matchup you'd still be expected to win
46 of your 100 games. The fact that it's 1 of 3 possible matchups means the actual affect, even if the imbalance is pretty significant, appears as a very tiny thing. When you factor in how ELO fudges this and drives it more towards 50% it's obvious that you can't just look at the data we have.
Your hypothetical would predict an underrepresentation of zergs at the high levels. This doesn't seem to be the case.
Besides the quote starts with, "I have the same feel, but the numbers don't support it..." and he goes on to say, "We are studying the issue and trying to figure out if we should do something..." That seems to imply that they're not sure of the data and need further analysis.
Well, they're approaching it scientifically and diplomatically. They're saying: "We're thinking about it." I'm doing the same thing. The current data does not seem to imply that zerg is underpowered in any way. I've yet to see numerical or statistical evidence that zerg IS underpowered.
|
How does the number of zerg players in diamond compared to the total number of zerg players have anything to do with this analysis?
The only thing that matters is how skilled these players are, and there is no objective method to measure that. Bnet rating is misleading because if the matchup were in fact imbalanced then zerg players would end up with lower ratings than their skill level deserves.
Simply put, you can't use this kind of statistical information to decide on balance (whether you use ladder or tournament data). Neither would this kind of aggregate information help fix any matchup. The kind of details needed here are specific builds/unit compositions that are producing high win rates, or an objective analysis of IdrA's claim that terran has 20 build orders which require different counters and cannot all be scouted in time.
|
How does the number of zerg players in diamond compared to the total number of zerg players have anything to do with this analysis?
Because it would show that something is allowing zerg to place in Diamond with a higher frequency then all the other races.
So are Zerg players just better :D?
|
On August 07 2010 08:34 Logo wrote:Show nested quote +Couldn't agree more, people are only seeing what they want to see. Z is very well represented in tournaments, both in participation numbers and in first or second place finishers. When a Z wins a tournament, people seem not to notice. IdrA sweeping the KotB, etc etc. Hell, last night's GosuCoaching #7 had a ZvZ finals (Sheth vs Slush). I guess when a Z wins it's not noteworthy at all, but when a T wins it's because of imba. [Citation Needed] Saying it doesn't make it true. If you want to say tournaments well represent each race then do it out and show that they do. If you want to say there is an imbalance in tournament wins by match-up then likewise show it.
Why? I'm Zerg, so it would seem I would have high interest in identifying an imbalance if such exists. However, an eyeball across the tournament rosters shows me an acceptable spread, and it is immediately plain that plenty of Zerg are winning or placing in them as well. These quick examinations are enough to satisfy me. You're the one claiming that the current situation is imbalanced, so the burden of proof falls on you, not me. Do your own work.
|
On August 07 2010 08:53 kajeus wrote:Show nested quote +On August 07 2010 08:34 Logo wrote:On August 07 2010 06:55 kajeus wrote:On August 07 2010 06:52 Logo wrote:On August 07 2010 06:46 kajeus wrote:On August 07 2010 06:29 Logo wrote:On August 07 2010 06:17 kajeus wrote:On August 07 2010 05:22 kGold wrote:
The only defense could be that while there are fewer Zerg players than the other races on Korea, there are more BETTER Zerg players per capita of Zerg players in Korea vs. the other races.
I can believe it because Zerg is the hardest race to play so the beginner-intermediate players will show more of a tendency to stay away from Zerg. Ok, here's the data that is directly relevant to YOUR hypothesis.. The problem with statistical analysis like this is it's not actually based on any sort of anything. If you aren't going to use some actual methods of statistical analysis with math to back it up, just don't even bother. For example you don't even account for the fact that Zerg is a smaller sample size so there's a higher chance that the result will deviate from the expected value. For all we know all three values are within an acceptable deviation from the expected value. Do you want me to do a t-test or something? I mean, I could... or you could. If you accept everything up to the point of demanding one t-test on the difference between 4223 (expected diamond zerg population) and 4059 (skill-adjusted diamond zerg population), then I'm doing ok. I'd like to do a simple binomial test on this difference to see the probability of that difference having popped up by chance. But I gotta read back over an old paper and remember how to do that. And I don't have time for that right now. D: So... I'll try to later. Or you could do it! ((By the way, on your point about ZvT, Dustin Browder has said that high-level ZvT is 50/50.)) Everything is 50/50 with good matchmaking regardless of balance =/. Also browder said that back in early/mid July. And no that's not really the only problem I have with the analysis. I mean the numbers aren't even valuable numbers to datamine. Simply put the data needed to statistically determine any sort of racial imbalance is not available to us as players. Your time would be better spent statistically analyzing tournament results. Why aren't the numbers valuable numbers to datamine? It's census data! And on the 50/50 point, two things: 1) We've established the zergs are well-represented at the top, yet STILL ZvT is 50/50, even though good zergs are playing good terrans; and 2) please give the Blizz staff a little more credit... They're not so stupid as to see 50/50 ZvT diamond results and ignore the possibility that good zergs are playing bad terrans and getting 50/50. And there have been no major nerfs to zerg or buffs to terran since his comment -- he made it AFTER the "ZvT is unfair" cries became so loud. Tournament results seem so much worse. They're extremely anecdotal (there aren't many) and there is no matchmaking component at all. It's not a system that churns out data like the ladder is. Actually, I'd say perceived tournament outcomes CAUSED this perception that TvZ is so imbalanced. People watch qxc turtle Sheth through three waves of ultras and call imba. Later, IdrA destroying everyone in KotB is completely ignored when Silver beats IdrA 2-0. So people are dumb about tournament results, imo. That's the opposite of science. This is just so wrong it's hard for me to start =/. Given that I'm essentially playing the role of a statistics textbook in this thread, that's an interesting reaction. Show nested quote +Tournaments are the #1 source of data to identify balance issues. While it's a smaller sample size we know that a) we're dealing with the highest caliber of players, b) all of said players are playing to the best of their ability because $ is on the line. Plus it's our biggest source of data of specific match-ups.
Second people selectively viewing tournament results has ABSOLUTELY NOTHING TO DO WITH THE VALUE OF LOOKING AT TOURNAMENT RESULTS. That's just so absurd I can't believe you said it. Just because some bozos decide to only focus on high profile zerg losses doesn't mean that taking an objective look at all tournament results is worthless. I mean is that really the argument you are putting forward? I'm not quite sure what you're so upset about. The fact of tournament results is that there are very few games played. This results in extremely small data pools, which is always bad. You can't conclude much scientifically from 10 tournament TvZ games dominated by 2 or 4 players. Bnet matchmaking is determined by performance. Tournament rosters are determined by invitation or popularity... not an objective system. This makes the "match-ups" rather anecdotal. But I'm not outright opposed to looking at tournament results. If you can prove, with numbers, that zerg is underrepresented or underperforming in recent tournaments, I'll listen. Show nested quote +Lastly your comments on 50/50 make absolutely no sense to me. Do you believe Blizzard is looking at every single game played and deciding that the players are of equal skill? How in the world could they possibly tell from the data that we have that there is a big balance issue? The Bnet matchmaking system determines ratings on the basis of past performance. If TvZ is 50/50 controlling for rating (at high levels, when lots and lots of games have been played), then TvZ is balanced -- unless zergs are underrepresented in diamond. They're not, relative to the total percentage of players who play zerg. Show nested quote +A lot of the affects of a matchup imbalance are going to be hidden by differences in player skill and the fact that only roughly 1/3rd of the games are played with that matchup. As an example of how much this affects things lets say you play 100 games vs a hypothetical evenly matched blob of players. If you have 50/50 for 2 matchups and 40/60 for one matchup you'd still be expected to win
46 of your 100 games. The fact that it's 1 of 3 possible matchups means the actual affect, even if the imbalance is pretty significant, appears as a very tiny thing. When you factor in how ELO fudges this and drives it more towards 50% it's obvious that you can't just look at the data we have. Your hypothetical would predict an underrepresentation of zergs at the high levels. This doesn't seem to be the case. Show nested quote +Besides the quote starts with, "I have the same feel, but the numbers don't support it..." and he goes on to say, "We are studying the issue and trying to figure out if we should do something..." That seems to imply that they're not sure of the data and need further analysis. Well, they're approaching it scientifically and diplomatically. They're saying: "We're thinking about it." I'm doing the same thing. The current data does not seem to imply that zerg is underpowered in any way. I've yet to see numerical or statistical evidence that zerg IS underpowered.
You treat diamond like it's 1 skill level. Diamond goes all the way from the lowly protoss who can only a-move gateway forces up to the pro level players. A prime example is the game I just played vs another Zerg. He makes 4 zerglings, a spine crawler on the edge of his own creep, and then a late batch of 2 roaches. When he losses over 1/2 his economy right then and there (and eventually the game) he says, "Omg lame rush." His rating? 468 Diamond with 64 wins. He's played over 100 Diamond rating games and doesn't even know the core basic fundamental of how ZvZ works? Give me a break. Diamond is not one skill level and is in fact quite a large spread of players of many different skill levels. "High Level" cannot be considered Diamond, there's just too many people in Diamond for that distinction to work. If you say took Diamond players about 500 ranking then you might start to have something worth going on.
While tournaments have their own issues when looking at stats they're way better than ladder. They take the highest skills of players and focus just on those. Representation isn't as much of an issue if you are looking for matchups and win %s. Do you really believe there are players out there who can play significantly better than IdrA, TLO, Tester and others but isn't being invited to a tournament? Do you really think there's a Zerg player out there who's completely solved ZvT but never players tournaments? Give me a break. Any player with that kind of jump in skill over our current roster would at the very least be showing up in spades on open tournaments of which there are plenty.
|
On August 07 2010 09:42 BillyMole wrote:Show nested quote +On August 07 2010 08:34 Logo wrote:Couldn't agree more, people are only seeing what they want to see. Z is very well represented in tournaments, both in participation numbers and in first or second place finishers. When a Z wins a tournament, people seem not to notice. IdrA sweeping the KotB, etc etc. Hell, last night's GosuCoaching #7 had a ZvZ finals (Sheth vs Slush). I guess when a Z wins it's not noteworthy at all, but when a T wins it's because of imba. [Citation Needed] Saying it doesn't make it true. If you want to say tournaments well represent each race then do it out and show that they do. If you want to say there is an imbalance in tournament wins by match-up then likewise show it. Why? I'm Zerg, so it would seem I would have high interest in identifying an imbalance if such exists. However, an eyeball across the tournament rosters shows me an acceptable spread, and it is immediately plain that plenty of Zerg are winning or placing in them as well. These quick examinations are enough to satisfy me. You're the one claiming that the current situation is imbalanced, so the burden of proof falls on you, not me. Do your own work.
How can you say this after calling people out on selectively looking at tournament results?
Also I'm not claiming anything, I'm just pointing out that people are using woefully inadequate methods of parsing data to act like they're proving a point.
|
On August 07 2010 02:58 Rah wrote:Show nested quote +On August 07 2010 02:50 WinHouse wrote:On August 07 2010 02:46 Rah wrote:
SC2 zerg players definitly win the bitchiest race award. Most of the kids complaining won't be happy until they can win on equal bases through a chokepoint with no micro. Zerg still wins games you know, even when played sub par. All of the high level zerg losses I've seen (since the mech nerf and ultralisk buff) have been due to an error or missed oppertunity on the zerg players part. Then they blame imbalance because it's the cool thing to do now. It's really gotten out of control.
p.s. the current state of zvz is awesome. Actually people bitch more about the bitching like you. No bitching involved from me, my post is pointing out the current state of 1/3 the community, and just maybe bringing it more to their attention. Well less than 1/3rd now since half of them have moved on to being terrible terran players instead of terrible zerg players. Thanks for the rage though? Some people can't be reasoned with when it's easier to pass off blame. That's what got this "imbalance" crap going in the first place.
Terrible post, how can you take 1/3 of the community as a whole (aka every zerg player) and say that every single one of them are whiny bitches.
And why can't you recognize that, maybe, some of the ''bitches'' might have a valid concern?
You sound like the one that is impossible to be reasoned with when you go around posting such statements.
I personaly found the OP kinda funny, but it obviously meant to put oil on the flames.
|
GUYS GUYS....... but do you really want chat rooms?
|
i am happy to present this Zerg Tears replay for the enjoyment of the community: [url blocked]
|
On August 07 2010 09:44 Logo wrote:Show nested quote +On August 07 2010 08:53 kajeus wrote:On August 07 2010 08:34 Logo wrote:On August 07 2010 06:55 kajeus wrote:On August 07 2010 06:52 Logo wrote:On August 07 2010 06:46 kajeus wrote:On August 07 2010 06:29 Logo wrote:On August 07 2010 06:17 kajeus wrote:On August 07 2010 05:22 kGold wrote:
The only defense could be that while there are fewer Zerg players than the other races on Korea, there are more BETTER Zerg players per capita of Zerg players in Korea vs. the other races.
I can believe it because Zerg is the hardest race to play so the beginner-intermediate players will show more of a tendency to stay away from Zerg. Ok, here's the data that is directly relevant to YOUR hypothesis.. The problem with statistical analysis like this is it's not actually based on any sort of anything. If you aren't going to use some actual methods of statistical analysis with math to back it up, just don't even bother. For example you don't even account for the fact that Zerg is a smaller sample size so there's a higher chance that the result will deviate from the expected value. For all we know all three values are within an acceptable deviation from the expected value. Do you want me to do a t-test or something? I mean, I could... or you could. If you accept everything up to the point of demanding one t-test on the difference between 4223 (expected diamond zerg population) and 4059 (skill-adjusted diamond zerg population), then I'm doing ok. I'd like to do a simple binomial test on this difference to see the probability of that difference having popped up by chance. But I gotta read back over an old paper and remember how to do that. And I don't have time for that right now. D: So... I'll try to later. Or you could do it! ((By the way, on your point about ZvT, Dustin Browder has said that high-level ZvT is 50/50.)) Everything is 50/50 with good matchmaking regardless of balance =/. Also browder said that back in early/mid July. And no that's not really the only problem I have with the analysis. I mean the numbers aren't even valuable numbers to datamine. Simply put the data needed to statistically determine any sort of racial imbalance is not available to us as players. Your time would be better spent statistically analyzing tournament results. Why aren't the numbers valuable numbers to datamine? It's census data! And on the 50/50 point, two things: 1) We've established the zergs are well-represented at the top, yet STILL ZvT is 50/50, even though good zergs are playing good terrans; and 2) please give the Blizz staff a little more credit... They're not so stupid as to see 50/50 ZvT diamond results and ignore the possibility that good zergs are playing bad terrans and getting 50/50. And there have been no major nerfs to zerg or buffs to terran since his comment -- he made it AFTER the "ZvT is unfair" cries became so loud. Tournament results seem so much worse. They're extremely anecdotal (there aren't many) and there is no matchmaking component at all. It's not a system that churns out data like the ladder is. Actually, I'd say perceived tournament outcomes CAUSED this perception that TvZ is so imbalanced. People watch qxc turtle Sheth through three waves of ultras and call imba. Later, IdrA destroying everyone in KotB is completely ignored when Silver beats IdrA 2-0. So people are dumb about tournament results, imo. That's the opposite of science. This is just so wrong it's hard for me to start =/. Given that I'm essentially playing the role of a statistics textbook in this thread, that's an interesting reaction. Tournaments are the #1 source of data to identify balance issues. While it's a smaller sample size we know that a) we're dealing with the highest caliber of players, b) all of said players are playing to the best of their ability because $ is on the line. Plus it's our biggest source of data of specific match-ups.
Second people selectively viewing tournament results has ABSOLUTELY NOTHING TO DO WITH THE VALUE OF LOOKING AT TOURNAMENT RESULTS. That's just so absurd I can't believe you said it. Just because some bozos decide to only focus on high profile zerg losses doesn't mean that taking an objective look at all tournament results is worthless. I mean is that really the argument you are putting forward? I'm not quite sure what you're so upset about. The fact of tournament results is that there are very few games played. This results in extremely small data pools, which is always bad. You can't conclude much scientifically from 10 tournament TvZ games dominated by 2 or 4 players. Bnet matchmaking is determined by performance. Tournament rosters are determined by invitation or popularity... not an objective system. This makes the "match-ups" rather anecdotal. But I'm not outright opposed to looking at tournament results. If you can prove, with numbers, that zerg is underrepresented or underperforming in recent tournaments, I'll listen. Lastly your comments on 50/50 make absolutely no sense to me. Do you believe Blizzard is looking at every single game played and deciding that the players are of equal skill? How in the world could they possibly tell from the data that we have that there is a big balance issue? The Bnet matchmaking system determines ratings on the basis of past performance. If TvZ is 50/50 controlling for rating (at high levels, when lots and lots of games have been played), then TvZ is balanced -- unless zergs are underrepresented in diamond. They're not, relative to the total percentage of players who play zerg. A lot of the affects of a matchup imbalance are going to be hidden by differences in player skill and the fact that only roughly 1/3rd of the games are played with that matchup. As an example of how much this affects things lets say you play 100 games vs a hypothetical evenly matched blob of players. If you have 50/50 for 2 matchups and 40/60 for one matchup you'd still be expected to win
46 of your 100 games. The fact that it's 1 of 3 possible matchups means the actual affect, even if the imbalance is pretty significant, appears as a very tiny thing. When you factor in how ELO fudges this and drives it more towards 50% it's obvious that you can't just look at the data we have. Your hypothetical would predict an underrepresentation of zergs at the high levels. This doesn't seem to be the case. Besides the quote starts with, "I have the same feel, but the numbers don't support it..." and he goes on to say, "We are studying the issue and trying to figure out if we should do something..." That seems to imply that they're not sure of the data and need further analysis. Well, they're approaching it scientifically and diplomatically. They're saying: "We're thinking about it." I'm doing the same thing. The current data does not seem to imply that zerg is underpowered in any way. I've yet to see numerical or statistical evidence that zerg IS underpowered. You treat diamond like it's 1 skill level. Diamond goes all the way from the lowly protoss who can only a-move gateway forces up to the pro level players. A prime example is the game I just played vs another Zerg. He makes 4 zerglings, a spine crawler on the edge of his own creep, and then a late batch of 2 roaches. When he losses over 1/2 his economy right then and there (and eventually the game) he says, "Omg lame rush." His rating? 468 Diamond with 64 wins. He's played over 100 Diamond rating games and doesn't even know the core basic fundamental of how ZvZ works? Give me a break. Diamond is not one skill level and is in fact quite a large spread of players of many different skill levels. "High Level" cannot be considered Diamond, there's just too many people in Diamond for that distinction to work. If you say took Diamond players about 500 ranking then you might start to have something worth going on.
I'm not treating diamond like it's one skill level. I'm treating it like it's the top 5% of the playerbase.
Matchmaking is not made on the basis of league. It is made on the basis of rating. Do you think that rating is a bad way of determining skill level? Because it's determined by who you beat and who you don't.
While tournaments have their own issues when looking at stats they're way better than ladder. They take the highest skills of players and focus just on those. Representation isn't as much of an issue if you are looking for matchups and win %s. Do you really believe there are players out there who can play significantly better than IdrA, TLO, Tester and others but isn't being invited to a tournament? Do you really think there's a Zerg player out there who's completely solved ZvT but never players tournaments? Give me a break. Any player with that kind of jump in skill over our current roster would at the very least be showing up in spades on open tournaments of which there are plenty.
Well, there are lots of players in the top 20 of every region's ladder who I've rarely seen in tournaments. And my entire point about Korea was that Korean zergs seemed to do pretty well against terrans. Again, this is an entirely subjective point -- they seemed to.
Anyway, Tester is like 10th on the Korean ladder, fyi. Who's first? Last time I looked (couple hours ago), it was CheckPrime, the zerg. Second place was another zerg. Third was a terran.
((EDIT: Tester is currently 22nd on the Korean ladder. CheckPrime is 1st.))
|
Listen to kajeus.
Whether you think TvZ is balanced or not it takes months of game data and tournaments with different map pools to come to a conclusion. Unless it's some major design flaw that was overlooked I doubt Blizzard is going to patch in anything new for a couple of months.
|
I'm just going to say this:
IdrA --> Thinks about switching to Terran, plays 1/3 of his time on his Terran account and clearly stated that zergs are not worth shit anymore
Dimaga --> Did some all ins every game since he knows he barely stand a chance vs them.
MoMaN --> Switched to Terran... straight pipe... he learned them in a couple days and now wins alot with them.
Artosis --> Eventho he will mostly repete everything that is said by IdrA.... he's still saying that zerg needs something
SLush --> Says that Terran is too strong economicaly and also says that tanks and thors are a pain to deal with. Also he said that zerg has no early harass and that it's hard to deal with all the crap that terran players can throw at thim.
Sheth --> I don't know his stances on the subject.
My stance --> I find that either tanks should be nerfed or thor air damage. The only viable counter to tanks are mutalisk which are taken out by 1-2 thors... yah yah spread your mutalisk blah blah blah... do you have to spread your marine balls? do you have to spread your tanks? I mean... ffs you'd need 20+ mutalisk to kill those 2 thors anyways. Also... hellions/banshee must be delayed... it's just stupid how fast hellions arrive and how devastating they are. Than banshee will normaly arrive just after. It's ridiculous.
So besides that... it's all good.
|
You know, all it really takes for me to be satisfied to make ZvT a lot more bearable would be to add in overkill on tanks.
Is that really so much to ask?
|
I remember when boxer bunker rushed yellow like 5 times in a row on a finals, people started saying "tvz imba nerf naw!1".... then came julyzerg, savior, etc...
|
On August 07 2010 10:04 kajeus wrote:Show nested quote +On August 07 2010 09:44 Logo wrote:On August 07 2010 08:53 kajeus wrote:On August 07 2010 08:34 Logo wrote:On August 07 2010 06:55 kajeus wrote:On August 07 2010 06:52 Logo wrote:On August 07 2010 06:46 kajeus wrote:On August 07 2010 06:29 Logo wrote:On August 07 2010 06:17 kajeus wrote:On August 07 2010 05:22 kGold wrote:
The only defense could be that while there are fewer Zerg players than the other races on Korea, there are more BETTER Zerg players per capita of Zerg players in Korea vs. the other races.
I can believe it because Zerg is the hardest race to play so the beginner-intermediate players will show more of a tendency to stay away from Zerg. Ok, here's the data that is directly relevant to YOUR hypothesis.. The problem with statistical analysis like this is it's not actually based on any sort of anything. If you aren't going to use some actual methods of statistical analysis with math to back it up, just don't even bother. For example you don't even account for the fact that Zerg is a smaller sample size so there's a higher chance that the result will deviate from the expected value. For all we know all three values are within an acceptable deviation from the expected value. Do you want me to do a t-test or something? I mean, I could... or you could. If you accept everything up to the point of demanding one t-test on the difference between 4223 (expected diamond zerg population) and 4059 (skill-adjusted diamond zerg population), then I'm doing ok. I'd like to do a simple binomial test on this difference to see the probability of that difference having popped up by chance. But I gotta read back over an old paper and remember how to do that. And I don't have time for that right now. D: So... I'll try to later. Or you could do it! ((By the way, on your point about ZvT, Dustin Browder has said that high-level ZvT is 50/50.)) Everything is 50/50 with good matchmaking regardless of balance =/. Also browder said that back in early/mid July. And no that's not really the only problem I have with the analysis. I mean the numbers aren't even valuable numbers to datamine. Simply put the data needed to statistically determine any sort of racial imbalance is not available to us as players. Your time would be better spent statistically analyzing tournament results. Why aren't the numbers valuable numbers to datamine? It's census data! And on the 50/50 point, two things: 1) We've established the zergs are well-represented at the top, yet STILL ZvT is 50/50, even though good zergs are playing good terrans; and 2) please give the Blizz staff a little more credit... They're not so stupid as to see 50/50 ZvT diamond results and ignore the possibility that good zergs are playing bad terrans and getting 50/50. And there have been no major nerfs to zerg or buffs to terran since his comment -- he made it AFTER the "ZvT is unfair" cries became so loud. Tournament results seem so much worse. They're extremely anecdotal (there aren't many) and there is no matchmaking component at all. It's not a system that churns out data like the ladder is. Actually, I'd say perceived tournament outcomes CAUSED this perception that TvZ is so imbalanced. People watch qxc turtle Sheth through three waves of ultras and call imba. Later, IdrA destroying everyone in KotB is completely ignored when Silver beats IdrA 2-0. So people are dumb about tournament results, imo. That's the opposite of science. This is just so wrong it's hard for me to start =/. Given that I'm essentially playing the role of a statistics textbook in this thread, that's an interesting reaction. Tournaments are the #1 source of data to identify balance issues. While it's a smaller sample size we know that a) we're dealing with the highest caliber of players, b) all of said players are playing to the best of their ability because $ is on the line. Plus it's our biggest source of data of specific match-ups.
Second people selectively viewing tournament results has ABSOLUTELY NOTHING TO DO WITH THE VALUE OF LOOKING AT TOURNAMENT RESULTS. That's just so absurd I can't believe you said it. Just because some bozos decide to only focus on high profile zerg losses doesn't mean that taking an objective look at all tournament results is worthless. I mean is that really the argument you are putting forward? I'm not quite sure what you're so upset about. The fact of tournament results is that there are very few games played. This results in extremely small data pools, which is always bad. You can't conclude much scientifically from 10 tournament TvZ games dominated by 2 or 4 players. Bnet matchmaking is determined by performance. Tournament rosters are determined by invitation or popularity... not an objective system. This makes the "match-ups" rather anecdotal. But I'm not outright opposed to looking at tournament results. If you can prove, with numbers, that zerg is underrepresented or underperforming in recent tournaments, I'll listen. Lastly your comments on 50/50 make absolutely no sense to me. Do you believe Blizzard is looking at every single game played and deciding that the players are of equal skill? How in the world could they possibly tell from the data that we have that there is a big balance issue? The Bnet matchmaking system determines ratings on the basis of past performance. If TvZ is 50/50 controlling for rating (at high levels, when lots and lots of games have been played), then TvZ is balanced -- unless zergs are underrepresented in diamond. They're not, relative to the total percentage of players who play zerg. A lot of the affects of a matchup imbalance are going to be hidden by differences in player skill and the fact that only roughly 1/3rd of the games are played with that matchup. As an example of how much this affects things lets say you play 100 games vs a hypothetical evenly matched blob of players. If you have 50/50 for 2 matchups and 40/60 for one matchup you'd still be expected to win
46 of your 100 games. The fact that it's 1 of 3 possible matchups means the actual affect, even if the imbalance is pretty significant, appears as a very tiny thing. When you factor in how ELO fudges this and drives it more towards 50% it's obvious that you can't just look at the data we have. Your hypothetical would predict an underrepresentation of zergs at the high levels. This doesn't seem to be the case. Besides the quote starts with, "I have the same feel, but the numbers don't support it..." and he goes on to say, "We are studying the issue and trying to figure out if we should do something..." That seems to imply that they're not sure of the data and need further analysis. Well, they're approaching it scientifically and diplomatically. They're saying: "We're thinking about it." I'm doing the same thing. The current data does not seem to imply that zerg is underpowered in any way. I've yet to see numerical or statistical evidence that zerg IS underpowered. You treat diamond like it's 1 skill level. Diamond goes all the way from the lowly protoss who can only a-move gateway forces up to the pro level players. A prime example is the game I just played vs another Zerg. He makes 4 zerglings, a spine crawler on the edge of his own creep, and then a late batch of 2 roaches. When he losses over 1/2 his economy right then and there (and eventually the game) he says, "Omg lame rush." His rating? 468 Diamond with 64 wins. He's played over 100 Diamond rating games and doesn't even know the core basic fundamental of how ZvZ works? Give me a break. Diamond is not one skill level and is in fact quite a large spread of players of many different skill levels. "High Level" cannot be considered Diamond, there's just too many people in Diamond for that distinction to work. If you say took Diamond players about 500 ranking then you might start to have something worth going on. I'm not treating diamond like it's one skill level. I'm treating it like it's the top 5% of the playerbase. Matchmaking is not made on the basis of league. It is made on the basis of rating. Do you think that rating is a bad way of determining skill level? Because it's determined by who you beat and who you don't. Show nested quote +While tournaments have their own issues when looking at stats they're way better than ladder. They take the highest skills of players and focus just on those. Representation isn't as much of an issue if you are looking for matchups and win %s. Do you really believe there are players out there who can play significantly better than IdrA, TLO, Tester and others but isn't being invited to a tournament? Do you really think there's a Zerg player out there who's completely solved ZvT but never players tournaments? Give me a break. Any player with that kind of jump in skill over our current roster would at the very least be showing up in spades on open tournaments of which there are plenty.
Well, there are lots of players in the top 20 of every region's ladder who I've rarely seen in tournaments. And my entire point about Korea was that Korean zergs seemed to do pretty well against terrans. Again, this is an entirely subjective point -- they seemed to. Anyway, Tester is like 10th on the Korean ladder, fyi. Who's first? Last time I looked (couple hours ago), it was CheckPrime, the zerg. Second place was another zerg. Third was a terran. ((EDIT: Tester is currently 22nd on the Korean ladder. CheckPrime is 1st.))
The only thing is you haven't proved they've done well against Terrans. You've just proved that Zerg are represented quite well in the ranking system and that they play vs Terrans ~34% of the time. I agree that seems to suggest that koreans do alright vs zerg, but it's still my same basic point that the data isn't enough to really prove anything. I also do agree with Ace that there hasn't been enough tournaments to completely decide definitively anything, but it'd at least be a start.
The thing about representation is that even if there are some strong ladder players who aren't playing in tournaments, there's no evidence that they'd be significantly better than the current pool to the point of skewing the data if it shows it's in Terran's favor.
It's a really tough thing to make a call on. As we know from BW you can leave a game for years and something can still evolve that shows it's balanced. Or you could leave a game for years and it'd still remain shit. I think either way the consensus is that ZvT plays like crap currently (for the Zerg at least) regardless of win % and Zerg really don't think they have many options to expand their play to make it feel better. If it's not imbalanced it still might warrant changes horizontally (not a straight buff for either side, but changes that make it more even for control of the game).
|
On August 07 2010 10:10 Konsume wrote:
I'm just going to say this:
IdrA --> Thinks about switching to Terran, plays 1/3 of his time on his Terran account and clearly stated that zergs are not worth shit anymore
Dimaga --> Did some all ins every game since he knows he barely stand a chance vs them.
MoMaN --> Switched to Terran... straight pipe... he learned them in a couple days and now wins alot with them.
Artosis --> Eventho he will mostly repete everything that is said by IdrA.... he's still saying that zerg needs something
SLush --> Says that Terran is too strong economicaly and also says that tanks and thors are a pain to deal with. Also he said that zerg has no early harass and that it's hard to deal with all the crap that terran players can throw at thim.
Sheth --> I don't know his stances on the subject..
MoMaN switched to terran? Could you link a news item for that, please? I was watching his stream like two days ago, and he was still zerg.
Did Dimaga actually say that he all-ins because he doesn't think he stands a chance versus toss, or are you just going off of what IdrA said? Because I know IdrA claims that, but IdrA spent his entire BW career proclaiming how badly OP protoss was.
Also, could you link some of SLush's comments on the subject?
I wonder what CheckPrime and Cool think, personally. My entire point in all of this was: a) Perception is not reality; and b) the Korean zergs seem to be doing pretty well.
The only thing is you haven't proved they've done well against Terrans. You've just proved that Zerg are represented quite well in the ranking system and that they play vs Terrans ~34% of the time. I agree that seems to suggest that koreans do alright vs zerg, but it's still my same basic point that the data isn't enough to really prove anything. I also do agree with Ace that there hasn't been enough tournaments to completely decide definitively anything, but it'd at least be a start.
The thing about representation is that even if there are some strong ladder players who aren't playing in tournaments, there's no evidence that they'd be significantly better than the current pool to the point of skewing the data if it shows it's in Terran's favor.
All I'm doing is providing evidence that zerg is not underpowered, man. B)
It sounds like you want a more recent in-depth comment from Blizzard on TvZ win-rates. I hope they give you that. But I also think they would definitely have done something if it were clear to them that TvZ is much more easily won by the T at high levels.
|
On August 07 2010 10:00 ToxNub wrote:
i am happy to present this Zerg Tears replay for the enjoyment of the community: [url blocked]
Good harvest, ToxNub. I was a bit surprised, though. I mean, I thought I was going to see tank and mech battles against 1a hydras but your opponent was raging about bio mech with no siege tanks. Siege tank-less zerg tears are hard to get.
|
On August 07 2010 10:15 kajeus wrote:Show nested quote +On August 07 2010 10:10 Konsume wrote:
I'm just going to say this:
IdrA --> Thinks about switching to Terran, plays 1/3 of his time on his Terran account and clearly stated that zergs are not worth shit anymore
Dimaga --> Did some all ins every game since he knows he barely stand a chance vs them.
MoMaN --> Switched to Terran... straight pipe... he learned them in a couple days and now wins alot with them.
Artosis --> Eventho he will mostly repete everything that is said by IdrA.... he's still saying that zerg needs something
SLush --> Says that Terran is too strong economicaly and also says that tanks and thors are a pain to deal with. Also he said that zerg has no early harass and that it's hard to deal with all the crap that terran players can throw at thim.
Sheth --> I don't know his stances on the subject.. MoMaN switched to terran? Could you link a news item for that, please? I was watching his stream like two days ago, and he was still zerg.
News? why news... he played all the last tournament as terran
here is a replay:
http://www.sc2rep.com/replays/show/id/70
On August 07 2010 10:15 kajeus wrote:
Did Dimaga actually say that he all-ins because he doesn't think he stands a chance versus toss, or are you just going off of what IdrA said?
Do you need him to say it? He litteraly went ALL IN in each games. I mean... I wouldn't say that he was deseperate if he did some all ins 1-2 games... but basicaly all his ZvT was ALL INS!... says alot... I mean... if you can't see it well sorry bro you must be one hell of a detective.
On August 07 2010 10:15 kajeus wrote:
Also, could you link some of SLush's comments on the subject?
I don't need a comment made by SLush.. HE'S HANGING IN MY VENT! We talk alot on vent and he's giving me lots of tips and such. Why would I need to link something from him? Just come and talk with him how about that?
|
|
|
|
|
|
|
|
EPT
