EPTHe said "DongRaeGu's weakest matchup is against Terran."
I thought, "wait a minute....isn't DRG suppose to be some Terran killer god?! Why is his weakest matchup ZvT?" It turns out Khaldor was right...
DongRaeGu
All: 67-30 (69.07%)
vT: 32-18 64.00%
vZ: 8-2 80.00%
vP: 27-10 72.97%
64% looks pretty terrible against the backdrop of 80% and 72%. So I shrugged it off...but it's been eating at me for over a week...I decided to investigate more...
The question was: Who has the best ZvT%? Is it possible to have a good ZvT percentage compared to TvZ?
Here are some findings
Top 6 Zergs by ELO using TLPD. I used the top 6 cutoff because nobody was really worth mentioning after Curious and Symbol...and then you get down to the top 20-ish on ELO...which by itself is no longer top tier play.
DRG vT: 32-18 64.00%
Nestea vT: 52-39 57.14%
Leenock vT: 62-40 60.78%
July vT: 29-19 60.42%
Curious vT: 25-19 56.82%
Symbol vT: 13-8 61.90%
Clearly DRG's ZvT outclasses everyone by at least 3-5%, DongRaeGu is indeed the best ZvT player in the world....statistic wise.
So...what about TvZ
Top 6 Terrans by ELO using TLPD
MMA vZ: 21-5 80.77%
MVP vZ: 43-21 67.19%
aLive vZ: 28-14 66.67%
MKP vZ: 39-21 65.00%
Taeja vZ: 58-29 66.67%
Gumiho vZ: 30-15 66.67%
....the discrepency is HUUUUUUGEE!!
Top Korean Terrans really can play well against Zerg...by a margin of greater than 5% win/loss ratio. How can this be? I am not sure it's a balance issue or design issue or just the meta game, but ZvT does not compare well to TvZ statistically in korea.
I come to the conclusion that....unless you're DRG, nobody's really any good at ZvT.
But then the stats themselves clearly are saying...DRG's not actually good at ZvT either...he's just good in general and it bleeds over to his ZvT.
Food for thought...
****************
Edit 3:
Actually...can anyone name me a top Terran Player on ELO (top 50) that has a bad TvZ? (as in 50% or lower)
Is it even possible to be bad at TvZ if you're in the GSL?
I will name you 5 Zergs in the top 50 that's terrible with ZvT
Zenio: 24-32 (42.86%)
Losira: 27-34 (44.26%)
Lucky: 16-21 (43.24%)
Coca: 18-24 (42.86%)
Seal: 12-18 (40.00%)
I removed BBBB from the list because 48% is pretty close to 50...benefit of the doubt.
(forGG only played 12 TvZ games, maybe it's not a good idea to count him)
******************
Top comment Highlights
+ Show Spoiler +
On March 16 2012 23:29 kckkryptonite wrote:
http://www.gomtv.net/records/playerInfo.gom?option=view&playerid=10791
July's GSL winrate for his 24 most recent vT's is 70.8%
http://www.gomtv.net/records/playerInfo.gom?option=view&playerid=10173
Nestea's sits at 63.9% for his most recent 36
http://www.gomtv.net/records/playerInfo.gom?option=view&playerid=22681
DRG has risen to 72% in his recent 25
Remember when 2rax was new and like, every Zerg was dying to it? Now you have Zergs like July who consistently crush it. Zergs in general die to it much less to it than before. And how about how much more useful the FG buff made Infestors? Thing's like this are what the overall statistics don't and can't account for, patches, meta-game shifts, maps, and because T/Z players of today are > T/Z players of yesteryear; increases in player skill. Just look at the Curious of early 2011 and Curious now. Look at Fruitdealer then and now/most recently.
In any case, statistics like these shouldn't be read into too much. For instance, they make Squirtle look like a PvZ savante (vZ 71%) , but then you look at his match history and he is just not facing the top-tier Zergs. If you go off the percentages alone, it looks like Squirtle should teach MC (vZ 53%) some PvZ, but then you look at the Zergs MC has been facing.
http://www.gomtv.net/records/playerInfo.gom?option=view&playerid=10791
July's GSL winrate for his 24 most recent vT's is 70.8%
http://www.gomtv.net/records/playerInfo.gom?option=view&playerid=10173
Nestea's sits at 63.9% for his most recent 36
http://www.gomtv.net/records/playerInfo.gom?option=view&playerid=22681
DRG has risen to 72% in his recent 25
Remember when 2rax was new and like, every Zerg was dying to it? Now you have Zergs like July who consistently crush it. Zergs in general die to it much less to it than before. And how about how much more useful the FG buff made Infestors? Thing's like this are what the overall statistics don't and can't account for, patches, meta-game shifts, maps, and because T/Z players of today are > T/Z players of yesteryear; increases in player skill. Just look at the Curious of early 2011 and Curious now. Look at Fruitdealer then and now/most recently.
In any case, statistics like these shouldn't be read into too much. For instance, they make Squirtle look like a PvZ savante (vZ 71%) , but then you look at his match history and he is just not facing the top-tier Zergs. If you go off the percentages alone, it looks like Squirtle should teach MC (vZ 53%) some PvZ, but then you look at the Zergs MC has been facing.
Makes me wonder if we should add an option to TLPD to list only "matches within X amount of months or since last patch"
+ Show Spoiler +
On March 16 2012 12:53 Mr. Nefarious wrote:
I compiled this for another thread I posted it in, I hope a cross post is ok since the information is directly relevant.
opinion starts here
+ Show Spoiler +
end opinion
A quick check of the best of the best Zerg and Terran players will reveal Vs T winrates only slightly above 50%, while the best Terrans look invincible with 70-80% winrates vs Z and P. All numbers taken straight from TLPD.
MMA
1v1 Record:
All: 83-51 (61.94%)
VT: 45-29 (60.81%)
vZ: 21-5 (80.77%)
VP: 17-17 (50.00%)
MVP
1v1 Record:
All: 124-61 (67.03%)
VT: 62-29 (68.13%)
vZ: 41-20 (67.21%)
VP: 21-12 (63.64%)
GuMiho
1v1 Record:
All: 84-57 (59.57%)
VT: 31-33 (48.44%)
vZ: 28-12 (70.00%)
VP: 25-12 (67.57%)
NesTea
1v1 Record:
All: 100-51 (66.23%)
vT: 50-39 (56.18%)
vZ: 28-5 (84.85%)
VP: 22-7 (75.86%)
LosirA
1v1 Record:
All: 67-57 (54.03%)
vT: 25-32 (43.86%)
vZ: 16-11 (59.26%)
VP: 26-14 (65.00%)
DRG
1v1 Record:
All: 58-26 (69.05%)
vT: 29-16 (64.44%)
vZ: 7-2 (77.78%)
VP: 22-8 (73.33%)
MC
1v1 Record:
All: 90-60 (60.00%)
vT: 43-30 (58.90%)
vZ: 22-19 (53.66%)
VP: 25-11 (69.44%)
HuK
1v1 Record:
All: 34-42 (44.74%)
vT: 12-17 (41.38%)
vZ: 13-14 (48.15%)
VP: 9-11 (45.00%)
HerO
1v1 Record:
All: 17-26 (39.53%)
vT: 4-11 (26.67%)
vZ: 10-10 (50.00%)
VP: 3-5 (37.50%)
As seen clearly above, the only player to break 60% winrate vs T is DRG. With that being said, it's also statistically his worst matchup by nearly 10%.
I compiled this for another thread I posted it in, I hope a cross post is ok since the information is directly relevant.
opinion starts here
+ Show Spoiler +
Terran is very strong at the highest level for two main reasons. 1. Terran benefits the most from extreme multitasking. If you have a spare 400apm, you can always marine split, always drop and get away, always stutter step well, you can get 15 kills with a banshee and get away with 5hp, you always can repair hurt units and always salvage buildings just in time. Almost every unit in the Terran army becomes exponentially stronger the better you can control it. Therefor, players that have very high APM and multitasking achieve a level of cost efficiency that is very difficult to match by lower level players. The other prong of my theory is actually very simple. Terran is typically the safest from cheese, and are very hard to attack directly if they are turtling. A higher average cost efficiency produced by excellent control combined with the ability to play very safe if turtling could lead to the highest level Terrans having the most stability as well as becoming exponentially stronger as multitasking and unit control skill increases.
This is just my theory based on watching GSL, playing games on the NA ladder and casually training.
Solution: Examine ways to make the other races more "micro-able". While some micro is needed for Z/P at the moment, it is not nearly as beneficial as extreme micro is for T. Instead of playing with damage numbers, analyze the unit design. Make units attack twice as fast but do the same overall DPS. Allow more units to cancel their attack to move away while still doing damage similar to the marine. These types of changes are obviously targeted solely at the highest level. A-moved roaches will do the same overall DPS in bronze as they will Korea GM, however if they shot twice as quickly they would be a hell of a lot more microable, despite doing the same overall DPS. This would allow additional functionality from players that have the APM to micro their army while maintaining production while not effecting the lowest leagues in the slightest as overall damage output stays the same.
This is just my theory based on watching GSL, playing games on the NA ladder and casually training.
Solution: Examine ways to make the other races more "micro-able". While some micro is needed for Z/P at the moment, it is not nearly as beneficial as extreme micro is for T. Instead of playing with damage numbers, analyze the unit design. Make units attack twice as fast but do the same overall DPS. Allow more units to cancel their attack to move away while still doing damage similar to the marine. These types of changes are obviously targeted solely at the highest level. A-moved roaches will do the same overall DPS in bronze as they will Korea GM, however if they shot twice as quickly they would be a hell of a lot more microable, despite doing the same overall DPS. This would allow additional functionality from players that have the APM to micro their army while maintaining production while not effecting the lowest leagues in the slightest as overall damage output stays the same.
end opinion
A quick check of the best of the best Zerg and Terran players will reveal Vs T winrates only slightly above 50%, while the best Terrans look invincible with 70-80% winrates vs Z and P. All numbers taken straight from TLPD.
MMA
1v1 Record:
All: 83-51 (61.94%)
VT: 45-29 (60.81%)
vZ: 21-5 (80.77%)
VP: 17-17 (50.00%)
MVP
1v1 Record:
All: 124-61 (67.03%)
VT: 62-29 (68.13%)
vZ: 41-20 (67.21%)
VP: 21-12 (63.64%)
GuMiho
1v1 Record:
All: 84-57 (59.57%)
VT: 31-33 (48.44%)
vZ: 28-12 (70.00%)
VP: 25-12 (67.57%)
NesTea
1v1 Record:
All: 100-51 (66.23%)
vT: 50-39 (56.18%)
vZ: 28-5 (84.85%)
VP: 22-7 (75.86%)
LosirA
1v1 Record:
All: 67-57 (54.03%)
vT: 25-32 (43.86%)
vZ: 16-11 (59.26%)
VP: 26-14 (65.00%)
DRG
1v1 Record:
All: 58-26 (69.05%)
vT: 29-16 (64.44%)
vZ: 7-2 (77.78%)
VP: 22-8 (73.33%)
MC
1v1 Record:
All: 90-60 (60.00%)
vT: 43-30 (58.90%)
vZ: 22-19 (53.66%)
VP: 25-11 (69.44%)
HuK
1v1 Record:
All: 34-42 (44.74%)
vT: 12-17 (41.38%)
vZ: 13-14 (48.15%)
VP: 9-11 (45.00%)
HerO
1v1 Record:
All: 17-26 (39.53%)
vT: 4-11 (26.67%)
vZ: 10-10 (50.00%)
VP: 3-5 (37.50%)
As seen clearly above, the only player to break 60% winrate vs T is DRG. With that being said, it's also statistically his worst matchup by nearly 10%.
Some Critiques and answers
1: The sample size is too small
+ Show Spoiler +
On March 16 2012 11:24 Mentalizor wrote:
When you "only" have a 50 game sample (DRG's ZvT is 50games) small mistakes, missmicroes can easilly cost 2-5 games. if just 5 games a lost due to mistakes, that will alter your statistics by 10% which is pretty much. Get bigger samples before comparing statistics. They are just not viable to look at.
When you "only" have a 50 game sample (DRG's ZvT is 50games) small mistakes, missmicroes can easilly cost 2-5 games. if just 5 games a lost due to mistakes, that will alter your statistics by 10% which is pretty much. Get bigger samples before comparing statistics. They are just not viable to look at.
The optimal minimum sample size is 20.
Ugh why do I feel like everytime I bring up stats in teamliquid, I need to teach a whole course of statistics to satisfy the whiners. This is the reason why I didn't want to spend time explaining earlier...but here goes.
Ok here's the breakdown:
The argument: The sample size is not large enough...
We have to understand why this is a problem in the first place.
This is related to the coin flip test...which is a comparison between True theoretical probability and actual probability.
Everyone knows that in a truly balanced coin (yes I know tails land more because head is heavier, but let's assume that the coin is fully balanced) the chances of heads or tails is 50/50.
However, if you flip the coin 8 times, you might get 5 heads and 3 tails, or even 7 heads and 1 tails. The reason this happens is because the actual outcome does not approach the theoretical outcome until very high number of samples are gained. This is related to the question at hand:
Are the number of games played by these top koreans high enough for their theoretical skill level to show?
I answer yes, 20 coinflips or greater tend to be the magical number in which the standard deviation improves significant enough for the gaussian distribution to be acceptable. Thus 20 games or greater is enough to probe how well a pro-gamer is skilled at a single matchup, as the chances of random deviation should decrease significantly when we attain 20 games or more. All these statistics presented in the opening post has more than 20 games. We are pretty safe to say that they matchup well with the player's capabilities.
So here are some pictures
Number of Tails
8 coinflips
![[image loading]](http://www.fourmilab.ch/rpkp/experiments/figures/bd8.png)
16 coinflips
![[image loading]](http://www.fourmilab.ch/rpkp/experiments/figures/bd16.png)
32 coinflips
![[image loading]](http://www.fourmilab.ch/rpkp/experiments/figures/bd32.png)
As you can see, the gaussian distribution gets "slimmer" the more coinflips there are.
This sliming down of the curve can be numerically expressed by the standard deviation. The bigger the sample size, the slimmer the standard deviation, which means the closer the actual probability approaches the theoretical probability.
The calculation for standard deviation is dependent on the inverse of sample size N...
The greater the N samples, the less fluctuations one is likely to see (meaning the standard deviation is smaller...which is what we want)
So why 20?
Because anything greater than 20 is great, but the total impact of N, itself, to the statistics decreases significantly over 20.
Call it diminishing returns.
Those of you who are telling me to go calculate the confidence interval have no idea what you're talking about.
2. The statistics are not relevant because it spans too many patches
+ Show Spoiler +
Edit 4:
Ok I need to address this question because people want an answer.
The statistics are indeed spanning the entire career of a pro-gamer.
In DRG's case it spans all the way back to 2011 GSTL S1.
A counterpoint to the whole "wow so it totally doesn't apply anymore" is that we are taking the CURRENT top 20 or so koreans by ELO. Obviously, if someone hasn't done well recently, their ELO will drop. You don't see fruitdealer or jinro being talked about in these group (even though they are top of the line in their time), because their ELO has fallen off and their data are no longer relevant. People with very high ELO tend to perform well in all matchups most recently. Yes, the treatment of these statistics isn't perfect, I agree that we need better data. But no better data are available at this point for the tippiest top of koreans (most of which don't even ladder or keep their ladder ID hidden), TLPD is well respected and it's there...might as well use it.
Also none of us have any idea how much the ghost patch will affect the matchup at the highest level of play.
You can speculate all you want, but there's no data to say anything.
We'll just have to wait and see.
On March 16 2012 18:56 aebriol wrote:
Just one question: for how long back does the sample size go?
It would be relevant to look maybe 2-3 months back, but ... patches etc, will really mess with the statistics. ZvT was damn near impossible for Z for a while here and there - not really the case right now.
Just one question: for how long back does the sample size go?
It would be relevant to look maybe 2-3 months back, but ... patches etc, will really mess with the statistics. ZvT was damn near impossible for Z for a while here and there - not really the case right now.
Ok I need to address this question because people want an answer.
The statistics are indeed spanning the entire career of a pro-gamer.
In DRG's case it spans all the way back to 2011 GSTL S1.
A counterpoint to the whole "wow so it totally doesn't apply anymore" is that we are taking the CURRENT top 20 or so koreans by ELO. Obviously, if someone hasn't done well recently, their ELO will drop. You don't see fruitdealer or jinro being talked about in these group (even though they are top of the line in their time), because their ELO has fallen off and their data are no longer relevant. People with very high ELO tend to perform well in all matchups most recently. Yes, the treatment of these statistics isn't perfect, I agree that we need better data. But no better data are available at this point for the tippiest top of koreans (most of which don't even ladder or keep their ladder ID hidden), TLPD is well respected and it's there...might as well use it.
Also none of us have any idea how much the ghost patch will affect the matchup at the highest level of play.
You can speculate all you want, but there's no data to say anything.
We'll just have to wait and see.
3. Wait, what's your point again?
+ Show Spoiler +
I'm arguing that currently... nobody is good at ZvT.
I mean, when the best ZvT player is absolutely terrible at ZvT compared to his other matchups(a difference of
almost 10% or more)
In contrast, everyone knows how to play TvZ...if you're GSL caliber, your TvZ tends to not fall below 50% winrate.
I have no idea why this is or how to fix it.
I mean, when the best ZvT player is absolutely terrible at ZvT compared to his other matchups(a difference of
almost 10% or more)
In contrast, everyone knows how to play TvZ...if you're GSL caliber, your TvZ tends to not fall below 50% winrate.
I have no idea why this is or how to fix it.
