|
On September 08 2011 03:33 Skwid1g wrote:Show nested quote +On September 08 2011 02:59 QTIP. wrote:On September 08 2011 02:53 Salteador Neo wrote:On September 08 2011 02:49 merlin101 wrote:On September 08 2011 02:39 Dommk wrote:On September 08 2011 02:37 QTIP. wrote:On September 08 2011 02:20 MrCon wrote:
Winrates of GSL code A qualifiers : (numbers from liquipedia)
TvZ : 25-37 = 40%
TvP : 42-57 = 42%
PvZ : 42-49 = 41%
Yup - some Protoss actually qualified this time! ^_^ Fun fact, Terran has more players in Code S than Protoss has in Code S and A combined  So what? There are more Terrans than Protoss and Zerg! Fun fact: Sample Size is waaay to small! GSL August, winrate of 31,3% (20-44), being 16-31 (34%) versus Terran and 4-13 (23,5%) versus Zerg. This is also a coincidence right? We all must be delusional! C'mon please have some dignity and accept the facts xD To be honest, Protoss players need to stop whining. Hongun, Trickster, Inca, Choya, ... you see how bad they all are? Your only good Protoss is MC. When he starts owning you all look so stupid for whin- Wait what??? MC to Code A? Anyway...like I was saying. MC is bad. He got stomped and he is slumping because he is too cocky. There are no good Protoss players -- you guys have been so abusive for so long and now you've been figured out. Seriously.. l2p. / sarcasm It's honestly pretty true though. A lot of BW toss (marineking included) switched to T/Z because of the beta and because of how BW was. That, combined with the large number of Terrans switching, means that there are just more skilled Terrans (and somewhat Zergs) than there are Protoss. That isn't saying that "ALL PROTOSS PLAYERS ARE INFERIOR," but more so that using simple "X number of Code S/A protoss" isn't really a good way to measure the state of balance. Let's be honest. Inca's PvZ is terrible. Choya always has been super gimmicky, and HongUn is prone to losing games he shouldn't (and winning games he shouldn't) because of his strange playstyle.
I don't disagree with anything you've said, it's certainly a valid theory.
However, given the recent trend of certain strategies being abused beyond reason (1-1-1) as well as both DT/SG openings being easily deflectable by a smart Zerg and Infestor/Ling shitting on 6-7 Warp Gate aggression, Protoss options are greatly limited. When a build as easy as 1-1-1 defeats Protoss players with alarming consistency, you need to consider attributing its success to the build inself instead of the Terran player's skill.
Perhaps these GSL Protoss players aren't as bad as they look -- but rather confined by an inflexible and largely predictable race?
This is also a valid theory.
|
HotS needs to give Protoss a fast cheap harass unit, something they sorely lack.
|
On September 08 2011 03:57 Zzoram wrote:
HotS needs to give Protoss a fast cheap harass unit, something they sorely lack.
Like the reaver and high templar (with a usable storm)?
Not every race is the same.
|
On September 08 2011 03:20 FairForever wrote:Show nested quote +On September 08 2011 03:17 Elean wrote:On September 08 2011 02:49 merlin101 wrote:On September 08 2011 02:39 Dommk wrote:On September 08 2011 02:37 QTIP. wrote:On September 08 2011 02:20 MrCon wrote:
Winrates of GSL code A qualifiers : (numbers from liquipedia)
TvZ : 25-37 = 40%
TvP : 42-57 = 42%
PvZ : 42-49 = 41%
Yup - some Protoss actually qualified this time! ^_^ Fun fact, Terran has more players in Code S than Protoss has in Code S and A combined So what? There are more Terrans than Protoss and Zerg! Fun fact: Sample Size is waaay to small! Out of 64 players (code A + code S), half of them are terrans. The sample size is more than enough to be almost certain there is either an imbalance or a biased in the sample (i.e. terran players are more skilled). From a statistic point of view, saying the sample size isn't large enough is irrelevant. Now, of course you can argue the terran players are better. There is no reason for it to be the case, but nobody can prove the contrary. I personnaly believe that terran players in general are much less skilled than the other races. And the argument for it is quite simple: you improve more when you loose/struggle. Terran have been dominating since the beginning, so in term of strategy they must have fell far behind the other races. And there are signs of this. For instance, at MLG people just realized how strong blue flame were, some people said there even were imbalance (which Blizzard apparently agreed). But blue flame helions have been here since the beginning, how comes it took 1 year to realize it ? Well the answer is simple, terran had plenty of other strategies enabling them to win, so their skills/strategies improved slowly. Same thing can be said about 1/1/1. And, it's not the end, I'm sure there are plenty of other stong build terrans haven't figured out because they are already winning with non-optimal strategies. Anyone just said ghosts ? Did the test involve using t-test? Or what? The problem is that if you can't test this like you test a coin (with p = 1/3), because we don't expect 1/3 of the GSL to be Terran. Our expectation is that the ratio of Terran:Total in the GSL is equal to the ratio of Terran:Total in either Korea, Pro-Gaming Teams, or the World (depending on what you want to use - personally, I'd use Pro-Gaming teams). I don't have the ratio but if half of the players on Pro teams in Korea played Terran, then we'd expect 1/2 to be Terran. EDIT: But yes, if we could find evidence to suggest that our hypothesis should be that p = 1/3, then we could very easily say that there is virtually no possible chance that Terran isn't OP.
I was intrigued by this, and went through the rosters of major KR teams on TLPD, just tallying up the numbers of each race. I don't necessarily think this is the right sample space, perhaps something like Korean GM would be better? In any case, it comes down to the following:
Terran: 60
Protoss: 55
Zerg: 48
Total: 163
Which gives a 37/34/29 distribution, percentage-wise.
One interesting fact I've noted, skimming through those player lists, is that most of the Terran names are known to me, and are either in Code S or Code A, while I haven't even heard of a lot of the Protoss players. So yeah, we might not get any more Terrans in GSL, because the pro teams simply don't have any more.  I imagine the teams try to maintain a healthy race distribution for practice purposes, so unless they go on a crazy Terran shopping spree, we might reach a "Code S for Terrans, Code A for the other races" balanced state.
On September 08 2011 03:43 QTIP. wrote:Show nested quote +On September 08 2011 03:33 Skwid1g wrote:On September 08 2011 02:59 QTIP. wrote:On September 08 2011 02:53 Salteador Neo wrote:On September 08 2011 02:49 merlin101 wrote:On September 08 2011 02:39 Dommk wrote:On September 08 2011 02:37 QTIP. wrote:On September 08 2011 02:20 MrCon wrote:
Winrates of GSL code A qualifiers : (numbers from liquipedia)
TvZ : 25-37 = 40%
TvP : 42-57 = 42%
PvZ : 42-49 = 41%
Yup - some Protoss actually qualified this time! ^_^ Fun fact, Terran has more players in Code S than Protoss has in Code S and A combined So what? There are more Terrans than Protoss and Zerg! Fun fact: Sample Size is waaay to small! GSL August, winrate of 31,3% (20-44), being 16-31 (34%) versus Terran and 4-13 (23,5%) versus Zerg. This is also a coincidence right? We all must be delusional! C'mon please have some dignity and accept the facts xD To be honest, Protoss players need to stop whining. Hongun, Trickster, Inca, Choya, ... you see how bad they all are? Your only good Protoss is MC. When he starts owning you all look so stupid for whin- Wait what??? MC to Code A? Anyway...like I was saying. MC is bad. He got stomped and he is slumping because he is too cocky. There are no good Protoss players -- you guys have been so abusive for so long and now you've been figured out. Seriously.. l2p. / sarcasm It's honestly pretty true though. A lot of BW toss (marineking included) switched to T/Z because of the beta and because of how BW was. That, combined with the large number of Terrans switching, means that there are just more skilled Terrans (and somewhat Zergs) than there are Protoss. That isn't saying that "ALL PROTOSS PLAYERS ARE INFERIOR," but more so that using simple "X number of Code S/A protoss" isn't really a good way to measure the state of balance. Let's be honest. Inca's PvZ is terrible. Choya always has been super gimmicky, and HongUn is prone to losing games he shouldn't (and winning games he shouldn't) because of his strange playstyle. I don't disagree with anything you've said, it's certainly a valid theory. However, given the recent trend of certain strategies being abused beyond reason (1-1-1) as well as both DT/SG openings being easily deflectable by a smart Zerg, Protoss options are greatly limited. When a build as easy as 1-1-1 defeats Protoss players with alarming consistency, you need to consider attributing its success to the build inself instead of the Terran player's skill. Perhaps these GSL Protoss players aren't as bad as they look -- but rather confined by an inflexible and largely predictable race? This is also a valid theory.
I kind of wish more players switched races, so these things could be investigated at least in individual cases. Right now, we only have Byun, who was terrible at Protoss, and then won Code A after switching to T...
|
Why don't they bring back the shield battery? or some type of mechanic of the sort..
|
Blazinghand
United States25550 Posts
On September 08 2011 04:22 morpheus2480 wrote:
Why don't they bring back the shield battery? or some type of mechanic of the sort..
An important part of shield battery was that it was a gateway tech building as opposed to a forge tech building, so protoss players going for fast expands or dealing with marine rushes could build these quick buildings to supplement their armies defensively. However, due to the bursty nature of the 1/1/1 attack, its unlikely that units will be cycling in and out of the fray easily enough for the shield battery to have a big impact.
It might be useful for immortals, but it's unclear how such a mechanic would interact with hardened shields.
|
On September 08 2011 04:02 Toadvine wrote:Show nested quote +On September 08 2011 03:20 FairForever wrote:On September 08 2011 03:17 Elean wrote:On September 08 2011 02:49 merlin101 wrote:On September 08 2011 02:39 Dommk wrote:On September 08 2011 02:37 QTIP. wrote:On September 08 2011 02:20 MrCon wrote:
Winrates of GSL code A qualifiers : (numbers from liquipedia)
TvZ : 25-37 = 40%
TvP : 42-57 = 42%
PvZ : 42-49 = 41%
Yup - some Protoss actually qualified this time! ^_^ Fun fact, Terran has more players in Code S than Protoss has in Code S and A combined So what? There are more Terrans than Protoss and Zerg! Fun fact: Sample Size is waaay to small! Out of 64 players (code A + code S), half of them are terrans. The sample size is more than enough to be almost certain there is either an imbalance or a biased in the sample (i.e. terran players are more skilled). From a statistic point of view, saying the sample size isn't large enough is irrelevant. Now, of course you can argue the terran players are better. There is no reason for it to be the case, but nobody can prove the contrary. I personnaly believe that terran players in general are much less skilled than the other races. And the argument for it is quite simple: you improve more when you loose/struggle. Terran have been dominating since the beginning, so in term of strategy they must have fell far behind the other races. And there are signs of this. For instance, at MLG people just realized how strong blue flame were, some people said there even were imbalance (which Blizzard apparently agreed). But blue flame helions have been here since the beginning, how comes it took 1 year to realize it ? Well the answer is simple, terran had plenty of other strategies enabling them to win, so their skills/strategies improved slowly. Same thing can be said about 1/1/1. And, it's not the end, I'm sure there are plenty of other stong build terrans haven't figured out because they are already winning with non-optimal strategies. Anyone just said ghosts ? Did the test involve using t-test? Or what? The problem is that if you can't test this like you test a coin (with p = 1/3), because we don't expect 1/3 of the GSL to be Terran. Our expectation is that the ratio of Terran:Total in the GSL is equal to the ratio of Terran:Total in either Korea, Pro-Gaming Teams, or the World (depending on what you want to use - personally, I'd use Pro-Gaming teams). I don't have the ratio but if half of the players on Pro teams in Korea played Terran, then we'd expect 1/2 to be Terran. EDIT: But yes, if we could find evidence to suggest that our hypothesis should be that p = 1/3, then we could very easily say that there is virtually no possible chance that Terran isn't OP. I was intrigued by this, and went through the rosters of major KR teams on TLPD, just tallying up the numbers of each race. I don't necessarily think this is the right sample space, perhaps something like Korean GM would be better? In any case, it comes down to the following: Terran: 60 Protoss: 55 Zerg: 48 Total: 163 Which gives a 37/34/29 distribution, percentage-wise. One interesting fact I've noted, skimming through those player lists, is that most of the Terran names are known to me, and are either in Code S or Code A, while I haven't even heard of a lot of the Protoss players. So yeah, we might not get any more Terrans in GSL, because the pro teams simply don't have any more. I imagine the teams try to maintain a healthy race distribution for practice purposes, so unless they go on a crazy Terran shopping spree, we might reach a "Code S for Terrans, Code A for the other races" balanced state.
Hmm... those results would say that Terran is OP (I didn't even bother running the test, the current proportion in Code S of Terran is so high).
Korea GM isn't a good idea seeing as a lot of GMs don't attempt to qualify for GSL and a lot of the top players (eg. Nestea) aren't in GM.
I think the best sample size would be to use current GSL + attempted September Qualifiers players. Obviously you're missing out on a few big names (eg. Moon, Lyn) and you're including some people who have virtually no chance of making GSL regardless of racial imbalances, but you can then identify this group as the pool of competitors interested in participating in the GSL.
|
protoss is wrongly balanced because of their super trollship
wild thesis
edit: on a more serious note i REALLY think that a useable mothership would make stargate tech more viable than phoenix harassment which no one ever died from or void rays, which are strong but glass
|
Remove warp gate, remove mules, remove larva inject. Makes the game like five times easier to balance.
|
On September 08 2011 04:33 FairForever wrote:Show nested quote +On September 08 2011 04:02 Toadvine wrote:On September 08 2011 03:20 FairForever wrote:On September 08 2011 03:17 Elean wrote:On September 08 2011 02:49 merlin101 wrote:On September 08 2011 02:39 Dommk wrote:On September 08 2011 02:37 QTIP. wrote:On September 08 2011 02:20 MrCon wrote:
Winrates of GSL code A qualifiers : (numbers from liquipedia)
TvZ : 25-37 = 40%
TvP : 42-57 = 42%
PvZ : 42-49 = 41%
Yup - some Protoss actually qualified this time! ^_^ Fun fact, Terran has more players in Code S than Protoss has in Code S and A combined So what? There are more Terrans than Protoss and Zerg! Fun fact: Sample Size is waaay to small! Out of 64 players (code A + code S), half of them are terrans. The sample size is more than enough to be almost certain there is either an imbalance or a biased in the sample (i.e. terran players are more skilled). From a statistic point of view, saying the sample size isn't large enough is irrelevant. Now, of course you can argue the terran players are better. There is no reason for it to be the case, but nobody can prove the contrary. I personnaly believe that terran players in general are much less skilled than the other races. And the argument for it is quite simple: you improve more when you loose/struggle. Terran have been dominating since the beginning, so in term of strategy they must have fell far behind the other races. And there are signs of this. For instance, at MLG people just realized how strong blue flame were, some people said there even were imbalance (which Blizzard apparently agreed). But blue flame helions have been here since the beginning, how comes it took 1 year to realize it ? Well the answer is simple, terran had plenty of other strategies enabling them to win, so their skills/strategies improved slowly. Same thing can be said about 1/1/1. And, it's not the end, I'm sure there are plenty of other stong build terrans haven't figured out because they are already winning with non-optimal strategies. Anyone just said ghosts ? Did the test involve using t-test? Or what? The problem is that if you can't test this like you test a coin (with p = 1/3), because we don't expect 1/3 of the GSL to be Terran. Our expectation is that the ratio of Terran:Total in the GSL is equal to the ratio of Terran:Total in either Korea, Pro-Gaming Teams, or the World (depending on what you want to use - personally, I'd use Pro-Gaming teams). I don't have the ratio but if half of the players on Pro teams in Korea played Terran, then we'd expect 1/2 to be Terran. EDIT: But yes, if we could find evidence to suggest that our hypothesis should be that p = 1/3, then we could very easily say that there is virtually no possible chance that Terran isn't OP. I was intrigued by this, and went through the rosters of major KR teams on TLPD, just tallying up the numbers of each race. I don't necessarily think this is the right sample space, perhaps something like Korean GM would be better? In any case, it comes down to the following: Terran: 60 Protoss: 55 Zerg: 48 Total: 163 Which gives a 37/34/29 distribution, percentage-wise. One interesting fact I've noted, skimming through those player lists, is that most of the Terran names are known to me, and are either in Code S or Code A, while I haven't even heard of a lot of the Protoss players. So yeah, we might not get any more Terrans in GSL, because the pro teams simply don't have any more. I imagine the teams try to maintain a healthy race distribution for practice purposes, so unless they go on a crazy Terran shopping spree, we might reach a "Code S for Terrans, Code A for the other races" balanced state. Hmm... those results would say that Terran is OP (I didn't even bother running the test, the current proportion in Code S of Terran is so high). Korea GM isn't a good idea seeing as a lot of GMs don't attempt to qualify for GSL and a lot of the top players (eg. Nestea) aren't in GM. I think the best sample size would be to use current GSL + attempted September Qualifiers players. Obviously you're missing out on a few big names (eg. Moon, Lyn) and you're including some people who have virtually no chance of making GSL regardless of racial imbalances, but you can then identify this group as the pool of competitors interested in participating in the GSL.
I actually specifically included the WeMadeFox players, as well as "Korean" representatives of foreign teams, like PuMa or HuK. Still, I agree with including the pool of players attempting to qualify as our sample space. Not exactly sure how you can access this data, since after the recent change, the qualifiers are a huge ordeal (500+ people) - the players you see on Liquipedia are just the tip of the iceberg.
That said, I don't see any reason why the racial distribution in the qualifier would be markedly different than in Korean Masters and GM. And those are pretty similar to the numbers I've gotten for the pro teams, off the top of my head.
A good argument in favor of using the pro team numbers, is that practically no players from outside those teams actually succeed in qualifying. There's a total of one exception to this rule at this moment - ShinyStar - whose parents apparently don't agree to him joining a team.
|
United States7483 Posts
Terran dominating the GSL should have been expected once players started really getting good. They're the most complete race with the most viable tech options and opening variations, they have the strongest defenses and entrenching abilities, their units have the highest dps, their ground spellcaster destroys both zerg and protoss lategame (as well as countering their spellcasters), they have the strongest all-ins and all-in variations, best unit synergy, and highest skill-cap.
In retrospect, it's obvious that it was going to dominate GSL.
Edit2:
Anyways, I honestly think the 1/1/1 push is fine, yeah sure it's a bit hard but these days most pros are going for economic builds as Protoss anyways aren't they?
Good luck scouting terran's build in time to figure out what type of build to go with. If terran expands and you don't as toss, you're dead, due to no good early aggression anymore. If terran does a 1/1/1 and you go for econ, you're dead, unless you're extremely careful, lucky, and terran screws up. Terran has tons of other builds they can do too, like cloaked banshee or hellion drops or even a 3 rax all-in etc. There is too much variation from terran and most of it looks similar when your probe is chased out by the first marine, so yeah, you don't find out for sure what terran is doing until it's too late most of the time. Good luck with that.
|
On September 08 2011 04:45 Toadvine wrote:Show nested quote +On September 08 2011 04:33 FairForever wrote:On September 08 2011 04:02 Toadvine wrote:On September 08 2011 03:20 FairForever wrote:On September 08 2011 03:17 Elean wrote:On September 08 2011 02:49 merlin101 wrote:On September 08 2011 02:39 Dommk wrote:On September 08 2011 02:37 QTIP. wrote:On September 08 2011 02:20 MrCon wrote:
Winrates of GSL code A qualifiers : (numbers from liquipedia)
TvZ : 25-37 = 40%
TvP : 42-57 = 42%
PvZ : 42-49 = 41%
Yup - some Protoss actually qualified this time! ^_^ Fun fact, Terran has more players in Code S than Protoss has in Code S and A combined So what? There are more Terrans than Protoss and Zerg! Fun fact: Sample Size is waaay to small! Out of 64 players (code A + code S), half of them are terrans. The sample size is more than enough to be almost certain there is either an imbalance or a biased in the sample (i.e. terran players are more skilled). From a statistic point of view, saying the sample size isn't large enough is irrelevant. Now, of course you can argue the terran players are better. There is no reason for it to be the case, but nobody can prove the contrary. I personnaly believe that terran players in general are much less skilled than the other races. And the argument for it is quite simple: you improve more when you loose/struggle. Terran have been dominating since the beginning, so in term of strategy they must have fell far behind the other races. And there are signs of this. For instance, at MLG people just realized how strong blue flame were, some people said there even were imbalance (which Blizzard apparently agreed). But blue flame helions have been here since the beginning, how comes it took 1 year to realize it ? Well the answer is simple, terran had plenty of other strategies enabling them to win, so their skills/strategies improved slowly. Same thing can be said about 1/1/1. And, it's not the end, I'm sure there are plenty of other stong build terrans haven't figured out because they are already winning with non-optimal strategies. Anyone just said ghosts ? Did the test involve using t-test? Or what? The problem is that if you can't test this like you test a coin (with p = 1/3), because we don't expect 1/3 of the GSL to be Terran. Our expectation is that the ratio of Terran:Total in the GSL is equal to the ratio of Terran:Total in either Korea, Pro-Gaming Teams, or the World (depending on what you want to use - personally, I'd use Pro-Gaming teams). I don't have the ratio but if half of the players on Pro teams in Korea played Terran, then we'd expect 1/2 to be Terran. EDIT: But yes, if we could find evidence to suggest that our hypothesis should be that p = 1/3, then we could very easily say that there is virtually no possible chance that Terran isn't OP. I was intrigued by this, and went through the rosters of major KR teams on TLPD, just tallying up the numbers of each race. I don't necessarily think this is the right sample space, perhaps something like Korean GM would be better? In any case, it comes down to the following: Terran: 60 Protoss: 55 Zerg: 48 Total: 163 Which gives a 37/34/29 distribution, percentage-wise. One interesting fact I've noted, skimming through those player lists, is that most of the Terran names are known to me, and are either in Code S or Code A, while I haven't even heard of a lot of the Protoss players. So yeah, we might not get any more Terrans in GSL, because the pro teams simply don't have any more. I imagine the teams try to maintain a healthy race distribution for practice purposes, so unless they go on a crazy Terran shopping spree, we might reach a "Code S for Terrans, Code A for the other races" balanced state. Hmm... those results would say that Terran is OP (I didn't even bother running the test, the current proportion in Code S of Terran is so high). Korea GM isn't a good idea seeing as a lot of GMs don't attempt to qualify for GSL and a lot of the top players (eg. Nestea) aren't in GM. I think the best sample size would be to use current GSL + attempted September Qualifiers players. Obviously you're missing out on a few big names (eg. Moon, Lyn) and you're including some people who have virtually no chance of making GSL regardless of racial imbalances, but you can then identify this group as the pool of competitors interested in participating in the GSL. I actually specifically included the WeMadeFox players, as well as "Korean" representatives of foreign teams, like PuMa or HuK. Still, I agree with including the pool of players attempting to qualify as our sample space. Not exactly sure how you can access this data, since after the recent change, the qualifiers are a huge ordeal (500+ people) - the players you see on Liquipedia are just the tip of the iceberg. That said, I don't see any reason why the racial distribution in the qualifier would be markedly different than in Korean Masters and GM. And those are pretty similar to the numbers I've gotten for the pro teams, off the top of my head. A good argument in favor of using the pro team numbers, is that practically no players from outside those teams actually succeed in qualifying. There's a total of one exception to this rule at this moment - ShinyStar - whose parents apparently don't agree to him joining a team.
Are you sure? I thought the listing on TL for the September data is complete (hence there were Byes, as there weren't enough players - am I incorrect in saying this?)
Maybe we should use only pro team players who attempted to qualify in September (or are already on a team). TBH though I don't think it's going to change our hypothesis significantly enough to not reject the hypothesis.
|
On September 08 2011 05:00 FairForever wrote:Show nested quote +On September 08 2011 04:45 Toadvine wrote:On September 08 2011 04:33 FairForever wrote:On September 08 2011 04:02 Toadvine wrote:On September 08 2011 03:20 FairForever wrote:On September 08 2011 03:17 Elean wrote:On September 08 2011 02:49 merlin101 wrote:On September 08 2011 02:39 Dommk wrote:On September 08 2011 02:37 QTIP. wrote:On September 08 2011 02:20 MrCon wrote:
Winrates of GSL code A qualifiers : (numbers from liquipedia)
TvZ : 25-37 = 40%
TvP : 42-57 = 42%
PvZ : 42-49 = 41%
Yup - some Protoss actually qualified this time! ^_^ Fun fact, Terran has more players in Code S than Protoss has in Code S and A combined So what? There are more Terrans than Protoss and Zerg! Fun fact: Sample Size is waaay to small! Out of 64 players (code A + code S), half of them are terrans. The sample size is more than enough to be almost certain there is either an imbalance or a biased in the sample (i.e. terran players are more skilled). From a statistic point of view, saying the sample size isn't large enough is irrelevant. Now, of course you can argue the terran players are better. There is no reason for it to be the case, but nobody can prove the contrary. I personnaly believe that terran players in general are much less skilled than the other races. And the argument for it is quite simple: you improve more when you loose/struggle. Terran have been dominating since the beginning, so in term of strategy they must have fell far behind the other races. And there are signs of this. For instance, at MLG people just realized how strong blue flame were, some people said there even were imbalance (which Blizzard apparently agreed). But blue flame helions have been here since the beginning, how comes it took 1 year to realize it ? Well the answer is simple, terran had plenty of other strategies enabling them to win, so their skills/strategies improved slowly. Same thing can be said about 1/1/1. And, it's not the end, I'm sure there are plenty of other stong build terrans haven't figured out because they are already winning with non-optimal strategies. Anyone just said ghosts ? Did the test involve using t-test? Or what? The problem is that if you can't test this like you test a coin (with p = 1/3), because we don't expect 1/3 of the GSL to be Terran. Our expectation is that the ratio of Terran:Total in the GSL is equal to the ratio of Terran:Total in either Korea, Pro-Gaming Teams, or the World (depending on what you want to use - personally, I'd use Pro-Gaming teams). I don't have the ratio but if half of the players on Pro teams in Korea played Terran, then we'd expect 1/2 to be Terran. EDIT: But yes, if we could find evidence to suggest that our hypothesis should be that p = 1/3, then we could very easily say that there is virtually no possible chance that Terran isn't OP. I was intrigued by this, and went through the rosters of major KR teams on TLPD, just tallying up the numbers of each race. I don't necessarily think this is the right sample space, perhaps something like Korean GM would be better? In any case, it comes down to the following: Terran: 60 Protoss: 55 Zerg: 48 Total: 163 Which gives a 37/34/29 distribution, percentage-wise. One interesting fact I've noted, skimming through those player lists, is that most of the Terran names are known to me, and are either in Code S or Code A, while I haven't even heard of a lot of the Protoss players. So yeah, we might not get any more Terrans in GSL, because the pro teams simply don't have any more. I imagine the teams try to maintain a healthy race distribution for practice purposes, so unless they go on a crazy Terran shopping spree, we might reach a "Code S for Terrans, Code A for the other races" balanced state. Hmm... those results would say that Terran is OP (I didn't even bother running the test, the current proportion in Code S of Terran is so high). Korea GM isn't a good idea seeing as a lot of GMs don't attempt to qualify for GSL and a lot of the top players (eg. Nestea) aren't in GM. I think the best sample size would be to use current GSL + attempted September Qualifiers players. Obviously you're missing out on a few big names (eg. Moon, Lyn) and you're including some people who have virtually no chance of making GSL regardless of racial imbalances, but you can then identify this group as the pool of competitors interested in participating in the GSL. I actually specifically included the WeMadeFox players, as well as "Korean" representatives of foreign teams, like PuMa or HuK. Still, I agree with including the pool of players attempting to qualify as our sample space. Not exactly sure how you can access this data, since after the recent change, the qualifiers are a huge ordeal (500+ people) - the players you see on Liquipedia are just the tip of the iceberg. That said, I don't see any reason why the racial distribution in the qualifier would be markedly different than in Korean Masters and GM. And those are pretty similar to the numbers I've gotten for the pro teams, off the top of my head. A good argument in favor of using the pro team numbers, is that practically no players from outside those teams actually succeed in qualifying. There's a total of one exception to this rule at this moment - ShinyStar - whose parents apparently don't agree to him joining a team. Are you sure? I thought the listing on TL for the September data is complete (hence there were Byes, as there weren't enough players - am I incorrect in saying this?) Maybe we should use only pro team players who attempted to qualify in September (or are already on a team). TBH though I don't think it's going to change our hypothesis significantly enough to not reject the hypothesis.
I was under the impression that there's a Bo1 round before the one shown as the first on Liquipedia. Might be wrong, don't remember where I got that. And there's no mention of it anywhere, so it may as well not be there at all, and I'm talking out of my ass.
But yeah, I can't imagine it will change the relative numbers all that much.
|
On September 08 2011 03:57 Zzoram wrote:
HotS needs to give Protoss a fast cheap harass unit, something they sorely lack.
We are getting one, hopefully HotS is coming not before late... :/
|
On September 07 2011 15:11 XD_Melchior wrote:Show nested quote +On September 07 2011 07:03 Hikari wrote:
My opinion of mules:
- Necessary to terran early game economy, esp before full mineral saturation.
- They offer a unique "investment" option for terrans. You feel slightly ahead, but not by much? Build an extra CC and turn it into an OC. 2.5 mules late you regained your investment and every extra mule become pure profit - with the added benefit of extra scv production (think of it as investing 550 for constant CB on probes ~ investment comes back in 3 minutes later)
- Very late game they allow terran to "live off" mules: leaving behind only maybe 12 scvs for gas = bigger army, but these situations 30 minutes into the game can be quite rare and unique.
- There is a tactical aspect to mules: you can save them up and float an OC over to a gold and spam mules. The gold will be depleted in record times. In combination of the "investment OC" option, this is a very strong economical benefit.
I've always thought that giving the OC a MULE range (like 9) would be a good fix. Terran can still use MULEs just as much in the early/mid game, the only difference is that they have to use them at the base the OC is at. (e.g. instead of going to base #2 and using 2 MULEs, they'll have to use 1 MULE at base #1 and 1 MULE at base #2. Minor inconvenience, but I assure you still nowhere as near as bad as larva injecting.) So what does that do? That helps take out the imbalance of dumping insane MULEs from all your OCs at a gold. AND the imbalance of taking a gold with a PF then dumping MULEs from other OCs there. AND the imbalance of floating to an island and dumping MULEs there. BTW, if your main runs you, you can always float that OC to another mining base to get double MULEs. Or, you can keep the OC there and have that as your designated scan base.
OMG, a limitation on mule drop point from the OC. you sir have created the most useful idea that doesn't nerf MULES at all directly but forces terran to be mobile and plan ahead with mineral consumption of bases, and thus makes terran like P and Z again.
This should be implemented immediately.
|
On September 08 2011 05:55 Truedot wrote:Show nested quote +On September 07 2011 15:11 XD_Melchior wrote:On September 07 2011 07:03 Hikari wrote:
My opinion of mules:
- Necessary to terran early game economy, esp before full mineral saturation.
- They offer a unique "investment" option for terrans. You feel slightly ahead, but not by much? Build an extra CC and turn it into an OC. 2.5 mules late you regained your investment and every extra mule become pure profit - with the added benefit of extra scv production (think of it as investing 550 for constant CB on probes ~ investment comes back in 3 minutes later)
- Very late game they allow terran to "live off" mules: leaving behind only maybe 12 scvs for gas = bigger army, but these situations 30 minutes into the game can be quite rare and unique.
- There is a tactical aspect to mules: you can save them up and float an OC over to a gold and spam mules. The gold will be depleted in record times. In combination of the "investment OC" option, this is a very strong economical benefit.
I've always thought that giving the OC a MULE range (like 9) would be a good fix. Terran can still use MULEs just as much in the early/mid game, the only difference is that they have to use them at the base the OC is at. (e.g. instead of going to base #2 and using 2 MULEs, they'll have to use 1 MULE at base #1 and 1 MULE at base #2. Minor inconvenience, but I assure you still nowhere as near as bad as larva injecting.) So what does that do? That helps take out the imbalance of dumping insane MULEs from all your OCs at a gold. AND the imbalance of taking a gold with a PF then dumping MULEs from other OCs there. AND the imbalance of floating to an island and dumping MULEs there. BTW, if your main runs you, you can always float that OC to another mining base to get double MULEs. Or, you can keep the OC there and have that as your designated scan base. OMG, a limitation on mule drop point from the OC. you sir have created the most useful idea that doesn't nerf MULES at all directly but forces terran to be mobile and plan ahead with mineral consumption of bases, and thus makes terran like P and Z again. This should be implemented immediately.
You would be seeing a lot of multi-CC bases once the first couple get mined out
|
On September 08 2011 06:06 Ovreel wrote:Show nested quote +On September 08 2011 05:55 Truedot wrote:On September 07 2011 15:11 XD_Melchior wrote:On September 07 2011 07:03 Hikari wrote:
My opinion of mules:
- Necessary to terran early game economy, esp before full mineral saturation.
- They offer a unique "investment" option for terrans. You feel slightly ahead, but not by much? Build an extra CC and turn it into an OC. 2.5 mules late you regained your investment and every extra mule become pure profit - with the added benefit of extra scv production (think of it as investing 550 for constant CB on probes ~ investment comes back in 3 minutes later)
- Very late game they allow terran to "live off" mules: leaving behind only maybe 12 scvs for gas = bigger army, but these situations 30 minutes into the game can be quite rare and unique.
- There is a tactical aspect to mules: you can save them up and float an OC over to a gold and spam mules. The gold will be depleted in record times. In combination of the "investment OC" option, this is a very strong economical benefit.
I've always thought that giving the OC a MULE range (like 9) would be a good fix. Terran can still use MULEs just as much in the early/mid game, the only difference is that they have to use them at the base the OC is at. (e.g. instead of going to base #2 and using 2 MULEs, they'll have to use 1 MULE at base #1 and 1 MULE at base #2. Minor inconvenience, but I assure you still nowhere as near as bad as larva injecting.) So what does that do? That helps take out the imbalance of dumping insane MULEs from all your OCs at a gold. AND the imbalance of taking a gold with a PF then dumping MULEs from other OCs there. AND the imbalance of floating to an island and dumping MULEs there. BTW, if your main runs you, you can always float that OC to another mining base to get double MULEs. Or, you can keep the OC there and have that as your designated scan base. OMG, a limitation on mule drop point from the OC. you sir have created the most useful idea that doesn't nerf MULES at all directly but forces terran to be mobile and plan ahead with mineral consumption of bases, and thus makes terran like P and Z again. This should be implemented immediately. You would be seeing a lot of multi-CC bases once the first couple get mined out
and? I have no problem with them bloating their infrastructure to get an advantage. that means they're fielding less units.
I highly doubt anyone would do that.
|
hi im in bronze league and i think that cheese is overpowered because it wins often and im not pro yet in my micro. in macro i always win but my micro is not pro yet so i lose vs cheese i think blizzard should nerf it.
sorry for bad english
hope i could help
|
On September 08 2011 06:08 Truedot wrote:Show nested quote +On September 08 2011 06:06 Ovreel wrote:On September 08 2011 05:55 Truedot wrote:On September 07 2011 15:11 XD_Melchior wrote:On September 07 2011 07:03 Hikari wrote:
My opinion of mules:
- Necessary to terran early game economy, esp before full mineral saturation.
- They offer a unique "investment" option for terrans. You feel slightly ahead, but not by much? Build an extra CC and turn it into an OC. 2.5 mules late you regained your investment and every extra mule become pure profit - with the added benefit of extra scv production (think of it as investing 550 for constant CB on probes ~ investment comes back in 3 minutes later)
- Very late game they allow terran to "live off" mules: leaving behind only maybe 12 scvs for gas = bigger army, but these situations 30 minutes into the game can be quite rare and unique.
- There is a tactical aspect to mules: you can save them up and float an OC over to a gold and spam mules. The gold will be depleted in record times. In combination of the "investment OC" option, this is a very strong economical benefit.
I've always thought that giving the OC a MULE range (like 9) would be a good fix. Terran can still use MULEs just as much in the early/mid game, the only difference is that they have to use them at the base the OC is at. (e.g. instead of going to base #2 and using 2 MULEs, they'll have to use 1 MULE at base #1 and 1 MULE at base #2. Minor inconvenience, but I assure you still nowhere as near as bad as larva injecting.) So what does that do? That helps take out the imbalance of dumping insane MULEs from all your OCs at a gold. AND the imbalance of taking a gold with a PF then dumping MULEs from other OCs there. AND the imbalance of floating to an island and dumping MULEs there. BTW, if your main runs you, you can always float that OC to another mining base to get double MULEs. Or, you can keep the OC there and have that as your designated scan base. OMG, a limitation on mule drop point from the OC. you sir have created the most useful idea that doesn't nerf MULES at all directly but forces terran to be mobile and plan ahead with mineral consumption of bases, and thus makes terran like P and Z again. This should be implemented immediately. You would be seeing a lot of multi-CC bases once the first couple get mined out and? I have no problem with them bloating their infrastructure to get an advantage. that means they're fielding less units. I highly doubt anyone would do that.
Whooooosh
|
Protoss can't expo at the moment without beeing too vulnerable/harmless.
e.g. forge fast expand. zerg goes for a quick 3rd and toss falls behind for miles.
transition into stargte/harrass: toss MUST deal tons of damage, to stay equal in economy.
pvt 1-1-1 or a concussive/stim/emp push at the 9 min. mark. it takes miracles to defend that.
problem imo:
toss needs more money, since the stuff is too expensive and the build times very long.
my suggestion:
make stalker 100/50. this gives toss the ability to expand safer. getting faster tech or a few more units/workers out.
in a 15 min. game: lets say there are 20 stalkers made, that would be 500 minerals more. that doesnt sound like too much of a buff.
an immortal range buff will not solve the current problem imo.
|
|
|
|
|
|
EPT
