|
On April 12 2013 06:59 SlixSC wrote:Show nested quote +On April 12 2013 05:41 Big J wrote:On April 12 2013 04:46 SlixSC wrote:aligulac is not very reliable. they are missing roughly 25% of all games played just in the GSL qualifiers alone. If you really want to have accurate statistics you can't just leave out 25% of all the games played in any given tournament. a) as long as the left out games are arbitrary it doesn't matter statistically b) qualifiers shouldn't be in there anyways, because everyone can participate and therefore you get many pro vs nonpro games which are won by the pro no matter what. And therefore the winrates turn out closer to 50:50 than they should be. c) I wish people would stop calling every statistic inaccurate for some reason or another. Yeah sure they are inaccurate. It's statistic. It's always going to be inaccurate to some degree. Be happy that we get stuff to work with, even if it is not perfect. You realize that I'm not just complaining but actually doing something and making more accurate statistics. And "as long as the left out games are arbitrary" is a meaningless statement, how could you possibly prove that they were left out arbitrarily. And a) is strongly contradicated by b), you understand the central limit theorem, but only apply it to a) but not to b). Which makes me think that you are heavily biased, for whatever reason.
a) doesn't contradict b)
The underlying assumption is that we only want pro-vs-pro games in out statistics, because pro-vs-nonpro doesn't give us any information.
And if there is a set limit like "all qualifier games don't count, no matter whether it is pro or nonpro", or "your first 5games won't make it into the stats, because up to then we don't consider you a highlevel player" we still don't screw the stats.
|
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup.
|
On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup.
So whether it is 1a or not I dont want to discuss... But how much of a choice does P have?
They dont have medivacs themself, they dont have speedlings or mutas or hellions. Against a bio T Protoss only can sit tight and mass up until they have enough to move out and leave stuf behind...
|
On April 12 2013 13:15 iky43210 wrote:Show nested quote +On April 12 2013 12:45 SlixSC wrote:On April 12 2013 12:21 iky43210 wrote:On April 12 2013 12:11 SlixSC wrote:On April 12 2013 12:05 Dudasc wrote:On April 12 2013 11:57 iky43210 wrote:On April 12 2013 10:07 Merkmerk wrote:On April 12 2013 06:59 SlixSC wrote:On April 12 2013 05:41 Big J wrote:On April 12 2013 04:46 SlixSC wrote:
[quote]
aligulac is not very reliable. they are missing roughly 25% of all games played just in the GSL qualifiers alone. If you really want to have accurate statistics you can't just leave out 25% of all the games played in any given tournament.
a) as long as the left out games are arbitrary it doesn't matter statistically b) qualifiers shouldn't be in there anyways, because everyone can participate and therefore you get many pro vs nonpro games which are won by the pro no matter what. And therefore the winrates turn out closer to 50:50 than they should be. c) I wish people would stop calling every statistic inaccurate for some reason or another. Yeah sure they are inaccurate. It's statistic. It's always going to be inaccurate to some degree. Be happy that we get stuff to work with, even if it is not perfect. You realize that I'm not just complaining but actually doing something and making more accurate statistics. And "as long as the left out games are arbitrary" is a meaningless statement, how could you possibly prove that they were left out arbitrarily. And a) is strongly contradicated by b), you understand the central limit theorem, but only apply it to a) but not to b). Which makes me think that you are heavily biased, for whatever reason. Oh shut the hell up. He provided ACTUAL STATISTICAL DATA that shows a significant imbalance in TvZ - one that appears to be getting worse based on looking at masters/GM TvZ win %s. You simply don't like the stat because you're a plat player like me and sometimes you get beaten by a Zerg who outplays you so therefore you think it must be balanced. Just hush. when we can have people like polt coming into NA GM with something like 90% winrates smashing through everybody, what use are those statistics? what you are getting is literally how many korean terran smurfs over at NA. Look at dragon, inno, polt, forgg etc. Just look at the GSL qualifier. Twice as many zerg pass over Terran, sure is imbalance showing. One can argue that the game is not figured out yet but denying the fact that terrans are over 60% win rate vs zerg in pro matches is just being deliberately stupid. I'm deliberately stupid then. https://docs.google.com/spreadsheet/ccc?key=0At0PE4rdhsI9dDE0cEprWkwwMGxQdTczTTlLNW1qX1E#gid=0It's actually 55.2% for April so far. Which is still in favor of Terran but let's actually stick with real numbers here and not just fabricate the numbers we want to use just to prove a point. I did mention no shady tourneys right? The only tourney on that list that isn't a complete joke is GSL and PL thus far. Almost as good as using playheim in part of calculating your statistics Every time you mix in foreign tourneys and korean tourneys together for statistics and what you get is not an accurate portray of anything, but rather who and what korean was invited to said foreign tourneys Adding any korean will create the bias Ok, statistically speaking TvZ win rate in GSL is 50.3%, where exactly did you get 60%+ from? On April 12 2013 12:32 lowercase wrote:Honestly, this spreadsheet is a bit of a mess. I can't understand what you're trying to show here. All I'm saying is that I watch a lot of Starcraft, and I have since Brood War, and anecdotally it feels like Terran is a bit strong right now, mostly in TvP. And I can tell you that Protoss has a 60.4% win rate in TvP this month, now what is more relevant your personal opinion or actual statistics? you're confusing me with someone else. I'm just having doubts that terran are broken like many ppl suggested (this thread, idra, catz etc)
Idra and Catz are not really the best to ask about balance. I mean Idra lost to dragon in a macro game so.... Just because they are "pros" doesn´t mean they know a lot about designing and balancing a complex game.
I think TvP balance is the same as in wol: The Terran has to win and the Protoss just has to not die. As long as Protoss has the best splash damage and the insanity of warpgate tech the lategame will be tough for terran.
The only things that really changed with hots in my opinion are these 2 situation:
Basetrades and low eco midgames
Basetrades as terran against P was always tough the way the macro mechanics work but it was kind of ok but with the MSC and the free recall P can kill the expansions/production and go back to defend a far away expansion. The recall just saves 20 or 30 seconds walking over the map doing no damage.
The MSC changes low eco situations too. For example a scenario dt rush against double drop which kills a lot of eco at both sides. In WoL the Terran was in pretty good shape in these situation because of mules and the fact that bio was very effective in low number battles against warpgate units. But with the nexus cannon there is no way to break the protoss front and they know this. So they are able to rebuild worker and tech at the same time without building any units and still survive little marine squads at the front.
|
On April 12 2013 18:54 USvBleakill wrote:Show nested quote +On April 12 2013 13:15 iky43210 wrote:On April 12 2013 12:45 SlixSC wrote:On April 12 2013 12:21 iky43210 wrote:On April 12 2013 12:11 SlixSC wrote:On April 12 2013 12:05 Dudasc wrote:On April 12 2013 11:57 iky43210 wrote:On April 12 2013 10:07 Merkmerk wrote:On April 12 2013 06:59 SlixSC wrote:On April 12 2013 05:41 Big J wrote:
[quote]
a) as long as the left out games are arbitrary it doesn't matter statistically
b) qualifiers shouldn't be in there anyways, because everyone can participate and therefore you get many pro vs nonpro games which are won by the pro no matter what. And therefore the winrates turn out closer to 50:50 than they should be.
c) I wish people would stop calling every statistic inaccurate for some reason or another. Yeah sure they are inaccurate. It's statistic. It's always going to be inaccurate to some degree. Be happy that we get stuff to work with, even if it is not perfect. You realize that I'm not just complaining but actually doing something and making more accurate statistics. And "as long as the left out games are arbitrary" is a meaningless statement, how could you possibly prove that they were left out arbitrarily. And a) is strongly contradicated by b), you understand the central limit theorem, but only apply it to a) but not to b). Which makes me think that you are heavily biased, for whatever reason. Oh shut the hell up. He provided ACTUAL STATISTICAL DATA that shows a significant imbalance in TvZ - one that appears to be getting worse based on looking at masters/GM TvZ win %s. You simply don't like the stat because you're a plat player like me and sometimes you get beaten by a Zerg who outplays you so therefore you think it must be balanced. Just hush. when we can have people like polt coming into NA GM with something like 90% winrates smashing through everybody, what use are those statistics? what you are getting is literally how many korean terran smurfs over at NA. Look at dragon, inno, polt, forgg etc. Just look at the GSL qualifier. Twice as many zerg pass over Terran, sure is imbalance showing. One can argue that the game is not figured out yet but denying the fact that terrans are over 60% win rate vs zerg in pro matches is just being deliberately stupid. I'm deliberately stupid then. https://docs.google.com/spreadsheet/ccc?key=0At0PE4rdhsI9dDE0cEprWkwwMGxQdTczTTlLNW1qX1E#gid=0It's actually 55.2% for April so far. Which is still in favor of Terran but let's actually stick with real numbers here and not just fabricate the numbers we want to use just to prove a point. I did mention no shady tourneys right? The only tourney on that list that isn't a complete joke is GSL and PL thus far. Almost as good as using playheim in part of calculating your statistics Every time you mix in foreign tourneys and korean tourneys together for statistics and what you get is not an accurate portray of anything, but rather who and what korean was invited to said foreign tourneys Adding any korean will create the bias Ok, statistically speaking TvZ win rate in GSL is 50.3%, where exactly did you get 60%+ from? On April 12 2013 12:32 lowercase wrote:Honestly, this spreadsheet is a bit of a mess. I can't understand what you're trying to show here. All I'm saying is that I watch a lot of Starcraft, and I have since Brood War, and anecdotally it feels like Terran is a bit strong right now, mostly in TvP. And I can tell you that Protoss has a 60.4% win rate in TvP this month, now what is more relevant your personal opinion or actual statistics? you're confusing me with someone else. I'm just having doubts that terran are broken like many ppl suggested (this thread, idra, catz etc) Idra and Catz are not really the best to ask about balance.... I agree on Catz, because he is playing almost different game on his level. But Idra apart from playing watches a lot of tournaments. And if you exclude his rage comments, he might be actually the best progamer to listen to. If you went through all Inside the game episodes and check his predictions and analyzes, you would find that he was mostly right. But majority of people listen just to his "protoss fucking op" and play deaf while he is not raging.
|
On April 12 2013 18:34 Big J wrote:Show nested quote +On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup. So whether it is 1a or not I dont want to discuss... But how much of a choice does P have? They dont have medivacs themself, they dont have speedlings or mutas or hellions. Against a bio T Protoss only can sit tight and mass up until they have enough to move out and leave stuf behind...
They have warp prisms, they have DT's, they have many harass options that people just aren't using.
|
On April 12 2013 19:26 Chaggi wrote:Show nested quote +On April 12 2013 18:34 Big J wrote:On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup. So whether it is 1a or not I dont want to discuss... But how much of a choice does P have? They dont have medivacs themself, they dont have speedlings or mutas or hellions. Against a bio T Protoss only can sit tight and mass up until they have enough to move out and leave stuf behind... They have warp prisms, they have DT's, they have many harass options that people just aren't using.
People are opening dts or putting them into play more now that the shrine is cheaper so I wouldn't say its not used and warp prism isn't the same as a medivac worth of units and stim so that's why it's not used as much. Not to mention it cuts into robo build time.
|
On April 12 2013 19:38 Sc2zero7 wrote:Show nested quote +On April 12 2013 19:26 Chaggi wrote:On April 12 2013 18:34 Big J wrote:On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup. So whether it is 1a or not I dont want to discuss... But how much of a choice does P have? They dont have medivacs themself, they dont have speedlings or mutas or hellions. Against a bio T Protoss only can sit tight and mass up until they have enough to move out and leave stuf behind... They have warp prisms, they have DT's, they have many harass options that people just aren't using. People are opening dts or putting them into play more now that the shrine is cheaper so I wouldn't say its not used and warp prism isn't the same as a medivac worth of units and stim so that's why it's not used as much. Not to mention it cuts into robo build time.
Of course it's not the same, it's a different race. But I posted this on another topic...
While I strongly agree that Protoss is pretty trash without hitting that critical mass (a horrible design outcome), Protoss harass options aren't as limited as you think. You mentioned that Terran can just kite the zealots. Well, sure, to be cost effective, we have to kite.
Imagine a situation where you warp in 4-6 zealots in a 3rd or 4th, drop a warp prism full of zealots in the main. Terran HAS to micro or commit a lot of units to each base's defense. And then the Protoss just pushes into maybe the natural or whatever they want to kill. That could very well be a killing blow. There isn't many Terrans that can defend an attack from multiple sides, and deal with a deathball that requires splitting, kiting and EMP's just to survive. You don't do one of them, and the Protoss lands a storm or two, you're pretty dead.
Of course that takes a lot of coordination, but Protoss losing a few zealots to push the Terran army out of position and force hard decisions is not a bad thing. At worst, if the Terran decides to say f it and push at Protoss, move back, turtle up, get a concave and land the good storms/feedback.
People seem to have this misconception that Protoss can't do anything until they have the perfect 200/200 deathball and honestly that's not true.
|
On April 12 2013 19:38 Sc2zero7 wrote:Show nested quote +On April 12 2013 19:26 Chaggi wrote:On April 12 2013 18:34 Big J wrote:On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup. So whether it is 1a or not I dont want to discuss... But how much of a choice does P have? They dont have medivacs themself, they dont have speedlings or mutas or hellions. Against a bio T Protoss only can sit tight and mass up until they have enough to move out and leave stuf behind... They have warp prisms, they have DT's, they have many harass options that people just aren't using. People are opening dts or putting them into play more now that the shrine is cheaper so I wouldn't say its not used and warp prism isn't the same as a medivac worth of units and stim so that's why it's not used as much. Not to mention it cuts into robo build time.
And this imo illustrates the problem quite some toss have: they are just too much focussed on min/maxing their deathball. Yes it cuts into robo build time and your main army is a bit weaker if you make a warp prism, but the idea is that you make that worth it by harassing.
As other example, I have had games with I had 6+ cloakshees vs a toss, and he just couldnt deal with their harrasment. So afterwards I asked him why on earth he didn't make a stargate and a single phoenix (was in WoL): He told me he had to use those resources on colossi. So he rather lost tons of resources both in direct losses and lost mining time than shut down the harrasment by making something that didn't fit in his deathball. Not to mention he also never realised that by investing that much in cloakshees my main army was also significantly weaker.
|
On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup.
Perhaps you need to work on your mindset? TvP is in a very good place for terran in the current metagame, perhaps you should ask for advice in the strategy forum? Your post certainly doesn't belong in a balance discussion.
|
On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup.
Then you're playing it wrong. Stop whining and figure out what you're doing wrong and get better.
|
On April 12 2013 16:25 Big J wrote:Show nested quote +On April 12 2013 06:59 SlixSC wrote:On April 12 2013 05:41 Big J wrote:On April 12 2013 04:46 SlixSC wrote:aligulac is not very reliable. they are missing roughly 25% of all games played just in the GSL qualifiers alone. If you really want to have accurate statistics you can't just leave out 25% of all the games played in any given tournament. a) as long as the left out games are arbitrary it doesn't matter statistically b) qualifiers shouldn't be in there anyways, because everyone can participate and therefore you get many pro vs nonpro games which are won by the pro no matter what. And therefore the winrates turn out closer to 50:50 than they should be. c) I wish people would stop calling every statistic inaccurate for some reason or another. Yeah sure they are inaccurate. It's statistic. It's always going to be inaccurate to some degree. Be happy that we get stuff to work with, even if it is not perfect. You realize that I'm not just complaining but actually doing something and making more accurate statistics. And "as long as the left out games are arbitrary" is a meaningless statement, how could you possibly prove that they were left out arbitrarily. And a) is strongly contradicated by b), you understand the central limit theorem, but only apply it to a) but not to b). Which makes me think that you are heavily biased, for whatever reason. a) doesn't contradict b) The underlying assumption is that we only want pro-vs-pro games in out statistics, because pro-vs-nonpro doesn't give us any information. And if there is a set limit like "all qualifier games don't count, no matter whether it is pro or nonpro", or "your first 5games won't make it into the stats, because up to then we don't consider you a highlevel player" we still don't screw the stats.
Of course it is a contradiction. Why would you ignore the central limit theorem in b) and assume that the ratio of pros to non-pros differs considerably from race to race? You are assuming that one (or even two) race(s) had better players (on average!) participate in the tournament than one (or two of the) other race(s).
You would essentially have to prove that there were more amateur T players participating in the tournament than amateur P or Z players, which then skewed the results in favor of P and Z. But the basic assumption is of course that, that wasn't the case and if you want to argue that it was you actually have to go out and prove it.
|
On April 12 2013 19:26 Chaggi wrote:Show nested quote +On April 12 2013 18:34 Big J wrote:On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup. So whether it is 1a or not I dont want to discuss... But how much of a choice does P have? They dont have medivacs themself, they dont have speedlings or mutas or hellions. Against a bio T Protoss only can sit tight and mass up until they have enough to move out and leave stuf behind... They have warp prisms, they have DT's, they have many harass options that people just aren't using.
Yeah, but not many cheap ones and most of them are not on the way to what P needs to hold/execute head on pushes. Unlike speedlings/speedroaches medivacs, biounits, mutalisks. A dark shrine + some DTs is not the same as just sending a drop or a runby - of stuff you have anyways - whenever you realize that you are not threatened.
|
Why don't protoss use that recall more, they can take out a base and just recall. I think they are just bad so the protoss you play are always worse then you in the league you are in.
|
On April 12 2013 20:23 SlixSC wrote:Show nested quote +On April 12 2013 16:25 Big J wrote:On April 12 2013 06:59 SlixSC wrote:On April 12 2013 05:41 Big J wrote:On April 12 2013 04:46 SlixSC wrote:aligulac is not very reliable. they are missing roughly 25% of all games played just in the GSL qualifiers alone. If you really want to have accurate statistics you can't just leave out 25% of all the games played in any given tournament. a) as long as the left out games are arbitrary it doesn't matter statistically b) qualifiers shouldn't be in there anyways, because everyone can participate and therefore you get many pro vs nonpro games which are won by the pro no matter what. And therefore the winrates turn out closer to 50:50 than they should be. c) I wish people would stop calling every statistic inaccurate for some reason or another. Yeah sure they are inaccurate. It's statistic. It's always going to be inaccurate to some degree. Be happy that we get stuff to work with, even if it is not perfect. You realize that I'm not just complaining but actually doing something and making more accurate statistics. And "as long as the left out games are arbitrary" is a meaningless statement, how could you possibly prove that they were left out arbitrarily. And a) is strongly contradicated by b), you understand the central limit theorem, but only apply it to a) but not to b). Which makes me think that you are heavily biased, for whatever reason. a) doesn't contradict b) The underlying assumption is that we only want pro-vs-pro games in out statistics, because pro-vs-nonpro doesn't give us any information. And if there is a set limit like "all qualifier games don't count, no matter whether it is pro or nonpro", or "your first 5games won't make it into the stats, because up to then we don't consider you a highlevel player" we still don't screw the stats. Of course it is a contradiction. Why would you ignore the central limit theorem in b) and assume that the ratio of pros to non-pros differs considerably from race to race? You are assuming that one (or even two) race(s) had better players (on average!) participate in the tournament than one (or two of the) other race(s). You would essentially have to prove that there were more amateur T players participating in the tournament than amateur P or Z players, which then skewed the results in favor of P and Z. But the basic assumption is of course that, that wasn't the case and if you want to argue that it was you actually have to go out and prove it.
Im not. Im assuming that the number of pros is actually quite even.
Lets say we have MU AvsB with 600-400 in favor of A in terms of progames. So balance should be considered 60-40. Now we hadd 300wins to each race from balanceirelevant pro vs nonpro games. 900-700, so suddenly the stats show 56-44.
(though this is of course dependend on the player numbers of each race. But the results will most likely be screwed, either they look too 50:50 or they look too onesided)
And Im not implying anything for these stats. I just want to point out that is better to not use open tournaments first rounds.
|
ZvT was broken for 12 months, with random euro/NA scrubs regularly taking games off top Koreans. How can anyone know if T is currently OP, or whether the trash just hasn't gotten cleaned out yet?
Are Stephano, Bly, and Vibe as good as some of the terrans they still take games off of?
And where are the NA/Euro terrans that are dominating with their new supposedly OP race?
Fact is, the top Korean zergs will adjust. Life has already shown how effective Muta/ling/bane can be vs widow mines and drops yet guys like Stephano still insist on building roach hydra....
|
On April 12 2013 20:29 Big J wrote:Show nested quote +On April 12 2013 19:26 Chaggi wrote:On April 12 2013 18:34 Big J wrote:On April 12 2013 18:24 Duncaaaaaan wrote:
I still hate how protoss can mass up a 1A deathball and A move, and terran completely melts to it.
TvP for me is 40%, such a dumb fucking matchup. So whether it is 1a or not I dont want to discuss... But how much of a choice does P have? They dont have medivacs themself, they dont have speedlings or mutas or hellions. Against a bio T Protoss only can sit tight and mass up until they have enough to move out and leave stuf behind... They have warp prisms, they have DT's, they have many harass options that people just aren't using. Yeah, but not many cheap ones and most of them are not on the way to what P needs to hold/execute head on pushes. Unlike speedlings/speedroaches medivacs, biounits, mutalisks. A dark shrine + some DTs is not the same as just sending a drop or a runby - of stuff you have anyways - whenever you realize that you are not threatened.
200 minerals + 400 minerals for zealots vs 100 minerals/100 gas for medivac + 50 * 8 = 400 minerals, it's very comparable and zealots are sturdier and you can reinforce with the warp prism consistently.
As far as getting the warp prism, yes it may not, Protoss's will need to harass a bit later, but the option is there, especially after establishing the 3rd. It may not be as fast but you can warp units in place and have them go in as soon as you see a move out, which is common among pros, as we just saw Life do in GSTL vs Maru. Every race doesn't need to have the same type of harassements, and honestly Protoss does just fine.
|
On April 12 2013 20:40 tskarzyn wrote:
ZvT was broken for 12 months, with random euro/NA scrubs regularly taking games off top Koreans. How can anyone know if T is currently OP, or whether the trash just hasn't gotten cleaned out yet?
Are Stephano, Bly, and Vibe as good as some of the terrans they still take games off of?
And where are the NA/Euro terrans that are dominating with their new supposedly OP race?
Fact is, the top Korean zergs will adjust. Life has already shown how effective Muta/ling/bane can be vs widow mines and drops yet guys like Stephano still insist on building roach hydra....
For NA/Euro Terrans to dominate, there needs to be some to begin with.
|
On April 12 2013 20:45 Chaggi wrote:Show nested quote +On April 12 2013 20:40 tskarzyn wrote:
ZvT was broken for 12 months, with random euro/NA scrubs regularly taking games off top Koreans. How can anyone know if T is currently OP, or whether the trash just hasn't gotten cleaned out yet?
Are Stephano, Bly, and Vibe as good as some of the terrans they still take games off of?
And where are the NA/Euro terrans that are dominating with their new supposedly OP race?
Fact is, the top Korean zergs will adjust. Life has already shown how effective Muta/ling/bane can be vs widow mines and drops yet guys like Stephano still insist on building roach hydra.... For NA/Euro Terrans to dominate, there needs to be some to begin with.
There were tons of EU/NA Terrans in the early days of WoL but they all quit/disappeared from the scene after Terran was nerfed into the ground. Terran has always had the highest skill ceiling, even in BW, and this is the reason that it is almost exclusively Koreans that play it at the pro level. Debating whether someone like Major is better than Vibe and so on is pointless, just look at Euro vs Korean representation.
|
On April 12 2013 20:40 Big J wrote:Show nested quote +On April 12 2013 20:23 SlixSC wrote:On April 12 2013 16:25 Big J wrote:On April 12 2013 06:59 SlixSC wrote:On April 12 2013 05:41 Big J wrote:On April 12 2013 04:46 SlixSC wrote:aligulac is not very reliable. they are missing roughly 25% of all games played just in the GSL qualifiers alone. If you really want to have accurate statistics you can't just leave out 25% of all the games played in any given tournament. a) as long as the left out games are arbitrary it doesn't matter statistically b) qualifiers shouldn't be in there anyways, because everyone can participate and therefore you get many pro vs nonpro games which are won by the pro no matter what. And therefore the winrates turn out closer to 50:50 than they should be. c) I wish people would stop calling every statistic inaccurate for some reason or another. Yeah sure they are inaccurate. It's statistic. It's always going to be inaccurate to some degree. Be happy that we get stuff to work with, even if it is not perfect. You realize that I'm not just complaining but actually doing something and making more accurate statistics. And "as long as the left out games are arbitrary" is a meaningless statement, how could you possibly prove that they were left out arbitrarily. And a) is strongly contradicated by b), you understand the central limit theorem, but only apply it to a) but not to b). Which makes me think that you are heavily biased, for whatever reason. a) doesn't contradict b) The underlying assumption is that we only want pro-vs-pro games in out statistics, because pro-vs-nonpro doesn't give us any information. And if there is a set limit like "all qualifier games don't count, no matter whether it is pro or nonpro", or "your first 5games won't make it into the stats, because up to then we don't consider you a highlevel player" we still don't screw the stats. Of course it is a contradiction. Why would you ignore the central limit theorem in b) and assume that the ratio of pros to non-pros differs considerably from race to race? You are assuming that one (or even two) race(s) had better players (on average!) participate in the tournament than one (or two of the) other race(s). You would essentially have to prove that there were more amateur T players participating in the tournament than amateur P or Z players, which then skewed the results in favor of P and Z. But the basic assumption is of course that, that wasn't the case and if you want to argue that it was you actually have to go out and prove it. Im not. Im assuming that the number of pros is actually quite even. Lets say we have MU AvsB with 600-400 in favor of A in terms of progames. So balance should be considered 60-40. Now we hadd 300wins to each race from balanceirelevant pro vs nonpro games. 900-700, so suddenly the stats show 56-44. (though this is of course dependend on the player numbers of each race. But the results will most likely be screwed, either they look too 50:50 or they look too onesided) And Im not implying anything for these stats. I just want to point out that is better to not use open tournaments first rounds.
Even if that was the case, the argument people make in this thread is that T>Z and T>P. When actual statistics show that T>Z (though not to the extent people think) and P>T. Even if I ignored R1 of the GSL Qualifiers this would have next to no impact on these statistics.
I accept your point but I don't think I'm justified in ignoring some of the games because some of the players "aren't pro enough". It's too subjective.
|
|
|
|
|
|
EPT
