On October 20 2009 12:30 zulu_nation8 wrote: how am i dismissing the line of reasoning? I'm saying 1, your sample size is bad, 2, when something that looks very much like variance occurs in a set of data you ignore everything and jump to the conclusion that something is "imbalanced." If what you actually wanted to bring to attention at first is indeed variance then you should be using words like trends, shift, or whatever, and not "imbalance."
Is this what all of your hostility in this thread comes from? The fact that I used one word instead of another? Fine. I wish I hadn't named this thread what I did.
My sample size is most definitely big enough for this difference in zerg winrate to be statistically significant. Unfortunately, I've only been able to calculate winrate on individual maps, as TLPD doesn't allow you to sort by matchup when you're looking at all the matches in the database.
show me your calculations that show the difference to be statistically significant
Like I said, I can't do it for the whole matchup without a lot of work. But I can do it for the combined stats of Neo Medusa, Destination, and Outsider since March (those are the only stats I have handy). BRB.
Ah, I'm on a Mac right now so I don't have access to Stata. This might have to wait.
I think race mu imbalances are irrelevant and it's all map dependant. Like you can make any map you want and have be super imbalanced towards any race in any mu you want. I think map makers have for a long time tried to find a map model that will give as balanced as possible results and their current model (desti/medusa macro maps) is slightly skewed and needs more tweaking
On October 20 2009 12:30 zulu_nation8 wrote: how am i dismissing the line of reasoning? I'm saying 1, your sample size is bad, 2, when something that looks very much like variance occurs in a set of data you ignore everything and jump to the conclusion that something is "imbalanced." If what you actually wanted to bring to attention at first is indeed variance then you should be using words like trends, shift, or whatever, and not "imbalance."
Is this what all of your hostility in this thread comes from? The fact that I used one word instead of another? Fine. I wish I hadn't named this thread what I did.
My sample size is most definitely big enough for this difference in zerg winrate to be statistically significant. Unfortunately, I've only been able to calculate winrate on individual maps, as TLPD doesn't allow you to sort by matchup when you're looking at all the matches in the database.
show me your calculations that show the difference to be statistically significant
Like I said, I can't do it for the whole matchup without a lot of work. But I can do it for the combined stats of Neo Medusa, Destination, and Outsider since March (those are the only stats I have handy). BRB.
Ah, I'm on a Mac right now so I don't have access to Stata. This might have to wait.
wow nice dodge
What kind of tests are you possible running that can't be done through any kind of excel clone? You don't need stata for algebra
OK I just found a much easier way to compile map matchup data! So when I get access to Stata, I'll have better data. I'll do this for all stats since March 1st, 2009.
Byzantium 3: 25-13 Byzantium 2: 30-11 Tears of the Moon: 1-0 New Autumn Wind: 3-1 Medusa: 34-23 Tau Cross: 7-7 Carthage 2: 2-4 Carthage: 0-1 Battle Royale: 4-5 Holy World: 4-3 Shades of Twilight: 1-3 Colosseum II: 2-4 Andromeda: 7-19 (?????) Neo Harmony: 5-0 God's Garden: 56-44 Carthage 3: 1-0 Outsider: 41-27 Neo Medusa: 34-25 Return of the King: 47-22 Eye of the Storm: 1-1 El Niño: 1-1 Destination: 110-72 (this changed significantly since the time of the OP... EVER OSL prelims used it) Tornado: 5-1 Outsider SE: 2-0 Moon Glaive: 2-3 Match Point: 3-4 Heartbreak Ridge: 90-64 Fighting Spirit: 6-3
On October 20 2009 12:30 zulu_nation8 wrote: how am i dismissing the line of reasoning? I'm saying 1, your sample size is bad, 2, when something that looks very much like variance occurs in a set of data you ignore everything and jump to the conclusion that something is "imbalanced." If what you actually wanted to bring to attention at first is indeed variance then you should be using words like trends, shift, or whatever, and not "imbalance."
Is this what all of your hostility in this thread comes from? The fact that I used one word instead of another? Fine. I wish I hadn't named this thread what I did.
My sample size is most definitely big enough for this difference in zerg winrate to be statistically significant. Unfortunately, I've only been able to calculate winrate on individual maps, as TLPD doesn't allow you to sort by matchup when you're looking at all the matches in the database.
show me your calculations that show the difference to be statistically significant
Like I said, I can't do it for the whole matchup without a lot of work. But I can do it for the combined stats of Neo Medusa, Destination, and Outsider since March (those are the only stats I have handy). BRB.
Ah, I'm on a Mac right now so I don't have access to Stata. This might have to wait.
wow nice dodge
What kind of tests are you possible running that can't be done through any kind of excel clone? You don't need stata for algebra
If you can do a probit test w/o a stats program, be my guest. I only know how to do it with Stata, sorry :\
EDIT: shouldn't be a probit test, actually. One sec... trying to figure out what test I need to run. I might be able to just do it in excel.
There are two issues. The first is what you mean by imbalance. If by imbalance you mean an advantage given to a certain race through how the game is designed, or some innate problem within the gameplay mechanics of starcraft, and thus something which should reveal itself in good statistical analysis, then and only then the term "statistically significant" comes into play here. You would need a large and significant enough sample to be able to call a game imbalanced, and hopefully you agree that 200 games played on three maps is not a big enough sample size from which to make any statements regarding the game engine of BW.
If this is not what you mean by imbalance, in that you only use the word in a casual sense to describe exactly what you presented in the op; how progamer zergs have been winning more games on destination, medusa, and heartbreak over a certain period of time, then you should have never used the word imbalance to begin with, and clarify in your op the exact conditions of your data. The only question that should be asked would then be, "why have progamer zergs been winning more games on destination, medusa, and heartbreak over the last however long months?" Some explanations could be that the six dragons started playing wow, or that the 3 hatch to 5 hatch build is hard to play against. Whatever the explanation is, you would have to present addition evidence, and you haven't.
On October 20 2009 13:58 zulu_nation8 wrote: There are two issues. The first is what you mean by imbalance. If by imbalance you mean an advantage given to a certain race through how the game is designed, or some innate problem within the gameplay mechanics of starcraft, and thus something which should reveal itself in good statistical analysis, then and only then the term "statistically significant" comes into play here. You would need a large and significant enough sample to be able to call a game imbalanced, and hopefully you agree that 200 games played on three maps is not a big enough sample size from which to make any statements regarding the game engine of BW.
If this is not what you mean by imbalance, in that you only use the word in a casual sense to describe exactly what you presented in the op; how progamer zergs have been winning more games on destination, medusa, and heartbreak over a certain period of time, then you should have never used the word imbalance to begin with, and clarify in your op the exact conditions of your data. The only question that should be asked would then be, "why have progamer zergs been winning more games on destination, medusa, and heartbreak over the last however long months?" Some explanations could be that the six dragons started playing wow, or that the 3 hatch to 5 hatch build is hard to play against. Whatever the explanation is, you would have to present addition evidence, and you haven't.
I'm done talking to you. You obviously haven't been reading my posts, such as the FUCKING OP, which clearly states that I think maps can correct any imbalance, which means that I OBVIOUSLY don't mean that I think there's "an advantage given to a certain race through how the game is designed" since I think that maps are what make the difference! And as for your second point, take a look at this post that I made one fucking hour ago:
On October 20 2009 12:30 zulu_nation8 wrote: how am i dismissing the line of reasoning? I'm saying 1, your sample size is bad, 2, when something that looks very much like variance occurs in a set of data you ignore everything and jump to the conclusion that something is "imbalanced." If what you actually wanted to bring to attention at first is indeed variance then you should be using words like trends, shift, or whatever, and not "imbalance."
Is this what all of your hostility in this thread comes from? The fact that I used one word instead of another? Fine. I wish I hadn't named this thread what I did.
I already admitted that the word "imbalanced" is not the ideal one here.
On October 20 2009 13:58 zulu_nation8 wrote: There are two issues. The first is what you mean by imbalance. If by imbalance you mean an advantage given to a certain race through how the game is designed, or some innate problem within the gameplay mechanics of starcraft, and thus something which should reveal itself in good statistical analysis, then and only then the term "statistically significant" comes into play here. You would need a large and significant enough sample to be able to call a game imbalanced, and hopefully you agree that 200 games played on three maps is not a big enough sample size from which to make any statements regarding the game engine of BW.
If this is not what you mean by imbalance, in that you only use the word in a casual sense to describe exactly what you presented in the op; how progamer zergs have been winning more games on destination, medusa, and heartbreak over a certain period of time, then you should have never used the word imbalance to begin with, and clarify in your op the exact conditions of your data. The only question that should be asked would then be, "why have progamer zergs been winning more games on destination, medusa, and heartbreak over the last however long months?" Some explanations could be that the six dragons started playing wow, or that the 3 hatch to 5 hatch build is hard to play against. Whatever the explanation is, you would have to present addition evidence, and you haven't.
I'm done talking to you. You obviously haven't been reading my posts, such as the FUCKING OP, which clearly states that I think maps can correct any imbalance, which means that I OBVIOUSLY don't mean that I think there's "an advantage given to a certain race through how the game is designed" since I think that maps are what make the difference! And as for your second point, take a look at this post that I made one fucking hour ago:
On October 20 2009 12:30 zulu_nation8 wrote: how am i dismissing the line of reasoning? I'm saying 1, your sample size is bad, 2, when something that looks very much like variance occurs in a set of data you ignore everything and jump to the conclusion that something is "imbalanced." If what you actually wanted to bring to attention at first is indeed variance then you should be using words like trends, shift, or whatever, and not "imbalance."
Is this what all of your hostility in this thread comes from? The fact that I used one word instead of another? Fine. I wish I hadn't named this thread what I did.
I already admitted that the word "imbalanced" is not the ideal one here.
you said that maps will "correct" imbalance? I don't understand how thats relevant at all to what I said? Whats the imbalance maps are correcting? If you're saying that the idea of imbalance wouldn't occur outside of maps then that would also be pretty stupid.
Yes you realize that you used the wrong word, but I don't think you understand why, thus I explained to you, it's not a simple "misuse", it's a misunderstanding of concept.
On October 20 2009 14:09 heyoka wrote: Your standard deviation is .49? I don't think you have the slightest clue what you're measuring man.
Don't be condescending; I'm an econ major. I'm measuring 1's and 0's. 1's are zerg wins and 0's are protoss wins. I put them seperately in a column in Excel and took the stdev of the column. Excel spat out .49
On October 20 2009 14:18 zulu_nation8 wrote: why would the null hypothesis be 50% though? I'm pretty sure the mean of zvp stats in every 7 month period in progaming is not 50%?
The concept of a null hypothesis works like this: I believe that the true win rate of zergs against Protoss is something greater than 50%. I have data to support this theory. The data gives me a figure of ~59%, but that stat MAY be due to chance. That's why I use the null hypothesis: "What is the chance that ZvP is actually totally balanced since March 1st, and the data is just a fluke?" The z-test tests the likelyhood that the data is a fluke.
The result was that it's very unlikely that ZvP is actually 50% and that the data was a fluke.
On October 20 2009 14:19 heyoka wrote: And you don't think there is anything wrong with your null hypothesis being "Distribution = .5" and then using a standard deviation of .49?
(for what its worth your null should also be "this <time period> is different than the previous history of progaming" - where ZvP is actually like 53% but thats a minor issue)
I think the only way we are going to resolve this is if you do the test and try to prove me wrong, cause I don't know where you're going with this.
EPT
