|
I apologize to everyone in this thread for taking the OP seriously. My mod senses are definitely off today.
-- Chill |
Having studied probability in the past, I have to agree with those saying that 100 attempts is not a large enough sample size to be conclusive.
If each possible position setup (close air, close ground, cross) are exactly 33% likely...then it's quite possible to get many of the same setup in a row. Those that claim that the probability decreases with each successive match are wrong...chances will remain 33%;
http://en.wikipedia.org/wiki/Gambler's_fallacy
Selecting an appropriet sample size is actually something that can be quite difficult. I'd suggest a sample size of 2,000 in order to generate a confidence level of 99%. 1,000 for a 95% confidence level. (A confidence of 100% is not possible)
|
On February 16 2011 05:08 Deadeight wrote:Show nested quote +On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. I know that if you analyse it statistically 100 samples is enough. The chance of getting that ratio is pretty low. And as Frozenserpent said the P value would be low. But (if I understand OPs post correctly) the sample was not random. He looked at his last 100 games that he'd played on it right? Or did he play an extra 100 games? Out of all the people who play it's pretty likely that this will have happened to someone. And when it does happen to someone that person will notice it. Would be good if there was a way to check this without checking individual replays.
On February 16 2011 04:45 the p00n wrote:
I found this pretty weird so I actually started testing it against the computer.
On February 16 2011 05:09 Liquid`Tyler wrote:Show nested quote +On February 16 2011 05:02 TBO wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. The problem here is that you have millions of players who play 100 games on metalopolis and a few of them will get extreme results (and those will post in the forums), even if it is 1/3 chance. If you have 100.000 people throwing a coin 15 times, you will get quite a few (6 in average) who will get a 15-0 or 0-15 result. Only if lots of people get the same results as the topic creator, one could assume it is conclusive. Show nested quote +On February 16 2011 05:05 Soma.bokforlag wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. you shouldnt call people retarded when you obviously doesnt understand statistics. if you make enough studdies some of them will turn out faulty results even if the method is correct 99.9% of the time. in this example, it is possible that the author of the thread felt "damn, i get alot of close positions" and therefore examined his stats.. the rest 99.99% of players which have a more even distribution never gave it a thought and probably are closer to correct ratios edit. TBO was quicker than me.. Ah, you guys aren't likely to get your PHD's when you can't even understand the OP's testing methods. He didn't examine the games that gave him the feeling that he got close positions. He got that feeling and ignored those games and loaded up 100 more games against the computer and recorded those results.
My bad.
|
On February 16 2011 05:09 Liquid`Tyler wrote:Show nested quote +On February 16 2011 05:02 TBO wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. The problem here is that you have millions of players who play 100 games on metalopolis and a few of them will get extreme results (and those will post in the forums), even if it is 1/3 chance. If you have 100.000 people throwing a coin 15 times, you will get quite a few (6 in average) who will get a 15-0 or 0-15 result. Only if lots of people get the same results as the topic creator, one could assume it is conclusive. Show nested quote +On February 16 2011 05:05 Soma.bokforlag wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. you shouldnt call people retarded when you obviously doesnt understand statistics. if you make enough studdies some of them will turn out faulty results even if the method is correct 99.9% of the time. in this example, it is possible that the author of the thread felt "damn, i get alot of close positions" and therefore examined his stats.. the rest 99.99% of players which have a more even distribution never gave it a thought and probably are closer to correct ratios edit. TBO was quicker than me.. Ah, you guys aren't likely to get your PHD's when you can't even understand the OP's testing methods. He didn't examine the games that gave him the feeling that he got close positions. He got that feeling and ignored those games and loaded up 100 more games against the computer and recorded those results.
Could still be coincidence.
The problem with this "test" is that it is a single sample, although the sample size is fairly large. Accurate conclusions can't be reached unless more samples like this one are taken.
I haven't conducted the hypothesis test for this sample but it's likely that he'd get a p-value less than 0.05. However, this particular sample could just be an outlier. Many more samples like this one would need to be taken to draw conclusions. Is that asking too much? Probably, but until it happens this sample really doesn't conclude anything.
|
Ok here are my stats. I only included ladder games I played as Zerg (got a few as random), not custom games. They're autosaved, so no discrimination.
Close : 16
Air : 14
Cross : 15
Total : 45 games
As you can see, that's pretty equal. However on the past month, i've had way more close position, but i cant draw conclusion from that.
|
Sample size of 100 proves nothing. Don't call people retarded and get a clue about what you are talking.
Thank you. Because isn't randomness random? You can never be sure. Even a million tests may still prove to get a 7:1:2 ratio. You never know!
|
On February 16 2011 04:52 magha wrote:Taking sample sizes of 100 would prove that nobody in history has ever won a lottery.
Lol
Lotto winning chances should theoretically be 1/# of ticket buyers.
Close Spawn chances should theoretically by 1/3
nice analogy man
|
On February 16 2011 05:01 Alpina wrote:Show nested quote +On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. Sample size of 100 proves nothing. Don't call people retarded and get a clue about what you are talking.
While I don't agree with calling people "fucking retarded", you need a better understanding of statistics before you tell people to get a clue. Sample size doesn't mean anything, it's p-value that matters. And you don't even need to calculate the p-value to know it's way less than 0.05 if something happens 72 times out of 100 when its expected value is 33.
|
On February 16 2011 05:12 Alpina wrote:Show nested quote +On February 16 2011 05:09 Liquid`Tyler wrote:On February 16 2011 05:02 TBO wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. The problem here is that you have millions of players who play 100 games on metalopolis and a few of them will get extreme results (and those will post in the forums), even if it is 1/3 chance. If you have 100.000 people throwing a coin 15 times, you will get quite a few (6 in average) who will get a 15-0 or 0-15 result. Only if lots of people get the same results as the topic creator, one could assume it is conclusive. On February 16 2011 05:05 Soma.bokforlag wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. you shouldnt call people retarded when you obviously doesnt understand statistics. if you make enough studdies some of them will turn out faulty results even if the method is correct 99.9% of the time. in this example, it is possible that the author of the thread felt "damn, i get alot of close positions" and therefore examined his stats.. the rest 99.99% of players which have a more even distribution never gave it a thought and probably are closer to correct ratios edit. TBO was quicker than me.. Ah, you guys aren't likely to get your PHD's when you can't even understand the OP's testing methods. He didn't examine the games that gave him the feeling that he got close positions. He got that feeling and ignored those games and loaded up 100 more games against the computer and recorded those results. So he could still be wrong, 100 games cannot prove anything, can they? People are saying that even if the chance of getting close positions is 1/3 the statistics made from 100 games can still be wrong..
You can ALWAYS be wrong.
The thing is, with these possible outcomes and this samples size test, its extremely unlikely.
|
On February 16 2011 04:59 RoyalCheese wrote:because : ![[image loading]](https://blog.startcom.org/wp-content/uploads/2008/05/debian-rng.jpg) !
Hahahahahahaha, I don't know why, but I cannot stop rofling from that.
|
whole lot of derp derp in this thread. Sample size is sufficient, an increase wouldn't be detrimental though.
As for the actual point of this thread, I hate this. It is Statistically significant, and strategically as well. Close positions always leads to a super gay cheese fest in my experience. I fist pump everytime I spawn cross position, and I can either 15pool 16 hatch, 1 rax expo, or gate cyber expo. Making it to the mid/late game can be difficult on close positions. Not saying it's difficult to win, it's just less fun to play in my opinion.
|
8751 Posts
On February 16 2011 05:12 Alpina wrote:Show nested quote +On February 16 2011 05:09 Liquid`Tyler wrote:On February 16 2011 05:02 TBO wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. The problem here is that you have millions of players who play 100 games on metalopolis and a few of them will get extreme results (and those will post in the forums), even if it is 1/3 chance. If you have 100.000 people throwing a coin 15 times, you will get quite a few (6 in average) who will get a 15-0 or 0-15 result. Only if lots of people get the same results as the topic creator, one could assume it is conclusive. On February 16 2011 05:05 Soma.bokforlag wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. you shouldnt call people retarded when you obviously doesnt understand statistics. if you make enough studdies some of them will turn out faulty results even if the method is correct 99.9% of the time. in this example, it is possible that the author of the thread felt "damn, i get alot of close positions" and therefore examined his stats.. the rest 99.99% of players which have a more even distribution never gave it a thought and probably are closer to correct ratios edit. TBO was quicker than me.. Ah, you guys aren't likely to get your PHD's when you can't even understand the OP's testing methods. He didn't examine the games that gave him the feeling that he got close positions. He got that feeling and ignored those games and loaded up 100 more games against the computer and recorded those results. So he could still be wrong, 100 games cannot prove anything, can they? People are saying that even if the chance of getting close positions is 1/3 the statistics made from 100 games can still be wrong..
You are correct to say that it's wrong to utter the words proof and conclusion at this point, but the 100 game sample size shows that it's more likely than not (>50% chance) that close positions happen more than 1/3 of the time.
edit: And yeah, it's a lot higher than 50%, but since I haven't done the math I don't know what it is exactly and will leave that to someone else.
|
It would be nice if we could make a small test where everybody could fill in their last game on Meta and note what positions it was --> that way we will get a ton of results which should be accurate
|
Pretty sure i dont spawn close position more often than the others. I dont know anything about statistics (besides 9th grade math  ), but i am almost 100% certain that 76% of my games arent close positions. I think this is just a coincidence tbh, had to happen to someone.
|
On February 16 2011 05:09 Liquid`Tyler wrote:Show nested quote +On February 16 2011 05:02 TBO wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. The problem here is that you have millions of players who play 100 games on metalopolis and a few of them will get extreme results (and those will post in the forums), even if it is 1/3 chance. If you have 100.000 people throwing a coin 15 times, you will get quite a few (6 in average) who will get a 15-0 or 0-15 result. Only if lots of people get the same results as the topic creator, one could assume it is conclusive. Show nested quote +On February 16 2011 05:05 Soma.bokforlag wrote:On February 16 2011 04:58 Frozenserpent wrote:
People are fucking retarded and need to learn some basic statistics before they say something like "not big enough sample size".
Even 50 can be sufficient to obtain a p-value < 5%.
In this case, p-value is definitely low enough to reject the assumption that it's 1/3rd.
Obviously if you want the exact ratios you'd want to expand on the larger sample size, but this is conclusive to determine that it's not 1/3rds. you shouldnt call people retarded when you obviously doesnt understand statistics. if you make enough studdies some of them will turn out faulty results even if the method is correct 99.9% of the time. in this example, it is possible that the author of the thread felt "damn, i get alot of close positions" and therefore examined his stats.. the rest 99.99% of players which have a more even distribution never gave it a thought and probably are closer to correct ratios edit. TBO was quicker than me.. Ah, you guys aren't likely to get your PHD's when you can't even understand the OP's testing methods. He didn't examine the games that gave him the feeling that he got close positions. He got that feeling and ignored those games and loaded up 100 more games against the computer and recorded those results.
indeed, sorry, i just wanted to lecture him for calling people retarded 
|
If my calculations are correct, the chance of this is less than 3 * 10^-15 ... not sure how much less likely you would want before saying the sample size is sufficient.
That being said, someone else should recreate this test before concluding something is wrong.
|
i think 100 is pretty good for this situation..hmm that is interesting indeed..
|
The sample size is fine, but I don't think the test is conclusive. If the test is accurate (the RNG is spitting out bad numbers) then you should be able to run the test again and get something reasonably close.
|
If someone wants to denounce the theory claiming that is based on statistical error. They need to run their own test. He gave a method in his first post.
Run 100 games against the computer and post your results. I'm not at home right now so I can't do it, or I would.
Statistically 100 is enough of a sample size. However there is still like a 5% chance that his numbers are off due to statistical error.
In a scientific study though, if you disagree with the original author of the study, its your job to run his test and show that your results are not in alignment with his.
-Cent
|
Calgary26003 Posts
Interesting. Instead of arguing about the reasonable sample size, let's either:
a) Figure out how we're going to test it with a larger sample size; or, prefereably
b) Figure out why Metalopolis is prone to spawning clone positions.
|
On February 16 2011 05:15 Vei wrote:
Lotto winning chances should theoretically be 1/# of ticket buyers.
Depends. If it's instant lottery, yeah.
But anything where you can chose numbers, it's quite possible to have no winners
|
|
|
|
|
|
EPT
