EPTI play fastest, where most 1v1's are random vs. random. I was wondering what the chances are that the game will become a mirror matchups. I have to say that I play almost no mirror matchups, but when I think about the chances, it seems as though the probability for mirror matchups is higher than the amount I receive. Why is this?
Probability of Mirror Matchups in rvr
| Forum Index > BW General |
|
Qizix
3 Posts
I play fastest, where most 1v1's are random vs. random. I was wondering what the chances are that the game will become a mirror matchups. I have to say that I play almost no mirror matchups, but when I think about the chances, it seems as though the probability for mirror matchups is higher than the amount I receive. Why is this? | ||
|
Jayme
United States5866 Posts
It's simply random. Edit: I'm a moron. Chances are 1/3. Lol lol | ||
]343[
United States10328 Posts
pick any race for 1st (probability 1), then probability of 2nd being same is 1/3. or 3 mirror matchups out of 3*3 matchups = 1/3? | ||
|
Senix
Germany149 Posts
| ||
|
Tracil
Australia505 Posts
| ||
|
Zherak
Norway256 Posts
The only thing I can say completely confidently is that it is not random. That is, you do not get mirrors every third game as one would expect. I seem to remember the odds coming out as 1 out of 9. Not as in 1/9 for TvT, but 1/9 for TvT, PvP and ZvZ all added together. If four players all go random, you will get 2/1/1 distribution nearly every single time. I ran a series of test because I considered it ridiculous for random to not actually be random, but, well, it isn't. Just like miss chance isn't 30% I guess. | ||
|
Qizix
3 Posts
For example, on vile gaming, there have been 2,791 matches played. Out of those 186 games have been mirror matches. That multiplied by nine equals 1674 matches. Why is it that the numbers aren't closer together? Edit: Didn't see the previous post, but that is what I was thinking of. | ||
|
]343[
United States10328 Posts
On March 16 2009 08:21 Zherak wrote: There was a huge thread about this a long time ago. The only thing I can say completely confidently is that it is not random. That is, you do not get mirrors every third game as one would expect. I seem to remember the odds coming out as 1 out of 9. Not as in 1/9 for TvT, but 1/9 for TvT, PvP and ZvZ all added together. If four players all go random, you will get 2/1/1 distribution nearly every single time. I ran a series of test because I considered it ridiculous for random to not actually be random, but, well, it isn't. Just like miss chance isn't 30% I guess. hmm I wouldn't be surprised if this were true however 2/1/1 has... multinomial(4, (2, 1, 1)) = 4!/2! = 12 ways [assuming the 2 race is fixed, so x3], and 1/3^4 = 81 probability/thing, so... the probability it's 2/1/1 is 4/9 (which is pretty high) :O | ||
|
Faronel
United States658 Posts
I guess it makes sense that it's not random since blizzard probably only but 1 byte of info for the random race generator: i.e. 1 byte = 8 choices --> pseudorandom http://en.wikipedia.org/wiki/Pseudorandom_number_generator | ||
|
]343[
United States10328 Posts
On March 16 2009 08:26 Faronel wrote: Zherak that idea is wrong in this case, since we're looking for mirror match ups. It's still 1/3 when at the end because the 1st choice can be anything from the 3/3/2 distribution and the 2nd choice is also from the 3/3/2 distribution, being that the zerg (2) 's low probability is offset by the protoss and terran probabilities, it will be 1/3. I guess it makes sense that it's not random since blizzard probably only but 1 byte of info for the random race generator: i.e. 1 byte = 8 choices --> pseudorandom http://en.wikipedia.org/wiki/Pseudorandom_number_generator wait so you're saying chance for zerg is only 1/4? interesting... | ||
|
Patriot.dlk
Sweden5462 Posts
![[image loading]](http://www.teamliquid.net/staff/Hot_Bid/3.jpg) | ||
|
Kaolla
China2999 Posts
On March 16 2009 08:28 Patriot.dlk wrote: This huge post has it all covered, written by blizzard guy: 2nd time ive seen that today, supposed to be funny? oo; | ||
|
Avidkeystamper
United States8554 Posts
| ||
|
Zherak
Norway256 Posts
http://www.teamliquid.net/forum/viewmessage.php?topic_id=37059¤tpage=2 That ought to cover it. | ||
MasterOfChaos
Germany2896 Posts
| ||
|
Zherak
Norway256 Posts
| ||
|
Nytefish
United Kingdom4282 Posts
| ||
|
WhuazGoodJaggah
Lesotho777 Posts
http://www.bwhacks.com/forums/starcraft-hacking-related/30728-random-race-test.html | ||
|
SonuvBob
Aiur21549 Posts
On March 16 2009 08:49 MasterOfChaos wrote: There is only possible explanation for the deviation from 1/3: Blizzard has invented the improbability drive. Even with infinite improbability Blizzard still can't release a game on time. | ||
| ||
