|
BGonline.org Forums
Some more stats
Posted By: Achim In Response To: Some stats (Did I ever say that i hate maths?) (Chuck Bower)
Date: Thursday, 5 July 2007, at 3:16 p.m.
Well ... I still did it wrong ... somehow. I took the wrong table, so you either need to divide all percentages by 2 and add to 50 or just read it here again and trust me:
gnu/jelly 5.88 100% (nearly)
gnu/snowie .69 74%
gnu/bgb 1.2 88%
bgb/jelly 2.15 98%
bgb/snowie 0.25 59.8%
snowie/jelly 4.56 100% (nearly)
Sorry in case I did confuse some people and thx for watching my progresses in statistics ;-).
Christopher Yep also mentioned some other stats regarding the all over results:
Total wins:
gnubg 1623
snowie 1557
bgb 1519
jelly 1301
equals ... 6000 (puh)
gnubg's std.err. 27.3 ---> 95% of all results will be between 1568 - 1678
snowie's std.err. 27.37 ---> 95% of all results will be between 1502 - 1612
bgb's std.err. 27.38 --->95% of all results will be between 1464 - 1574
jelly's std.err 27.1 ---> 95% of all results will be between 1246 - 1356
So the only thing we can be pretty confident of is that jellyfish is ranked 4th and gnubg is likely to be ranked 1st (if all 4 bots play against each other!).
It differs when we take only 3 bots (gnubg/snowie/bgb) into account:
gnubg 1030 wins
bgb 985 wins
snowie 985 wins
... 3000 (puh)
gnubg's std.err. 22.35 --->985 - 1075
bgb's std. err: 22.35 --->940 - 1030
snowie's std.err. 2235--->940 -1030
so no confidence at all besides that BGBlitz has reached the Olymp also. Though I don't understand bgblitz' result against jellyfish.
Ciao
Achim
PS. Don't trust my numbers, calculate it for your own ;-).
|
BGonline.org Forums is maintained by Stick with WebBBS 5.12.