Author: Robert Hyatt
Date: 11:52:52 01/29/04
Go up one level in this thread
On January 29, 2004 at 13:49:24, Stefan Zipproth wrote: >>>1 Fritz 8 >>>½½0½1½0000000½1½00101½010½10½110½11011111111000110½01½11½½½1½10½100½½101110½1½½½0½101½001011½½001010 >>>51.5/100 >>>2 Shredder 8 >>>½½1½0½1111111½0½11010½101½01½001½00100000000111001½10½00½½½0½01½011½½010001½0½½½1½010½110100½½110101 >>>48.5/100 >>> >>>Note, that there is one 8 and one 7 wins in row - stunning, when engines are so >>>similar in strength. >> >>I'm not sure if I like this very much. The highest thing under that is one time >>000 and 111 each. But statistically, I think that lower chains should be more >>frequent. So when it goes 7 or 8 in a row, I would be very suspicious, and try >>to check out many possible reasons. >>> >>>Jouni > >Why? I'm testing permanently, and rows are perfectly normal. From a mathematical >point of view, the probability for a row of length n is approximately >0.33^(n-1). (with n=1, this makes 100 %) > >So if you play 8 games the probability that all have the same result is about >0.05 %. Within a series of 100 games, this probability raises to a value that is >still unlikely, but no reason for doubts. A row of 4-5 games in a tourney of 100 >games can be expected. That's the reason why a test of 10 games says nothing, >imagine this row was at the beginning of the tourney, which is of course >possible... > >Of course the above formula bases on two approximations: the probabilities for >win, loss and draw are each 33 %, and the two engines are of equal strength. > >Regards, >Stefan it is just like a random number generator. If the generator does not generate a "run" here and there, it looks good, but is really bad. After all, when you flip a coin, you will get several heads or tails in succession at points in the experiment.
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.