Author: Robin Smith
Date: 21:03:30 07/01/02
Go up one level in this thread
On July 01, 2002 at 23:43:54, Dann Corbit wrote: >On July 01, 2002 at 23:37:21, Robin Smith wrote: > >>On July 01, 2002 at 15:21:09, Dann Corbit wrote: >> >>>On July 01, 2002 at 14:55:41, Albert Silver wrote: >>> >>>>On July 01, 2002 at 14:49:47, pavel wrote: >>>> >>>>>On July 01, 2002 at 14:47:40, ERIQ wrote: >>>>> >>>>>>I'm not sure about you guys but I'm in shock !! >>>>> >>>>>Because how the PGN output looks? >>>>> >>>>>;) >>>>>cheers, >>>>>pavs >>>> >>>>Either that or because Hiarcs 8 on an Athlon 1200 lost to Junior 7 on a K6/450 >>>>in a previous SSDF match. >>> >>>In a span of 4 games, we can expect anything to happen. In the Ridderkerk >>>tournament, Quark 1.76 blasted Crafty 18.15 4-0. I ran the same programs over >>>the weekend at G/90 and got: >>> >>> Program Elo + - Games Score Av.Op. Draws >>>1 Crafty-18.15 : 2522 143 105 24 56.2 % 2478 37.5 % >>>2 Quark v1.76beta : 2478 105 143 24 43.8 % 2522 37.5 % >>> >>>I don't know why anyone would be in shock about any 4-0 result unless there were >>>at least a 500 ELO difference. >> >>Even for a 500 rating point difference, the weaker side should get 24% of the >>points. If we simplify the math by ignoring draws, four games in a row would be >>.24^4=.0033 (or .33% or 3.3 times out of 1000 four game matches). So I am >>still not sure 4 wins in a row for a 500 rating point difference should cause >>shock. > >I think it is 5% for 500 points. 24% is about 200 points difference. > >So that would give 0.0532402^4 = 0.0000080344973720078203276816 >Which is about 1 in 124463 trials. It could happen, but rather shocking, I >think. At least surprising. Yes, you are right, I somehow grabbed the expected result for a 200 point difference. 1 in 124463 trials is definately shocking, assuming you a have a much smaller number of trials. At least it would shock me.
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.