Author: blass uri
Date: 09:38:33 07/28/00
Go up one level in this thread
On July 28, 2000 at 11:26:36, Harald Faber wrote: >On July 28, 2000 at 11:11:33, blass uri wrote: > >>On July 28, 2000 at 08:07:32, Harald Faber wrote: >> >>>On July 28, 2000 at 06:49:44, Terry Ripple wrote: >>> >>>>On July 28, 2000 at 05:58:53, Terry Ripple wrote: >>>> >>>>>On July 28, 2000 at 03:06:29, Harald Faber wrote: >>>>> >>>>>>On July 28, 2000 at 01:05:53, Terry Ripple wrote: >>>>>> >>>>>>>Used an AMD K6-2, 266Mhz, 64Ram, Ponder off, 16Mb Hash per engine. >>>>>>> >>>>>>>If anyone cares to see some or all of the games, i will be glad to post them. >>>>>>> This match shows how close the strengths are between these two fine engines! >>>>>>> >>>>>>>Best regards, >>>>>>>Terry >>>>>>> >>>>>>>Blitz:5' 2000 >>>>>>> >>>>>>> >>>>>>>1 Fritz 6 158.0/306 >>>>>>>2 Hiarcs 7.32 148.0/306 >>>>>> >>>>>>Again for interest: What is the score after 10, 20, 50, 100 and 200 games from >>>>>>your tournament? >>>>> >>>>>Hi Harold, >>>>> It's interesting to see how the scores change, and proves that you need many >>>>>games to try and determine the strengths of different engines! Here's the info: >>>>> >>>>>Regards,Terry >>>>> >>>>> >>>>> Fritz 6 H7.32 >>>>> Scores at 10 games : 2.5 7.5 >>>>> " 20 " : 10.0 10.0 >>>>> " 50 " : 26.5 23.5 >>>>> " 100 " : 52.0 48.0 >>>>> " 150 " : 79.5 70.5 >>>>> " 200 " : 105.5 94.5 >>>>> " 250 " : 130.5 119.5 >>>>> " 306 " : 158.0 148.0 >>>> >>>> I would like to comment from the scores that it proves that at a 20 game match >>>>it doesn't mean anything! Between 50 and 100 games the score for Fritz moves up >>>>only slightly from a +3 to +4 points in its favor. But at 150 games the score >>>>climbs considerably to a +9 point advantage over Hiarcs7.32. Then it starts >>>>tapering off between the 200 and 250 games with a point advantage of +11 for >>>>Fritz 6 and then drops to a +10 point advantage at 306 games. So, for two - >>>>engines that are close to the same strength it shows that you need 50 to 100 >>>>games, and if the engines are "extremly" close in strength it would be wise to >>>>play anywhere from 150 to 200 games! This of course is based on playing 5 min. >>>>per side for the whole game, as slower time controls might change this factor! >>>> This of course is only from the view of a non-expert's opinion, and by - >>>>presenting this information that was requested by Harold i hope to get some >>>>interesting points of view from our members. >>>> >>>>Best regards, >>>>Terry >>> >>> >>>Hi Terry, >>>Thanks a lot for the info. >>>It reflects all I have expected and what I have got in my own testing games. >>>I hope this info opens many eyes who conclude MUCH too early. You may get an >>>idea of weaknesses or strengths of new programs or versions after few games but >>>this is only subjective. >>>And this result proves me right with my estimation that the top 10 programs play >>>at one level. (I know that you played 5min/game but the result won't be much >>>different with longer time controls) >>>BTW the +10 difference in favour of Fritz after 306 games just means 51.6% >>>winning percentage, that is why I say one level. >> >>I disagree. > > >I knew that. :-) >You almost always disagree with anything. :-) > > >>Hiarcs is better in blitz and I expect result of 60% for Fritz6 in tournament >>time control based on the ssdf rating. > > >"Expect" is very vague and pure speculation. >And BTW a possible 60%-score would not say that Fritz is stronger, this would >have to be proven against other opponents too... > > >>I discovered that hiarcs suffers from a learning bug at tournament time >>control(there are cases when it needs more time to see something or cannot see >>something because it does not get the right information from the hash tables) >>and I guess that it is the main reason that hiarcs is worse at tournament time >>control. >>Uri > > >I disagree. :-) >Do you have enough games to prove this? :-) one game is enough to prove a bug and I proved by one game of Hiarcs7.32 against Crafty17.11(3 hours per 40 moves) that hiarcs has a learning bug(In one case hiarcs needed a long time to solve the fail low because it failed low at depth 10 and I found after the game that if I give hiarcs only the position without playing it fails low at depth 9 and can find a better move faster and if I play at slower time control than 3 hours/40 moves it fails low at depth 11 and has not enough time to search to find a better move) There was also another move in the same game that I checked and found that I cannot reproduce it. I suspect that hiarcs can play better in long time control if it clears the hash tables after every move but unfortunately I do not know about a way to do it(learning parameter off does not help). Uri
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.