Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Hydra 3000?

Author: Günther Simon

Date: 08:09:06 06/24/05

Go up one level in this thread


On June 24, 2005 at 10:57:55, James T. Walker wrote:

>On June 24, 2005 at 10:42:35, James T. Walker wrote:
>
>>On June 24, 2005 at 00:56:57, Uri Blass wrote:
>>
>>>On June 24, 2005 at 00:33:45, Amir wrote:
>>>
>>>>On June 24, 2005 at 00:25:03, Terry McCracken wrote:
>>>>
>>>>>On June 24, 2005 at 00:06:37, Amir wrote:
>>>>>
>>>>>>On June 23, 2005 at 23:45:59, Terry McCracken wrote:
>>>>>>
>>>>>>>On June 23, 2005 at 19:56:17, Amir wrote:
>>>>>>>
>>>>>>>>With what score Hydra would have to be beat Adams to have 3000 performance
>>>>>>>>rating? (if it's even possible at all). Hydra also played a match against Evgeny
>>>>>>>>Vladimirov. What was Hydra's performance rating in that match?
>>>>>>>>
>>>>>>>>Can we calculate Hydra's FIDE rating from these two matches against humans?
>>>>>>>
>>>>>>>No you can't calculate Hydras' rating FIDE or otherwise on so few games and get
>>>>>>>anything meaningful.
>>>>>>
>>>>>>You can calculate performance rating of the players in the match or tournament.
>>>>>>You can also probably calculate provisional FIDE rating. After this match, Hydra
>>>>>>will have at least 13 games against FIDE rated players.
>>>>>
>>>>>Then why ask the question...the part in brackets? Also, you can't get a
>>>>>meaningful performance rating or FIDE or whatever and now I'm wasting my time on
>>>>>you repeating myself!
>>>>>
>>>>>Would you like me to answer it again?? Are you trying to be annoying??
>>>>>
>>>>>Or are just trolling like I stated before?
>>>>
>>>>"Performance" ratings are not "meaningful" but you can still calculate them for
>>>>a match or tournament. My question was very clear: with what score Hydra will
>>>>have to beat Adams to get close to 3000 performance rating in *this* match. The
>>>>question was not about "meaning" but of math. If anyone knows how to calculate
>>>>performance rating, he can  answer my question, regardless how meaningless the
>>>>answer is.
>>>
>>>
>>>Hydra needs to score exactly 4.92 points out of 6 games to get performance of
>>>3000.
>>>
>>>I used the following site to calculate performance
>>>
>>>http://www.uschess.org/ratings/calculator.html
>>>
>>>I printed 6 times 2737(Adams rating) and changed the total score until I got
>>>exactly 3000 after clicking performance.
>>>
>>>Uri
>>
>>Hello Uri,
>>Since Hydra cannot score 4.92 points that figure is not important.  If you plug
>>4.5 into the program you get a performance rating of 2928.  If you use 5 points
>>out of 6 you get a performance rating of 3017.  There is nothing wrong with
>>either of these calculations that I can see.
>>Jim
>
>Of course this forumla is for USCF ratings.  The FIDE uses something entirely
>different for "performance ratings".
>Jim

No, _all_ chess organisations use a quite similar method for calculating
the performance.
Actually the difference will not be more than a very few Elo.

Almost all rely on a similar precalculated table. This one is for
German DWZ, but it is identically to FIDEs'.

DWZ-Diff.P(D) betterP(D)  worse    DWZ-Diff.P(D)   betterP(D)  worse
  0-3           0,50       0,50        189-197       0,75       0,25
  4-10          0,51       0,49        198-206       0,76       0,24
 11-17          0,52       0,48        207-215       0,77       0,23
 18-25          0,53       0,47        216-225       0,78       0,22
 26-32          0,54       0,46        226-235       0,79       0,21
 33-39          0,55       0,45        236-245       0,80       0,20
 40-46          0,56       0,44        246-256       0,81       0,19
 47-53          0,57       0,43        257-267       0,82       0,18
 54-61          0,58       0,42        268-278       0,83       0,17
 62-68          0,59       0,41        279-290       0,84       0,16
 69-76          0,60       0,40        291-302       0,85       0,15
 77-83          0,61       0,39        303-315       0,86       0,14
 84-91          0,62       0,38        316-328       0,87       0,13
 92-98          0,63       0,37        329-344       0,88       0,12
 99-106         0,64       0,36        345-357       0,89       0,11
107-113         0,65       0,35        358-374       0,90       0,10
114-121         0,66       0,34        375-391       0,91       0,09
122-129         0,67       0,33        392-411       0,92       0,08
130-137         0,68       0,32        412-432       0,93       0,07
138-145         0,69       0,31        433-456       0,94       0,06
146-153         0,70       0,30        457-484       0,95       0,05
154-162         0,71       0,29        485-517       0,96       0,04
163-170         0,72       0,28        518-559       0,97       0,03
171-179         0,73       0,27        560-619       0,98       0,02
180-188         0,74       0,26        620-735       0,99       0,01
                                        >735         1,00       0,00

Guenther



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.