Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: DT and DB (prototype) vs Computers

Author: Robert Hyatt

Date: 09:10:37 01/24/00

Go up one level in this thread


On January 24, 2000 at 09:21:45, Chris Carson wrote:

>On January 23, 2000 at 22:28:55, Robert Hyatt wrote:
>
><Big Snip>
>
>>>But, my guess is we will hear some blah, blah, blah, wait for
>>>some mythical chip, don't believe the data, watch me pull a
>>>rabbit out out of my hat - statements.  :)
>>>
>>
>>Here's a rabbit or three:  Which commercial program has produced a USCF rating
>>of 2550+?  Which commercial program has produced a TPR of 2650 over 25
>>consecutive games?
>>
>>When you can point to one, I'll listen carefully.  I am not interested in TPRs
>>of 2600 for 5 games, followed by a TPR of 2300 for the next 5, etc.
>>
>>When you can point to a program that has come close to that, let me know and
>>provide games as Monty did in his book.  And remember that this was all on
>>1986 hardware.  DB/DB2 didn't have anything to do with those results, although
>>they produced some pretty good TPRs against the best player in the world.
>>
>
>Date	Player	        FIDE	Program	PS	TPR	25 Games
>Feb-97	Ernst	        2450	Rebel 8	0	2325	2546
>Mar-97	Corvi	        2306	Rebel 8	0.5	2306	2558	*
>Apr-97	Van Geet	2320	Rebel 9	1	2651	2567	*
>Apr-97	Loeffler	2410	Rebel 9	1	2651	2562	*
>Apr-97	Polgar	        2470	Rebel 9	0.5	2651	2553	*
>Apr-97	Timoshchenko	2515	Rebel 9	1	2651	2549
>Apr-97	Kosashvili	2560	Rebel 9	0	2651	2544
>Apr-97	Kuijf	        2475	Rebel 9	1	2651	2535
>Jun-97	Akesson	        2485	Rebel 8	1	2325	2531
>Jun-97	Akesson	        2485	Rebel 8	0	2325	2539
>Nov-97	Santos	        2190	Rebel 9	1	2683	2543
>Nov-97	Pena	        2330	Rebel 9	1	2683	2533
>Nov-97	Dominguez	2340	Rebel 9	0.5	2683	2524
>Nov-97	Mateo	        2408	Rebel 9	1	2683	2518
>Nov-97	Hernandez	2320	Rebel 9	1	2683	2512
>Dec-98	Dominguez	2340	Rebel 10 0.5	2533	2508
>Dec-98	Pinal	        2330	Rebel 10 1	2533	2511
>Dec-98	Alvarado	2180	Rebel 10 1	2533	2513
>Dec-98	Fernendez	2180	Rebel 10 1	2533	2515
>Dec-98	Mateo	        2415	Rebel 10 1	2533
>Dec-98	Montes	        2130	Rebel 10 0.5	2533
>Dec-98	Jaquez	        2220	Rebel 10 1	2533
>May-99	Rohde	        2585	Reb-Cen	0	2185
>Jul-99	Sorin	        2524	Reb-Cen	0.5	2537
>Jul-99	Anand	        2795	Reb-Cen	0.5	2605
>Jul-99	Anand	        2795	Reb-Cen	0	2605
>Aug-99	Rohde	        2585	Reb-Cen	0.5	2537
>Sep-99	Hoffman	        2501	Reb-Cen	0	2537
>Sep-99	Zapolskis	2399	Reb-Cen	1	2427
>Sep-99	Ruzele	        2515	Reb-Cen	0.5	2537
>Sep-99	Grabliauskas	2466	Reb-Cen	0.5	2533
>Sep-99	Krivonosov	2432	Reb-Cen	0.5	2432
>Oct-99	Scherbakov	2541	Reb-Cen	1	2537
>Oct-99	Zagorskis	2486	Reb-Cen	1	2537
>Oct-99	Krivonosov	2432	Reb-Cen	0.5	2432
>Oct-99	Sakalauskas	2450	Reb-Cen	0	2427
>Oct-99	Grabliauskas	2466	Reb-Cen	0	2442
>Dec-99	Baburin	        2593	Reb-Cen	0.5	2537
>Jan-00	Scherbakov	2552	Reb-Cen	0	2537
>Jan-00	Russek	        2398	Reb-Cen	1	2591
>Jan-00	Russek	        2398	Reb-Cen	0.5	2591
>Jan-00	Russek	        2398	Reb-Cen	0.5	2591
>Jan-00	Russek	        2398	Reb-Cen	1	2591
>Jan-00	Nur	        2300	Reb-Cen	1	2442
>			Average		        2534
>                        (+20, =15, -9)  44 FIDE rated games!
>                        Better than USCF and not G/30 or
>                        G/60.


I don't understand your last column.  What does it mean.  IE you have
data for that after one game...  which would be impossible if it means
TPR over last 25 games...  IE rebel lost the first game it played in your
list, yet the 25 game TPR is 2546... so I assume I am not understanding
your chart.

Maybe that number is for that game plus the next 24?

Also what are the time controls.  I haven't seen this many games played
at 40/2hr...





>
>Notes:  DT TPR and 25 game rating was USCF.  Rebel rating
>        is based on FIDE rated games.  FIDE=USCF-100.


This is incorrect.  I gave a URL to find out how to adjust FIDE to USCF.

>
>        * games would be over USCF 2650.  Four times not once.  :)
>
>        Thus Rebel outperformed DT.  I have not included any
>        Rebel G/30 or G/60 data.  I will now go back and add
>        that data, the TPR and 25 game average will increase.
>
>Also, I include the actual ratings for opponents,
>M. Newborn did not, only TPR's.
>
>Best Regards,
>Chris Carson



This page took 0.01 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.