Author: Enrique Irazoqui
Date: 15:51:13 08/04/00
Just got it: THE SSDF RATING LIST 2000-08-04 74012 games played by 209 computers Rating + - Games Won Oppo ------ --- --- ----- --- ---- 1 Fritz 6.0 128MB K6-2 450 MHz 2631 28 -27 673 67% 2504 2 Junior 6.0 128MB K6-2 450 MHz 2601 25 -24 864 67% 2478 3 Chess Tiger 12.0 DOS 128MB K6-2 450 MHz 2573 30 -29 569 63% 2481 4 Fritz 5.32 128MB K6-2 450 MHz 2553 31 -30 557 62% 2467 5 Nimzo 7.32 128MB K6-2 450 MHz 2549 29 -28 613 62% 2463 6 Goliath Light 128MB K6-2 450 MHz 2534 48 -48 210 51% 2528 7 Hiarcs 7.32 128MB K6-2 450 MHz 2533 31 -31 519 60% 2460 8 Junior 5.0 128MB K6-2 450 MHz 2526 29 -28 598 58% 2467 9 SOS 128MB K6-2 450 MHz 2516 57 -55 159 58% 2456 10 Nimzo 99 128MB K6-2 450 MHz 2501 29 -29 581 54% 2475 11 Crafty 17.07/CB 128MB K6-2 450 MHz 2499 27 -27 651 51% 2496 12 Fritz 5.32 64MB P200 MMX 2477 20 -20 1208 57% 2429 12 Hiarcs 7.32 64MB P200 MMX 2477 25 -24 815 60% 2404 14 Chessmaster 6000 64MB P200 MMX 2473 61 -53 184 76% 2278 15 MChess Pro 8.0 128MB K6-2 450 MHz 2470 34 -35 418 44% 2511 16 Fritz 5.0 PB29% 67MB P200 MMX 2459 23 -22 1005 66% 2342 17 Hiarcs 7.0 64MB P200 MMX 2458 21 -21 1106 55% 2420 18 Nimzo 99 64MB P200 MMX 2447 23 -23 885 51% 2439 19 Junior 5.0 64MB P200 MMX 2433 22 -22 1010 51% 2427 20 Nimzo 98 58MB P200 MMX 2423 22 -22 1038 58% 2367 21 Rebel 9.0 47MB P200 MMX 2419 24 -23 900 61% 2340 22 Hiarcs 6.0 49MB P200 MMX 2417 24 -24 829 56% 2373 23 Rebel 8.0 51MB P200 MMX 2409 23 -23 887 50% 2408 24 MChess Pro 6.0 41MB P200 MMX 2407 26 -25 749 54% 2378 25 Shredder 2.0 58MB P200 MMX 2396 21 -21 1054 48% 2408 26 MChess Pro 7.1 46MB P200 MMX 2394 22 -22 1042 53% 2371 27 Genius 5.0 DOS 46MB P200 MMX 2393 21 -21 1093 52% 2378 28 MChess Pro 8.0 64MB P200 MMX 2390 27 -27 681 53% 2366 29 Chess Tiger 11.8 Pentium 90 MHz 2387 45 -45 242 52% 2375 30 Gandalf 3.0 64MB P200 MMX 2364 41 -40 307 59% 2296 31 Kallisto II 64MB P200 MMX 2342 35 -35 403 52% 2327 32 Rebel 9.0 Pentium 90 MHz 2334 23 -23 890 47% 2356 33 Hiarcs 6.0 Pentium 90 MHz 2332 18 -18 1437 51% 2328 34 Genius 5.0 DOS Pentium 90 MHz 2329 18 -18 1558 47% 2348 35 MChess Pro 6.0 Pentium 90 MHz 2309 17 -17 1726 45% 2343 36 Nimzo 3.5 Pentium 90 MHz 2293 22 -22 998 46% 2322 37 Chessmaster 5000 Pentium 90 MHz 2287 49 -45 240 67% 2162 37 Junior 4.0 Pentium 90 MHz 2287 22 -22 1035 42% 2341 39 Shredder 1.0 Pentium 90 MHz 2282 59 -58 145 53% 2262 40 R30 v. 2.5 2274 41 -38 343 69% 2135 41 CometA90 64MB P200 MMX 2251 37 -39 358 36% 2351 42 Fritz 4.0 Pentium 90 MHz 2234 40 -39 324 60% 2163 43 WChess 1.06 Pentium 90 MHz 2230 20 -20 1222 39% 2308 44 Meph Genius 68 030 33 MHz 2198 45 -44 248 55% 2161 45 Berlin Pro 68 020 24 MHz 2125 24 -24 850 58% 2071 45 Meph RISC 2 1 MB 2125 62 -66 125 39% 2205 47 Mephisto Montreux ARM 14 MHz 512K 2099 29 -28 689 73% 1930 48 Atlanta SH7000 20 MHz 2093 31 -29 580 67% 1967 49 Sapphire II 2013 35 -33 444 63% 1917 50 Milano Pro SH7000 20 MHz 1974 33 -32 469 61% 1895 6 Goliath Light 128MB K6-2 450 MHz, 2534 Junior6 K6450 12-28 Ch.Ti12 K6450 9-13 Nimz732 K6450 3.5-4.5 Hiar732 K6450 15-14 Nimzo99 K6450 26.5-13.5 Craf17.07 K62 23-17 MCP8 K6-2 450 15.5-11.5 MCP 6 P200MMX 2.5-1.5 9 SOS 128MB K6-2 450 MHz, 2516 Hiar732 K6450 6-10 Nimzo99 K6450 10-4 Fritz532 P200 10.5-7.5 Hiarcs7 P200X 22.5-15.5 Junior5 P200X 18.5-13.5 190 P200MMX 23.5-13.5 MCP 6 P200MMX 2-2 15 MChess Pro 8.0 128MB K6-2 450 MHz, 2470 Fritz6 K6-450 9-35 Junior6 K6450 15-25 204 K6-450 3.5-4.5 Nimz732 K6450 19-25 Goliath K6450 11.5-15.5 Hiar732 K6450 1-4 Junior5 K6450 15.5-24.5 Nimzo99 K6450 20.5-21.5 Craf17.07 K62 7-15 Fritz532 P200 17.5-28.5 Hiarcs7 P200X 21-19 Junior5 P200X 16.5-3.5 193 P200MMX 27.5-12.5 The SSDF rating list provides information about the relative strength of chess programs, when tested in the way SSDF does, but does not necessarily say which ELO-rating a certain program would achieve after having played hundreds of tournament games against human players. How good or bad the individual correlation between SSDF- and ELO-ratings is, will most likely never be established. So many games against humans will never be played. Apart from establishing relative ratings, we have had the ambition that the general level of the list would be fairly realistic, compared to human ratings. From our start in 1984 we have used tournament games against Swedish chess players to calibrate the list. At some points we have discarded older games, believing that human chess players with time have become better to exploit the weaknesses of chess programs. Until the latest rating list the level of the list has been unchanged from summer 1991, and was based on 337 tournament games against Swedish players between 1987 and 1991. Regrettably it has not been possible for us to play any more games for many years now. For some time we had the general impression that the level of the list was rather OK. But during the latest years it has become more and more obvious that the best programs on the latest hardware don't get as high ELO-ratings as our list could be interpreted to predict. If this is due to differences between Swedish- and ELO- ratings, to the "human learning effect", to some kind of "spreading effect" in a computer-computer list or a com- bination of these and perhaps other factors, we don't know. It is difficult to find a perfect solution, but we have chosen to correlate the level of the list to the results of tournament games between computers and ELO-rated humans, played during the latest years. For us it has been very convenient to use Chris Carsons compilation of such games. Calculations based on these games indicate that the level of the list is about 100 points too high. So from now on we have lowered the list with 100 points! Our hope is that the SSDF-ratings of the top entrants as a group now are better correlated with ELO-ratings. If the rating-inflation to a large part is due to a "spreading-effect", there is now a certain possibility that the older and weaker entrants of the list would play better against humans than their SSDF-ratings could indi- cate. But having to choose, we prefer to secure that the top programs have as correct ratings as possible. It is interesting to see how well chess programs play against each other, but it's even more fascinating to see what they can achieve against humans! I hope that more games against strong humans will be played in the future, and that Chris Carson will continue to collect these games, so that the level of the SSDF list can be more securely established. Compared to the latest rating list in early April we now have 1953 more tournament games and three new entrants. Marty Hirschs MChess Pro 8.0 has been replayed on K6-2 450 MHz. After 418 games it has 2470, which is 80 points more than on Pentium MMX 200 MHz. The difference between these two hardwares has in average been 79 points, so the result is as expected. Completely new on the list is Michael Borgstädts Goliath Light K6-2 450 MHz. It is played under the Fritz surface using the opening book general.ctg from Fritz 6. It has got a rating of 2534, which gives it a sixth place on the list! SOS K6-2 450 MHz is written by Rudolf Huber and is also played with the opening book of Fritz 6. After 159 games it has 2516 and a ninth place! Crafty 17.07 K6-2 has lost 24 points compared to the latest list and Nimzo 99 K6-2 has 22 points less. Fritz 6 K6-2 has gone up 10 points and Junior 6 K6-2 has increased 12 points. Next official list will be made in September or October. Thoralf Karlsson
This page took 0.02 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.