Author: James Swafford
Date: 09:48:57 05/17/99
Go up one level in this thread
On May 17, 1999 at 04:31:41, Will Singleton wrote: >Under 2600 Club >May 17 1999 > >Program ICC Blitz Change Games Platform Author > >LambChop 2586 +9 1 P/133 Pete McKenzie >TDChess 2578 P/400 Jon Baxter >Rajah 2578 +123 30 P/166 V. Manohararajah >PostModernist 2546 -87 10 K6/300 Andrew Williams >Amateur 2529 -36 9 Mac/300 Will Singleton >EXchess 2475 Cel/400 Dan Homan >Galahad 2450 +92 26 K6/400 James Swafford >Hossa 2426 -12 286 P/350 Steffen Jakob >Grok 2426 -2 68 P/300 Peter Kappler >Arasan 2374 -166 70 K6/233 Jon Dart >BugChess 2326 P/450 Erik van het Hof > > > ICC account names that differ from the program name: > ---------------------------------------------------------------------------- > ArasanX (Arasan), jlong (Galahad). > ---------------------------------------------------------------------------- > > >Notable > >LambChop tops the list this week, at 2586. Interesting to note that LambChop >runs on a P133, while Chopster (P3/500) and Beadle (P2/266), running the same >program, are rated 2590 and 2585, respectively. This probably shows the >importance of opponent selection, since their formulas differ. And, too, >blitz ratings can swing a lot. But the fact that the three accounts are >within 5 points of each other is remarkable. > >Hossa was again the most active program, while Rajah made the largest gain >with 123 points. Also gaining quite a lot was Galahad, and based on some >of its games this week, I expect similar gains in the future. > >One new high wes recorded, by Galahad (2462). I attribute that to two things. One was a bug fix, which was related to my thread based i/o routine. It would strike every twenty games or so. I also modified my extensions w/ the positions Dann posted. I use fractional extensions, where one ply=20. I had all the extensions except one-reply-to-check set to 15. The one-rep was a full 20. I set then _all_ to 20, and my program plays much better. I'm sure I should study this more closely, but right now I'm wondering just how much fractional extensions really help. Doesn't appear to be much. In each case, the maximum extension is one ply, except in cases of one-reply. I still have a bug to chase, though. The prog was playing ComputerTal2, and was +10 pawns or so in the endgame, but lost on time after it hung up. Haven't played a game since. :-( -- James > > >Changes > >Rookie goes inactive. > > >-------------------------- >u2600 Position of the Week > >This is from the game Amateur-Phalanx21 (pikozrout), played this week on >fics. The correct move is an intuitive choice for humans, and likely for >some programs. Even if yours likes it initially, see how long it takes for >the score to jump up. (Ne4 was played in the actual game.) > >r3k2r/ppp2ppp/3bb3/q3p3/2P5/P1NP1P2/1P1B2PP/R1Q1K2R w KQkq - > >Amateur 0.9, g3/300 -- > 2 -80 0.02 1883 > 3 -73 0.05 2630 Kg1 Bc5 Be3 Kc8 > 3 -63 0.07 3217 Ne4 Qb6 Nxd6 Qxd6 > 4 -69 0.14 7969 Ne4 Qa6 Kg1 Kc8 Nxd6 Rxd6 > 5 -54 0.25 14896 Ne4 Qa6 Kg1 Kc8 Re1 > 6 -55 0.50 31673 Ne4 Qa6 Kg1 Kc8 Re1 Be7 > 7 -46 2.54 107992 Ne4 Qa6 Bc3 Kc8 Qe3 f6 Kg1 > 8 -41 15.07 642093 Ne4 Qa6 Qc3 Kc8 Ng5 Bc5 Qxe5 Bd4 > 9 -41 37.43 1662808 Ne4 Qa6 Qc3 Kc8 Ng5 Bc5 Qxe5 Bd4 Qa5 > 9 90 56.42 2593903 b4 Qa6 c5 Bxc5 Pxc5 Kc8 Bg5 Rxd3
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.