Author: Harald Faber
Date: 06:29:29 06/16/99
Go up one level in this thread
On June 16, 1999 at 08:13:13, blass uri wrote: >>>>>I know the Mchess result, and of course I'm happy as larry. What happened then, >>>>>was it a real blunder/bug, or a horizon thing? >>>> >>>>It was certainly a horizon. In that position there was no way out, MCP just >>>>needed some more time/plys to realize that the end is near. >>>>I have seen this several times, programs have problems evaluating passed pawns. >>>>Anyway, MCP is not the only one... :-) >>> >>>I looked at the game and did not see +3 evaluation. >>> >>>what was the losing mistake of MCP in this game? >> >>Who knows? Maybe ...Ne3 wasn't correct? >> >>>can you post the relevant position or the number of move (I am not going to >>>analyze every move from the end of the game for a long time to discover it) >>>Uri >> >>Francesca - MChess >>WCCC99 Paderborn (3), 15.06.1999 >> >>1.e4 e6 2.d4 d5 3.Sc3 Lb4 4.e5 c5 5.a3 Lxc3+ 6.bxc3 Se7 7.Dg4 Dc7 8.Dxg7 Tg8 >>9.Dxh7 cxd4 10.Se2 Sbc6 11.f4 Ld7 12.Dd3 dxc3 13.Le3 d4 14.Sxd4 Sxd4 15.Lxd4 Lc6 >>16.Dxc3 Sd5 17.Dd2 Tg4 18.Le2 Txg2 19.Tg1 Txg1+ 20.Lxg1 0-0-0 21.Td1 Tg8 22.Lf2 >>La4 23.Tb1 Td8 24.c4 Se7 25.Dc3 Sd5 26.Df3 Se7 27.Tb4 Lc6 28.De3 b6 29.Db3 Dd7 >>30.Dc2 Sf5 31.c5 b5 32.Dd1 De8 33.Dc2 Dd7 34.Da2 Lb7 35.Dc2 Dc6 36.Lf1 Df3 37.c6 >>Se3 38.cxb7+ Kb8 39.Lxe3 Dxe3+ 40.De2 Dc1+ 41.Kf2 Td2 42.h3 Txe2+ 43.Lxe2 a6 >>44.Lf3 Dxa3 45.Te4 Dc5+ 46.Kg3 b4 47.Te1 a5 48.Td1 Dc7 49.h4 Ka7 50.h5 a4 51.h6 >>f6 52.Td4 De7 53.Tc4 fxe5 54.Tc8 Dxb7 55.Lxb7 1-0 >> >>Oops, sorry, it is copy&paste via CB7 in German notation. Hope you get it >>anyway. Fail-low was after 49.h4 or 50.h5, don't remember exactly. From ply to >>ply the eval dropped from +3.xx or even +4.xx down to -12 or so... >>Anyway the position was lost. > >I looked at the position after 49.h4 with Junior5.4 and with hiarcs7 and I do >not see a sign for 3.xx advantageor even 1.xx advantage for black. Sometimes evaluation between programs differ. ;-) And it is well known that MCP evaluates positions very optimistic. And last not least MCP underestimated the h- and of course the b-pawn. I thought and hoped that this problem was fixed with V.6 or V.7. Maybe this is too hard to program because it is never easy to say if passed pawn(s) will win or not. Even for humans. >The evaluation after few minutes with Junior is 0.35 pawns for black and it >suggests 49...b3. >Uri I still think the position was lost for black.
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.