Author: Albert Silver
Date: 10:11:21 08/16/04
I've been playing around with the settings of ProDeo, as many here have, testing
it specifically on the Nunn sets, so as to avoid noise created by opening books.
The opponent of choice has been Junior 8, since it has consistently given ProDeo
a hard time. This new group of settings I've tried seems slightly promising so
far, but it's still too early to say, and even if it does, it bears testing
against other engines. Here are a couple of positions from one of the games:
[D]r1r3k1/3b1pp1/pp2p2p/n2pP3/P2P3Q/2qB4/2P2PPP/RRB3K1 w - - 0 24
Interestingly enough, Junior 8 never saw this one coming, and after testing it a
bit, I noted that some engines see it quite easily and others not. The move is
24.Bxh6! giving white a very strong attack. Note that the engine has to show it
really holds this. I say this because when testing Hiarcs 9, it held Bxh6 but
then switched to 24.Qg3 for a ply at ply depth 10. I had let it sit on it
because the eval kept dropping:
24.Qe7 Bxa4
= (0.21) Depth: 5/19 00:00:00 26kN
24.Bb2 Qd2 25.Ba3 Nc4 26.Bb4
² (0.37) Depth: 5/19 00:00:00 32kN
24.Bxh6 gxh6
± (0.77) Depth: 5/19 00:00:00 46kN
24.Bxh6 gxh6 25.Qxh6 Be8 26.Bh7+
± (1.02) Depth: 5/21 00:00:00 95kN
24.Bxh6 gxh6 25.Qxh6 Be8 26.Rd1 Ra7 27.h4 Qb2 28.Bh7+ Kh8
± (1.23) Depth: 6/21 00:00:00 181kN
24.Bxh6 gxh6 25.Qxh6 Be8 26.Rd1
± (1.23) Depth: 7/23 00:00:02 566kN
24.Bxh6 gxh6
± (0.98) Depth: 8/25 00:00:08 2327kN
24.Bxh6 gxh6 25.Qxh6 Be8 26.Rd1 Nc4 27.Bh7+ Kh8 28.Bf5+ Kg8 29.Bd3 Rab8 30.Bh7+
Kh8
± (0.92) Depth: 8/27 00:00:11 3029kN
24.Bxh6 gxh6
² (0.67) Depth: 9/27 00:01:00 15985kN
24.Bxh6 gxh6 25.Qxh6 Be8 26.Bh7+ Kh8 27.Re1 Nc4 28.Bd3+ Kg8
² (0.66) Depth: 9/29 00:01:06 17405kN
24.Bxh6 gxh6 25.Qxh6 Be8 26.Bh7+ Kh8 27.Rd1 Nc6 28.Bd3+ Kg8 29.Bh7+
² (0.63) Depth: 10/29 00:04:48 74586kN
24.Qg3 Qe1+ 25.Bf1 Qc3 26.Qxc3 Rxc3 27.Rxb6 Rxc2
² (0.69) Depth: 10/30 00:10:05 135105kN
24.Qg3 Qe1+ 25.Bf1 Qc3 26.Qxc3 Rxc3 27.Rxb6 Rxc2
² (0.64) Depth: 11/30 00:14:33 187261kN
24.Bxh6 gxh6
± (0.94) Depth: 11/30 00:15:36 202735kN
A couple of moves later, after 24.Bxh6 gxh6 25.Qxh6 Be8 (only move), another
good test position comes up regarding the follow-up to the attack:
[D]r1r1b1k1/5p2/pp2p2Q/n2pP3/P2P4/2qB4/2P2PPP/RR4K1 w - - 0 26
The move here is 26.Rf1! There is more than one way to conduct the attack here,
and I'll admit I haven't examined them all exhaustively, but I believe bringing
the rook to f1, covering the back rank and allowing the a1 rook to join the
fray, is the most logical.
[D]r1r1b1k1/5p2/pp2p2Q/3pP3/P1nP4/2qB4/2P2PPP/R4RK1 w - - 0 27
After 26.Rf1 Nc4 seems best, but here is a really bizarre thing: When I look at
this, the most logical follow-up to me seems to be 27.Rae1, threatening 28.Bxc4
and 29.Re3, yet none of the engines really likes it. At least not at any
playable time-control. However, in the end, DF8 confirmed 27.Rae1 as best:
27.Rad1!
+- (1.93) Depth: 10/29 00:00:04 3453kN
27.Rad1 Nxe5 28.Qh7+ Kf8 29.Qh8+ Ke7 30.dxe5 Kd7 31.Qh4 Qxe5
+- (2.21) Depth: 10/29 00:00:04 4205kN
27.Rad1 Nxe5 28.Qh7+ Kf8 29.Qh8+ Ke7 30.dxe5 Kd7 31.a5 Qxa5 32.Ra1 Qc3 33.Bxa6
+- (2.40) Depth: 11/31 00:00:07 6400kN
27.Rad1 Nxe5 28.Qh7+ Kf8 29.Qh8+ Ke7 30.dxe5 Kd7 31.a5 b5 32.f4 Kc7 33.Qf8 Kb7
34.Qe7+
+- (2.42) Depth: 12/35 00:00:11 11229kN
27.Rad1 Nxe5 28.Qh7+ Kf8 29.Qh8+ Ke7 30.dxe5 Kd7 31.a5 b5 32.f4 Kc7 33.Qf8 Kb7
34.Qe7+ Rc7
+- (2.42) Depth: 13/36 00:00:23 23252kN
27.Rad1 Nxe5 28.Qh7+ Kf8 29.Qh8+ Ke7 30.dxe5 Kd7 31.Qf6 Kc7 32.Rb1
+- (2.50) Depth: 14/39 00:00:56 58954kN
27.Rad1 Nxe5 28.Qh7+ Kf8 29.Qh8+ Ke7 30.dxe5 Kd8 31.Rb1 Kc7 32.Rb3 Qd4 33.Qf6
Qxa4 34.Rfb1 b5
+- (2.55) Depth: 15/41 00:03:29 219995kN
27.Rad1 Nxe5 28.Qh7+ Kf8 29.Qh8+ Ke7 30.dxe5 Kd8 31.Rb1 Kc7 32.Rb3 Qd4 33.Qf6
Qxa4
+- (2.55) Depth: 16/44 00:08:15 528574kN
27.Rae1!
+- (2.56) Depth: 16/46 00:26:59 1705842kN
27.Rae1!
+- (2.71) Depth: 16/46 00:34:06 2109132kN
27.Rae1 Ra7 28.Bxc4 Qxc2 29.Re3 Qg6 30.Rg3 Rxc4 31.Rxg6+ fxg6 32.Qg5 Rd7
+- (3.00) Depth: 16/54 00:54:34 3322121kN
Albert
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.