Author: Albert Silver
Date: 13:12:35 09/14/04
Go up one level in this thread
On September 14, 2004 at 15:30:21, Eduard Nemeth wrote:
>Both settings You can find now on my Homepage, but Pro Deo 1.0vff is my style
>That I like (vff2 is only Experimental at time).
>
>I played 7 games blitz against Fritz 8.... Ok, blitz only but see the game 7.
>
>First with the Experimental Version ff2 (4 games).
Glad you're having so much fun with the style. Same here. I should add that
whenever I test a setting or style, I compare the results with those of the
default setting.
>Diagram 1:
>
>Game 7 (ok. only Blitz but see it!):
>Pro Deo 1.0vff - Fritz 8, P600, Blitz:10' 2004
>
>[D]r1b2rk1/6pp/pq1p1p2/1p2pP1Q/3bP3/3P3P/PP1BN1P1/4RR1K b - - 0 1
>
>Pro Deo 1.0vff gibt den Bauer auf b2 auf!? White gives the pawn on b2!?
It's interesting of course, but is it good? My current experimental settings
likes this move also for a number of plies but stops liking it at 13 plies, and
then prefers Bc1 again.
>Diagram 2:
>
>Pro Deo 1.0vff - Fritz 8, P600, Blitz:10' 2004
>
>[D]r4rk1/1b4pp/pq3p2/1pbppP1Q/4P3/1R1P2RP/P2BN1P1/7K w - - 0 1
>
>d4! An fantastic and dynamic move by Pro Deo 1.0vff on blitz time only on P600.
>(I like such games)
Note that the default style also plays this in 11 seconds.
Analysis by Pro Deo 1.0:
26.Nc3 Bf2 27.Rg4 dxe4 28.dxe4 Qd4 29.Bh6 Rf7 30.Qxf7+ Kxf7 31.Rxg7+ Ke8
µ (-0.89) Depth: 9 00:00:10 2310kN
26.Rg4 Rf7 27.Rh4 h6 28.Rg4 Kf8 29.Kh2 dxe4 30.dxe4 Rd8
µ (-0.81) Depth: 9 00:00:11 2639kN
26.d4
µ (-0.81) Depth: 9 00:00:11 2928kN
26.d4 exd4 27.Bh6 Rf7 28.Bxg7 Rxg7 29.Rxg7+ Kxg7 30.Rg3+ Kf8 31.Qxh7 Ke8 32.Rg8+
Bf8
= (0.19) Depth: 9 00:00:12 3619kN
26.d4 Be7 27.Qh6 Rf7 28.dxe5 dxe4 29.e6 Kh8 30.Qh5
² (0.52) Depth: 10 00:00:14 5223kN
26.d4 Be7 27.Qh6 Rf7 28.dxe5 Bc5 29.e6 Re7 30.Qxf6 Rf8 31.Qh4 dxe4
± (1.00) Depth: 11 00:00:17 8579kN
>Pro Deo 1.0vff - Fritz 8, P600, Blitz:10' 2004
>
>[D]r4rk1/1b4pp/pq3p1B/1pbp1P1Q/3pP3/1R4RP/P3N1P1/7K b - - 0 1
>
>Pro Deo 1.0vff says here +3! and game was over soon!
Analysis by Pro Deo 1.0:
27.Bh6 Rf7 28.Bxg7 Rxg7 29.Rxg7+ Kxg7 30.Rg3+ Kf8 31.Qxh7 Ke8 32.Rg8+ Bf8 33.Qg7
= (0.05) Depth: 7 00:00:01 239kN
27.Bh6 Rf7 28.Bxg7 Rxg7 29.Rxg7+ Kxg7 30.Rg3+ Kf8 31.Qxh7 Ke8 32.Rg7 Bf8 33.Rxb7
+- (2.66) Depth: 8 00:00:01 950kN
27.Bh6 Qc7 28.Bxg7 Qxg7 29.Rxg7+ Kxg7 30.Nf4 Rfc8 31.Rg3+ Kh8 32.Qf7 Rg8 33.Qxb7
Rxg3
+- (2.79) Depth: 9 00:00:03 1597kN
27.Bh6 Rf7 28.Bxg7 Qc7 29.Bxf6+ Kf8 30.e5 Rc8 31.a3 Be7
+- (2.83) Depth: 10 00:00:05 3337kN
27.Bh6 Rf7 28.Bxg7 Rxg7 29.Rxg7+ Kxg7 30.Rg3+ Kf8 31.Qxh7 Ke8 32.Rg7 Bf8 33.Rxb7
Qd6 34.Nxd4 b4
+- (3.05) Depth: 11 00:00:09 8136kN
Please don't misunderstand me. The game that was played could only have been
played with your style, so nice job, but when a move or evaluation is proposed
as special to the style, the first thing I do is compare it with the default
settings. I am just as strict when evaluating my own you can be sure.
BTW, I did quite a bit of experimenting with that MISC_17 setting and although
the ply depth looks impressive, I think it misses more moves than it finds, and
takes longer to find more moves than it is faster in others. Those were my first
impressions in any case after my testing. I get better results using Ed's
untested:
[Pruning = MISC_03] * move ordering speedup
[Pruning = MISC_21] * Lazy Eval (tuning)
[Checks Depth = 1] * limit QS checks to 1 ply
Albert
>
>GREAT!
>
>I like this style! :-)
>
>Sorry,
>Eduard
>
>My World:
>
>http://www.beepworld.de/members37/computerschach/
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.