Author: Joseph Ciarrochi
Date: 12:17:53 12/29/05
Go up one level in this thread
that is an amazing difference uri. the SP setting buys four ply!!! But perhaps
they don't use SP because they lose strength in other areas? Would you suggest
that maybe in the endgame, they switch to SP setting
in any case, we probably would not have made this interesing discovery without
Sune's test suite :)
On December 29, 2005 at 12:08:50, Uri Blass wrote:
>On December 29, 2005 at 12:01:27, Uri Blass wrote:
>
>>On December 29, 2005 at 11:48:48, Sune Larsson wrote:
>>
>>>Thanks to Joseph who has played 80 games Rybka-F9 and Rybka-Fruit - endgame
>>>test. A lot of interesting information has been collected. One could expect good
>>>and bad endings by these programs - but maybe not this...
>>>
>>>[Event "Rybka-Fritz9 NunnE 40 4'"]
>>>[Site "Wollongong"]
>>>[Date "2004.12.30"]
>>>[Round "28.1"]
>>>[White "Fritz 9"]
>>>[Black "Rybka 1.0 Beta 32-bit"]
>>>[Result "1-0"]
>>>[Annotator "1.06;0.29"]
>>>[SetUp "1"]
>>>[FEN "6k1/2p3np/1p1p2p1/3P4/1PPK1R2/6PB/7P/4r3 w - - 0 1"]
>>>[PlyCount "11"]
>>>[EventDate "2005.12.28"]
>>>
>>>{GenuineIntel 447 MHz W=16.4 ply; 236kN/s; 477 TBAs B=18.4 ply; 65kN/s} 1. Bd7
>>>{1.06/13 132} Nh5 {0.29/14 55} 2. Be6+ {(Rf3) 4.03/15 115} Rxe6 {
>>>(Kg7) 2.72/16 191} 3. dxe6 {7.12/16 74} Nxf4 {10.00/21 1126} 4. gxf4 {
>>>7.02/16 61} Kf8 {9.94/18 23} 5. Kd5 {7.39/16 55} Ke7 {11.03/19 311} 6. b5 {
>>>7.58/18 84} 1-0
>>>
>>>Rybka's first move 1.-Nh5???? is a mega-gross error that loses right on the
>>>spot. And showing only a small advantage for white (+0.29). I hope it looks
>>>different with faster hardware. Misjudging the pawn endgame or what?
>>>
>>>/S
>>
>>Rybka needs depth 18 to find that Nh5 is losing
>>After 7 minutes of search it still did not change it's mind.
>>
>>Fritz 9 - Rybka 1.0 Beta 32-bit, Rybka-Fritz9 NunnE 40 4' 2005
>>[D]6k1/2pB2np/1p1p2p1/3P4/1PPK1R2/6P1/7P/4r3 b - - 0 1
>>
>>Analysis by Rybka 1.0 Beta 32-bit:
>>
>>1...Nh5
>> ± (0.74) Depth: 3 00:00:00
>>1...Nh5
>> ² (0.59) Depth: 4 00:00:00
>>1...Nh5 2.Rf3
>> ² (0.63) Depth: 5 00:00:00
>>1...Nh5 2.Rf3 Kg7
>> ² (0.54) Depth: 6 00:00:00 5kN
>>1...Nh5 2.Rf3 Kg7 3.Re3
>> ² (0.37) Depth: 7 00:00:00 11kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3
>> ² (0.51) Depth: 8 00:00:00 22kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6
>> ² (0.34) Depth: 9 00:00:00 52kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6 5.Ba4
>> ² (0.35) Depth: 10 00:00:00 103kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6 5.Re2 Rc1+ 6.Kd3
>> ² (0.36) Depth: 11 00:00:01 184kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6 5.Re2 h6 6.Ba4
>> ² (0.34) Depth: 12 00:00:02 380kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6 5.Re2 Kf7 6.Be6+
>> ² (0.28) Depth: 13 00:00:04 947kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6 5.Re2 Kf7 6.Be6+
>> ² (0.29) Depth: 14 00:00:07 1551kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Rc1+ 5.Kb3 Nf6 6.Re2
>> = (0.25) Depth: 15 00:00:16 3288kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Rc1+ 5.Kb3 Nf6 6.Re2
>> = (0.25) Depth: 16 00:00:25 5133kN
>>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6 5.Re2 Rf1 6.Bb5
>> = (0.24) Depth: 17 00:00:49 10104kN
>>1...Nh5 2.Be6+ Rxe6 3.dxe6 Nxf4 4.gxf4 h6 5.Kd5 Kf8 6.b5
>> +- (3.97) Depth: 18 00:02:15 30129kN
>>
>>(, 29.12.2005)
>
>Next lines in the analysis
>
>
>1...Re2 2.Kd3 Re7 3.Bg4 h5 4.Bc8 Rf7 5.Ke3 Re7+ 6.Kf3
> ² (0.61) Depth: 18 00:08:11 105113kN
>1...Re7 2.Bg4 h5 3.Bf3 Nf5+ 4.Kc3 Re3+ 5.Kd2 Ra3 6.Bxh5
> = (0.19) Depth: 18 00:10:43 137450kN
>1...Re7 2.Bg4 h5 3.Bf3 Nf5+ 4.Kc3 Re3+ 5.Kd2 Ra3 6.Bxh5
> = (0.20) Depth: 19 00:11:03 141845kN
>
>(, 29.12.2005)
>
>For comparison slightly positional shows the following analysis
>
>Fritz 9 - Rybka 1.0 Beta 32-bit, Rybka-Fritz9 NunnE 40 4' 2005
>6k1/2pB2np/1p1p2p1/3P4/1PPK1R2/6P1/7P/4r3 b - - 0 1
>
>Analysis by Rybka 1.0 Beta 32-bit:
>
>1...Nh5
> ± (0.74) Depth: 3 00:00:00
>1...Nh5
> ² (0.59) Depth: 4 00:00:00
>1...Nh5 2.Rf3
> ² (0.63) Depth: 5 00:00:00
>1...Nh5 2.Rf3 Kg7
> ² (0.54) Depth: 6 00:00:00 5kN
>1...Nh5 2.Rf3 Kg7 3.Re3
> ² (0.37) Depth: 7 00:00:00 11kN
>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3
> ² (0.51) Depth: 8 00:00:00 22kN
>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6
> ² (0.34) Depth: 9 00:00:00 51kN
>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6 5.Ba4
> ² (0.35) Depth: 10 00:00:00 103kN
>1...Nh5 2.Rf3 Kg7 3.Re3 Rd1+ 4.Kc3 Nf6 5.Re2 Kf7
> ² (0.35) Depth: 11 00:00:01 185kN
>1...Nh5 2.Rf3 Kg7 3.Ba4 Nf6 4.h4 Kf7 5.Re3 Rc1
> ² (0.27) Depth: 12 00:00:02 489kN
>1...Nh5 2.Be6+ Rxe6 3.dxe6 Nxf4 4.gxf4 c6 5.b5 cxb5 6.cxb5
> ² (0.34) Depth: 13 00:00:05 1134kN
>1...Nh5 2.Be6+ Rxe6 3.dxe6 Nxf4 4.gxf4 h6 5.Kd5 Kf8 6.b5
> +- (1.44) Depth: 14 00:00:09 2083kN
>1...Re2 2.Kd3 Re7 3.Bg4 h5 4.Bd1 Rf7 5.Ke4 Re7+ 6.Kf3
> ² (0.60) Depth: 14 00:00:17 3627kN
>
>(, 29.12.2005)
>
>It seems that vasik believes that these 4 plies are not important(otherwise he
>would use slightly positional and not the default personality in the
>tournament).
>
>Uri
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.