Author: Richard Pijl
Date: 01:25:14 09/15/04
Go up one level in this thread
On September 14, 2004 at 19:50:10, GuyHaworth wrote:
>
>I had occasion to return to a DEEP JUNIOR -v- DEEP FRITZ position of the 'Deep'
>C-C match, Game 4, 25/04/2001:
>
>r7/8/5R2/2k5/8/3R1K2/6P1/8 w:
>Play continued
>70. Rc6+ Kb5 71. Rc5+ Kb4 72. Rb5+ Kc4 73. Rd4+ Kc3 74. Rc5+ K*d4
>
>I guess DJ is trying to drop into a won KRPKR endgame, DF is trying to avoid
>doing that as long as possible, and DJ is assuming that that is what DF will do.
>
>However, if someone has the KRRPKR DTM EGT to hand, we can perhaps see how the
>DTM-minimaxing line goes from this position.
>
>Incidentally, at the time, Tony Werten suggested that with a deeper search, DJ
>should have gone:
>70. g4 Kc4 71. Rc6 Kxd3 and mate in 121 plies
>Is this the best line that can be identified today with KRPKR EGTs?
The Baron with 5 men EGTB's prints a mate score almost instantly (although not
the shortest path available):
ply time nodes score pv
4( 9)+ 0:00.11 2651 MATE55 Rf6-c6 (Kc5-b5) (Rd3-b3) (Kb5xc6)
4( 9)& 0:00.14 3384 MATE55 Rf6-c6 Kc5-b5 Rd3-b3 Kb5xc6
4(11). 0:00.19 6593 MATE55 Rf6-c6 Kc5-b5 Rd3-b3 Kb5xc6
5(12)& 0:00.22 9219 MATE55 Rf6-c6 Kc5-b5 Rd3-b3 Kb5xc6
5(13)& 0:00.39 20964 MATE40 Rf6-f5 Kc5-b6 Rf5-b5 Kb6-c6 Rb5-c5 Kc6-b6
Rd3-d6 Kb6-b7 Rc5-c7 Kb7xc7
5(13)& 0:00.44 27887 MATE39 Rd3-d5 Kc5-b4 Rf6-b6 Kb4-c4 Rd5-c5 Kc4-d4
Rb6-d6 Kd4xc5
5(13). 0:00.44 27901 MATE39 Rd3-d5 Kc5-b4 Rf6-b6 Kb4-c4 Rd5-c5 Kc4-d4
Rb6-d6 Kd4xc5
6(13)& 0:00.52 43981 MATE39 Rd3-d5 Kc5-b4 Rf6-b6 Kb4-c4 Rd5-c5 Kc4-d4
Rb6-d6 Kd4xc5
6(13). 0:00.83 64779 MATE39 Rd3-d5 Kc5-b4 Rf6-b6 Kb4-c4 Rd5-c5 Kc4-d4
Rb6-d6 Kd4xc5
7(16)& 0:01.08 92138 MATE39 Rd3-d5 Kc5-b4 Rf6-b6 Kb4-c4 Rd5-c5 Kc4-d4
Rb6-d6 Kd4xc5
7(16)& 0:01.49 139054 MATE38 Rd3-c3 Kc5-d4 Rf6-f5 Ra8-a7 Rc3-c4 Kd4-d3
Rf5-d5 Kd3xc4
7(16). 0:02.44 313828 MATE38 Rd3-c3 Kc5-d4 Rf6-f5 Ra8-a7 Rc3-c4 Kd4-d3
Rf5-d5 Kd3xc4
8(18)& 0:03.41 529863 MATE36 Rd3-c3 Kc5-d4 g2-g4 Ra8-h8 Rc3-c5 Rh8-e8
(Rf6-d6) (Kd4xc5)
8(18). 0:05.55 904516 MATE36 Rd3-c3 Kc5-d4 g2-g4 Ra8-h8 Rc3-c5 Rh8-e8
(Rf6-d6) (Kd4xc5)
9(21)& 0:06.55 1155255 MATE36 Rd3-c3 Kc5-d5 g2-g4 Ra8-a1 Rc3-c5 Kd5-d4
Rf6-d6 Kd4xc5
9(21)& 0:07.84 1483680 MATE35 Rf6-c6 Kc5-b5 Rc6-c8 Ra8xc8
9(21)& 0:09.31 2021208 MATE34 g2-g4 Ra8-h8 Rd3-c3 Kc5-d4 (Kf3-f4)
(Rh8-d8)
(Rc3-d3) (Kd4-c4) (Rf6-c6) (Kc4-b5)
(Rd3-b3) (Kb5xc6)
9(21). 0:09.31 2027861 MATE34 g2-g4 Ra8-h8 Rd3-c3 Kc5-d4 (Kf3-f4)
(Rh8-d8)
(Rc3-d3) (Kd4-c4) (Rf6-c6) (Kc4-b5)
(Rd3-b3) (Kb5xc6)
10(22)& 0:09.72 2246718 MATE34 g2-g4 Ra8-h8 Rd3-c3 Kc5-d4 Kf3-f4 Rh8-d8
Rc3-d3 Kd4-c4 Rf6-c6 Kc4-b5 Rd3-b3 Kb5xc6
10(22). 0:10.44 2537445 MATE34 g2-g4 Ra8-h8 Rd3-c3 Kc5-d4 Kf3-f4 Rh8-d8
Rc3-d3 Kd4-c4 Rf6-c6 Kc4-b5 Rd3-b3 Kb5xc6
11(23)& 0:14.22 3998632 MATE34 g2-g4 Ra8-a1 Rf6-c6 Kc5-b5 Rc6-c1 Ra1-a2
Rd3-b3 Kb5-a4 Rc1-c7 Ra2-d2 Rc7-a7 Ka4xb3
11(23). 0:17.45 4955766 MATE34 g2-g4 Ra8-a1 Rf6-c6 Kc5-b5 Rc6-c1 Ra1-a2
Rd3-b3 Kb5-a4 Rc1-c7 Ra2-d2 Rc7-a7 Ka4xb3
12(26)& 0:22.80 7070109 MATE33 g2-g4 Ra8-a1 Rf6-c6 Kc5-b5 Rc6-c1 Ra1xc1
12(26). 0:28.47 9316295 MATE33 g2-g4 Ra8-a1 Rf6-c6 Kc5-b5 Rc6-c1 Ra1xc1
13(26)& 0:46.81 17804369 MATE32 g2-g4 Ra8-a1 Rf6-c6 Kc5-b5 g4-g5 Ra1-f1
Kf3-g4 Rf1-g1 Rd3-g3 Rg1-e1 Rg3-b3 Kb5-a4
Rc6-c4 Ka4-a5 Rc4-a4 Ka5xa4
13(27). 0:55.39 22114229 MATE32 g2-g4 Ra8-a1 Rf6-c6 Kc5-b5 g4-g5 Ra1-f1
Kf3-g4 Rf1-g1 Rd3-g3 Rg1-e1 Rg3-b3 Kb5-a4
Rc6-c4 Ka4-a5 Rc4-a4 Ka5xa4
14(30)& 2:22.56 68711171 MATE30 g2-g4 Ra8-a1 g4-g5 Ra1-f1 Kf3-g2 Rf1-e1
g5-g6 Re1-e2 Kg2-f3 Re2-e8 Rd3-c3 Kc5-b4
Rf6-b6 Kb4-a5 g6-g7 Re8-d8 Rc3-a3 Ka5xb6
14(30). 2:39.42 78386392 MATE30 g2-g4 Ra8-a1 g4-g5 Ra1-f1 Kf3-g2 Rf1-e1
g5-g6 Re1-e2 Kg2-f3 Re2-e8 Rd3-c3 Kc5-b4
Rf6-b6 Kb4-a5 g6-g7 Re8-d8 Rc3-a3 Ka5xb6
without EGTB's the Baron cannot find a mate within an hour on a Xeon 2.8, but it
will play g4 almost instantly. The line found in an hour:
18(42)& 61:46.74 -1433891367 +10.50 g2-g4 Kc5-c4 Rd3-d7 Ra8-a1 g4-g5 Kc4-c5
Kf3-g4 Kc5-b4 g5-g6 Ra1-g1 Kg4-f4 Rg1-f1
Kf4-e5 Rf1-g1 g6-g7 Rg1-e1 Ke5-f5 Re1-f1
Kf5-g6 Rf1-g1 Kg6-f7 Rg1xg7 Kf7xg7
Richard.
>
>I wonder how this will compare with the KRRPKR DTM-minimaxing line.
>
>g
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.