Nice one

Discussion of anything and everything relating to chess playing software and machines.

Moderators: hgm, Dann Corbit, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
Post Reply
Jack Lad
Posts: 8557
Joined: Wed Mar 08, 2006 9:01 pm
Location: UK

Nice one

Post by Jack Lad » Wed Mar 18, 2009 3:32 pm

[d] rnb1qk1r/pp3ppp/2p5/3pP1B1/8/2N3Q1/PPP2PPP/R4RK1 w

Shredder Classic takes 27 seconds to find the best move. Try to solve this one in your head first before testing you engine. 8-)
Now cracks a noble heart.—Good night, sweet Princess, And flights of angels sing thee to thy rest!

Damir
Posts: 2472
Joined: Mon Feb 11, 2008 2:53 pm
Location: Denmark
Full name: Damir Desevac

Re: Nice one

Post by Damir » Wed Mar 18, 2009 4:53 pm

No need to use the engine, Bf6 !! is the winning move.....

Michael Sherwin
Posts: 3196
Joined: Fri May 26, 2006 1:00 am
Location: WY, USA
Full name: Michael Sherwin

Re: Nice one

Post by Michael Sherwin » Wed Mar 18, 2009 5:00 pm

RomiChess (one core) && Glaurung 2.1 (four core) find Nc3xd5 in just over 1 second at a search depth of 14.

What I find interesting is that Fruit 2.1 finds the same move (also in just over one second) at a depth of 9!
If you are on a sidewalk and the covid goes beep beep
Just step aside or you might have a bit of heat
Covid covid runs through the town all day
Can the people ever change their ways
Sherwin the covid's after you
Sherwin if it catches you you're through

User avatar
michiguel
Posts: 6401
Joined: Thu Mar 09, 2006 7:30 pm
Location: Chicago, Illinois, USA
Contact:

Re: Nice one

Post by michiguel » Wed Mar 18, 2009 5:31 pm

Michael Sherwin wrote:RomiChess (one core) && Glaurung 2.1 (four core) find Nc3xd5 in just over 1 second at a search depth of 14.

What I find interesting is that Fruit 2.1 finds the same move (also in just over one second) at a depth of 9!
An old version of Gaviota that I have here in this computer finds it a ply 8, Bf6 at ply 7. Takes longer to find it though!
I do not think there is any merit anymore at what ply an engine finds what with all the heavy pruning that many engines do.

Miguel

Code: Select all

         3   1       0.0    -1.00  1.Nxd5 cxd5
         6   1       0.0    +0.95  1.f4
        10   1       0.0    +1.00  1.f3
        14   1       0.0    +1.02  1.a4
        20   1       0.0    +1.03  1.h3
        66   1       0.0    +1.04  1.Qf4
        80   1:      0.0    +1.04  1.Qf4
       146   2       0.0    +0.79  1.Qf4 Kg8
       379   2:      0.0    +0.79  1.Qf4 Kg8
       613   3       0.0    +0.82  1.Qf4 Kg8 2.Bh4
      2937   3:      0.0    +0.82  1.Qf4 Kg8 2.Bh4
      3626   4       0.0    +0.66  1.Qf4 Kg8 2.Bh4 Be6
     11984   4:      0.1    +0.66  1.Qf4 Kg8 2.Bh4 Be6
     15554   5       0.1    +0.78  1.Qf4 Kg8 2.Rfe1 Na6 3.Qf3
     44546   5:      0.2    +0.78  1.Qf4 Kg8 2.Rfe1 Na6 3.Qf3
     63088   6       0.2    +0.64  1.Qf4 Kg8 2.a3 Nd7 3.Rfe1 Nc5
    186424   6:      0.6    +0.64  1.Qf4 Kg8 2.a3 Nd7 3.Rfe1 Nc5
    274236   7       0.9    +0.62  1.Qf4 h6 2.Bh4 g5 3.Qf6 Kg8 4.Bg3
    917469   7       2.9      :-)  1.Bf6
    937263   7       3.0    +1.81  1.Bf6 Rg8 2.Nxd5 gxf6 3.Qa3+ c5 4.Qxc5+
                                   Kg7 5.Nxf6
    943905   7:      3.0    +1.81  1.Bf6 Rg8 2.Nxd5 gxf6 3.Qa3+ c5 4.Qxc5+
                                   Kg7 5.Nxf6
    972293   8       3.1      :-)  1.Bf6
   1068841   8       3.3    +2.45  1.Bf6 Rg8 2.Nxd5 gxf6 3.Qa3+ Kg7 4.Nc7
                                   Qxe5 5.Nxa8
   1573237   8       4.9      :-)  1.Nxd5
   1655795   8       5.1    +3.10  1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
                                   4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rae1 Kc7
                                   7.Re7+ Bd7 8.Qxh7 Qxf6 9.Qxf7 Qxf7
                                   10.Rxf7
   1661843   8:      5.1    +3.10  1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
                                   4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rae1 Kc7
                                   7.Re7+ Bd7 8.Qxh7 Qxf6 9.Qxf7 Qxf7
                                   10.Rxf7
   1882353   9       5.8    +3.10  1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
                                   4.Qg7+ <=TRANS
   2185510   9&#58;      6.6    +3.10  1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
                                   4.Qg7+ <=TRANS
   3071317  10       9.3    +3.34  1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
                                   4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Qg7 Qxc2
                                   7.Qxf7+ Kc6 8.Rad1 Qxb2 9.Qxd5+ Kc7
   3696518  10&#58;     10.9    +3.34  1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
                                   4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Qg7 Qxc2
                                   7.Qxf7+ Kc6 8.Rad1 Qxb2 9.Qxd5+ Kc7
   5736664  11      17.1    +3.28  1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
                                   4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Qg7 Kc7
                                   7.Qxf7+ Bd7 8.Qxh7 Qxf6
  13289556  11&#58;     38.1    +3.28  1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
                                   4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Qg7 Kc7
                                   7.Qxf7+ Bd7 8.Qxh7 Qxf6
  19809532  12      56.9    +3.28  1.Nxd5 cxd5 2.Bf6 g6 3.Bxh8 Kg8 4.Bf6
                                   Nd7 5.Bg5 Qxe5 6.Qxe5 Nxe5 7.Rfe1

Michael Sherwin
Posts: 3196
Joined: Fri May 26, 2006 1:00 am
Location: WY, USA
Full name: Michael Sherwin

Re: Nice one

Post by Michael Sherwin » Wed Mar 18, 2009 6:22 pm

Are the reductions in Fruit and Glaurung that much different? Or does Fruit have extensions that ...

If Fruit has extensions that Glaurung does not have, then can Glaurung benefit from adding those extensions to the last few plys of it's search with out loosing much depth?

This is what I meant by, "interesting".
If you are on a sidewalk and the covid goes beep beep
Just step aside or you might have a bit of heat
Covid covid runs through the town all day
Can the people ever change their ways
Sherwin the covid's after you
Sherwin if it catches you you're through

User avatar
Eelco de Groot
Posts: 4279
Joined: Sun Mar 12, 2006 1:40 am
Location: Groningen

Re: Nice one

Post by Eelco de Groot » Thu Mar 19, 2009 2:54 pm

This type of position is one where Rebel excels at so probably the Rebel-type extensions will help here.

Good old Pro Deo 1.5, with Judith Polgar settings: 0.4 seconds on Athlon 2009MHZ, 192Mb.

Code: Select all

00&#58;00&#58;00.1	-1,10	1	7	Nxd5 cxd5 
00&#58;00&#58;00.1	-1,03	1	15	Bf6 gxf6 exf6 
00&#58;00&#58;00.1	0,52	1	17	Rae1 
00&#58;00&#58;00.1	0,68	1	19	Rad1 
00&#58;00&#58;00.1	0,25	2	159	Rad1 Nd7 
00&#58;00&#58;00.1	0,50	2	266	f4 Nd7 
00&#58;00&#58;00.1	0,45	3	909	f4 Bf5 Rac1 
00&#58;00&#58;00.2	0,51	3	2039	Ne2 Nd7 f4 
00&#58;00&#58;00.2	0,41	4	4269	Ne2 Nd7 f4 c5 
00&#58;00&#58;00.2	0,41	4	15423	e6 
00&#58;00&#58;00.2	0,87	4	16190	e6 Qxe6 Rae1 
00&#58;00&#58;00.2	0,87	4	17011	Bf6 
00&#58;00&#58;00.2	1,29	4	17563	Bf6 Rg8 Nxd5 gxf6 Qxg8+ Kxg8 Nxf6+ Kf8 Nxe8 Kxe8 
00&#58;00&#58;00.2	1,29	5	20550	Bf6 Rg8 Nxd5 gxf6 Qxg8+ Kxg8 Nxf6+ Kf8 Nxe8 Kxe8 
00&#58;00&#58;00.2	1,29	6	34290	Bf6 Rg8 Nxd5 gxf6 Qxg8+ Kxg8 Nxf6+ Kf8 Nxe8 Kxe8 
00&#58;00&#58;00.2	1,29	7	66483	Bf6 Rg8 Nxd5 gxf6 Qxg8+ Kxg8 Nxf6+ Kf8 Nxe8 Kxe8 
00&#58;00&#58;00.3	1,51	8	151492	Bf6 Rg8 Nxd5 gxf6 Qxg8+ Kxg8 Nxf6+ Kf8 Nxe8 Kxe8 f4 Nd7 
00&#58;00&#58;00.4	1,51	8	220722	Nxd5 
00&#58;00&#58;00.5	1,91	8	327151	Nxd5 cxd5 Bf6 gxf6 exf6 Qd8 Qg7+ Ke8 Qxh8+ Kd7 Qxd8+ Kxd8 Rae1 
00&#58;00&#58;00.7	2,05	9	601044	Nxd5 cxd5 Bf6 gxf6 exf6 Qd8 Qg7+ Ke8 Qxh8+ Kd7 Qg7 Kc7 Qxf7+ Nd7 Rae1 
00&#58;00&#58;01.5	2,05	10	1730384	Nxd5 cxd5 Bf6 gxf6 exf6 Qd8 Qg7+ Ke8 Qxh8+ Kd7 Qg7 Kc7 Qxf7+ Nd7 Rae1 
00&#58;00&#58;05.3	2,32	11	6646746	Nxd5 cxd5 Bf6 gxf6 exf6 Qd8 Qg7+ Ke8 Qxh8+ Kd7 Qg7 Kd6 Qxf7 Nd7 Rae1 Kc7 Qxh7 Qxf6 
00&#58;00&#58;12.7	2,32	12	16418905	Nxd5 cxd5 Bf6 gxf6 exf6 Qd8 Qg7+ Ke8 Qxh8+ Kd7 Qg7 Kd6 Qxf7 Nd7 Rae1 Kc7 Qxh7 Qxf6 
00&#58;00&#58;40.2	2,34	13	53220164	Nxd5 cxd5 Bf6 g6 Qe3 Kg8 Qh6 Qf8 Qxf8+ Kxf8 Bxh8 Nc6 Rfd1 Kg8
Number of plies does not say very much, for instance look at the disastrous result from latest Blueberry: 5 plies and 2 minutes 22 seconds.

[D]rnb1qk1r/pp3ppp/2p5/3pP1B1/8/2N3Q1/PPP2PPP/R4RK1 w - -

Engine: Blueberry Beta 4 DM70 Build 416 (Athlon 2009 MHz, 128 MB)
by F. Letouzey, T. Gaksch, E. de Groot

1/01 0:00 +3.90 1.Rab1 (3)

1/01 0:00 +4.14 1.Rad1 (7)

1/01 0:00 +5.18 1.Rae1 (9)

2/16 0:00 +3.41 1.Rae1 Kg8 2.Bf6 Qf8 (27.406)

3/33 0:01 +3.94 1.Rae1 Bf5 2.Na4 (865.101) 607

4/37 0:02 +3.32 1.Rae1 Bf5 2.Qf4 Bxc2 3.Re2 Nd7
4.Rxc2 Qxe5 5.Qb4+ Kg8 6.Qxb7 (1.182.611) 590

5/49 0:17 +2.38 1.Rae1 Bf5 2.Qf3 Bg6 3.e6 Kg8 4.Qg4 fxe6
5.Rxe6 (10.440.176) 596

5/49 2:21 +3.23 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rae1 Kc7
7.Re7+ Bd7 8.Rxf7 Qxc2 9.Rxd7+ Nxd7
10.Qxa8 Nxf6 11.Qxa7 Qxb2 12.Qc5+ Kb8
13.Qd6+ Ka7 (79.688.596) 563

5/49 2:22 +3.23 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rae1 Kc7
7.Re7+ Bd7 8.Rxf7 Qxc2 9.Rxd7+ Nxd7
10.Qxa8 Nxf6 11.Re1 Qxb2 12.f3 Nd7
13.Qxa7 (80.454.506) 564

5/58 2:40 +3.23 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rae1 Kc7
7.Re7+ Bd7 8.Rxf7 Qxc2 9.Rxd7+ (92.753.900) 576

6/58 2:44 +3.22 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rad1 Kc7
7.Qxh7 Qxf6 8.Rxd5 Qxb2 9.Qxf7+ Bd7 (95.144.218) 577

7/58 3:52 +3.22 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rad1 Kc7
7.Qxh7 Qxf6 8.Rxd5 Qxb2 9.Qxf7+ Bd7 (133.102.056) 573

8/58 9:48 +3.58 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.c4 d4
7.Rad1 Qc5 8.Rfe1 Qxc4 9.Re7+ Kd6
10.Qd8+ Nd7 11.Rxd7+ Bxd7 12.Qxa8 (330.367.491) 561

9/58 12:35 +3.23 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rae1 Kc7
7.Re7+ Bd7 8.Rxf7 Qxc2 9.Rxd7+ Nxd7
10.Qxa8 Nxf6 11.Qxa7 Qxb2 12.Qc5+ Kb8
13.Qd6+ Ka7 (424.281.511) 561

10/58 44:55 +3.23 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rae1 Kc7
7.Re7+ Bd7 8.Rxf7 Qxc2 9.Rxd7+ Nxd7
10.Qxa8 Nxf6 11.Qxa7 Qxb2 12.Qc5+ Kb8
13.Qd6+ Ka7 (1.520.591.695) 564

11/59 106:38 +3.20 1.Nxd5 cxd5 2.Bf6 g6 3.Bxh8 Kg8
4.Bf6 Bf5 5.Rad1 Nd7 6.Rxd5 Rc8
7.Qh4 Rc6 8.Rd4 h5 9.Ra4 Nxf6
10.exf6 Rxc2 11.Rxa7 Rxb2 (3.604.622.019) 563


best move: Nc3xd5 time: 160:01.563 min n/s: 561.307 CPU 100.0% n/s(1CPU): 561.307 nodes: 5.389.340.000

But, not all is lost, I was a bit desperate after above result with build 416 but today I rewound 41 versions and got a better result with build 375. All later versions extend the PV with null window searches, but as Fruit shows "Keep It Simple Stupid" still works best here :)

rnb1qk1r/pp3ppp/2p5/3pP1B1/8/2N3Q1/PPP2PPP/R4RK1 w - -

Engine: Blueberry Beta 4 DM70 Build 375 (Athlon 2009 MHz, 128 MB)
by F. Letouzey, T. Gaksch, E. de Groot

1/01 0:00 +3.90 1.Rab1 (3)

1/01 0:00 +4.14 1.Rad1 (7)

1/01 0:00 +5.18 1.Rae1 (9)

2/17 0:00 +2.93 1.Rae1 h6 2.Be3 Kg8 3.Bd4 Bf5 (8.349)

3/30 0:00 +2.63 1.Rae1 Bf5 2.Qf4 Bxc2 3.Qb4+ Kg8
4.Qxb7 Nd7 5.f4 (273.680)

3/34 0:01 +3.14 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Qxh7 Qxf6
7.Rad1 Kc6 (860.233) 550

3/35 0:01 +3.14 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Qg7 Kc7
7.Rfe1 Qxc2 8.Qg3+ Kb6 9.Qd6+ Ka5
10.Qxd5+ Ka6 (987.197) 550

3/44 0:02 +3.14 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rae1 Kc7
7.Re7+ Nd7 8.Qg7 Qxc2 9.Qg3+ Kb6
10.Qd6+ Ka5 11.Qxd5+ Nc5 12.Rxf7 (1.748.424) 575

4/44 0:02 +2.91 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rfe1 Qxc2
7.Re7+ Kd6 8.Qd8+ Bd7 9.Rxd7+ Nxd7
10.Qxa8 (1.830.466) 621

5/44 0:05 +3.30 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Rae1+ Kd8 6.Qxh8+ Kc7
7.Re7+ Kb6 (3.252.212) 612

6/44 0:11 +3.22 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rad1 Kc7
7.Qxh7 Qxf6 8.Rxd5 Qxb2 9.Qxf7+ Bd7 (6.924.363) 622

7/46 0:37 +3.22 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Rad1 Kc7
7.Qxh7 Qxf6 8.Rxd5 Qxb2 9.Qxf7+ Bd7 (23.165.913) 622

8/46 1:55 +3.34 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qd8
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Qg7 Kd6
7.Qxf7 Nd7 8.Rad1 Nxf6 9.Rfe1 Kc6 (71.405.485) 618

9/46 3:16 +3.47 1.Nxd5 cxd5 2.Bf6 gxf6 3.exf6 Qc6
4.Qg7+ Ke8 5.Qxh8+ Kd7 6.Qg7 Kd6
7.Qg3+ Kd7 8.Rfe1 (120.593.806) 615

best move: Nc3xd5 time: 3:26.156 min n/s: 614.810 CPU 100.0% n/s(1CPU): 614.810 nodes: 126.670.000

Eelco
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan

rightrook
Posts: 1452
Joined: Wed Mar 08, 2006 7:45 pm

Re: Nice one

Post by rightrook » Fri Mar 20, 2009 7:08 pm

Hi...The best move is really.....Nxd5...!!

Ruffian 1.0.5 finds Bf6 at only 8 ply...but at 14 ply...changes to
Nxd5.... :lol:

regards
Robert

Post Reply