Seperating the men from the toys - test position

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

Moderator: Ras

Howard E
Posts: 261
Joined: Wed Mar 08, 2006 8:49 pm

Seperating the men from the toys - test position

Post by Howard E »

in this case toys = chess programs
So easy for humans to spot.

[d] 8/4p3/rr1p1p2/bkp1PPp1/1p1P2Pp/p1PK3P/PPB5/8 w - -
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Seperating the men from the toys - test position

Post by bob »

Howard E wrote:in this case toys = chess programs
So easy for humans to spot.

[d] 8/4p3/rr1p1p2/bkp1PPp1/1p1P2Pp/p1PK3P/PPB5/8 w - -
I assume you are looking for 1. c4+ Kc6 2. d5+ K-any 3. d5 and then putting the B on b3 to lock things up? Crafty will play this instantly, but it has no clue it is a draw because it will take a ton of move to either reach the 50 move counter or cause black to repeat...

Also looks like c4, d5 and b3 will work too...
Last edited by bob on Mon Apr 06, 2009 10:09 pm, edited 1 time in total.
IWB
Posts: 1539
Joined: Thu Mar 09, 2006 2:02 pm

Re: Seperating the men from the toys - test position

Post by IWB »

bob wrote:
Howard E wrote:in this case toys = chess programs
So easy for humans to spot.

[d] 8/4p3/rr1p1p2/bkp1PPp1/1p1P2Pp/p1PK3P/PPB5/8 w - -
I assume you are looking for 1. c4+ Kc6 2. d5+ K-any 3. d5 and then putting the B on b3 to lock things up? Crafty will play this instantly, but it has no clue it is a draw because it will take a ton of move to either reach the 50 move counter or cause black to repeat...
This one is even better (found by Shredder instantly)



29/44 0:24 -0.29 1.Ba4+ Kxa4 2.b3+ Kb5 3.c4+ Kc6
4.d5+ Kb7 5.e6 (77.504.863) 3179 SB:8

best move: Bc2-a4 time: 0:29.047 min n/s: 3.164.246 CPU 396.0% n/s(1CPU): 799.052 nodes: 91.810.596 SB: 11

It locks up everything 'forced' without the Bishop!

BYe
Ingo
IWB
Posts: 1539
Joined: Thu Mar 09, 2006 2:02 pm

Re: Seperating the men from the toys - test position

Post by IWB »

Hello

I just checked with a few top engines (Top 3) and to my suprise they all have problems finding Ba4+! The drawish eval is even more difficult.

So this is the output of the latest official Shredder:

[d]8/4p3/rr1p1p2/bkp1PPp1/1p1P2Pp/p1PK3P/PPB5/8 w - -

Code: Select all

Engine: Shredder WM Edition Bonn (1024 MB)
by Stefan Meyer-Kahlen

22/34  0:01   -0.29    1.Ba4+ Kxa4 2.b3+ Kb5 3.c4+ Kc6 
                       4.d5+ Kd7 5.e6+ Ke8 6.Ke4 Rb5 7.Kf3 Rb7 
                       8.Ke3 Ra8 9.Ke4 Bb6 10.Ke3 Kf8 
                       11.Ke4 Kg7 12.Ke3 Rc7 13.Ke4 Rcc8 
                       14.Ke3 (7.402.951) 5322 

23/35  0:02   -0.29    1.Ba4+ Kxa4 2.b3+ Kb5 3.c4+ Kc6 
                       4.d5+ Kd7 5.e6+ Ke8 6.Ke4 Rb5 7.Kf3 Rb7 
                       8.Ke3 Ra8 9.Ke4 Bb6 10.Ke3 Kf8 
                       11.Ke4 Kg7 12.Ke3 Rc7 13.Ke4 Rcc8 
                       14.Ke3 (11.708.703) 5390 

24/36  0:03   -0.29    1.Ba4+ Kxa4 2.b3+ Kb5 3.c4+ Kc6 
                       4.d5+ Kd7 5.e6+ Ke8 6.Ke4 Rb5 7.Kf3 Rb7 
                       8.Ke3 Ra8 9.Ke4 Bb6 10.Ke3 Kf8 
                       11.Ke4 Kg7 12.Ke3 Rc7 13.Ke4 Rcc8 
                       14.Ke3 (19.970.505) 5325 
best move: Bc2-a4 time: 0:04.469 min n/s: 5.289.217 CPU 394.6% n/s(1CPU): 1.340.399 nodes: 23.473.548 SB: 1

1s and -0.29 eval - not too bad

BYe
Ingo
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Seperating the men from the toys - test position

Post by bob »

IWB wrote:Hello

I just checked with a few top engines (Top 3) and to my suprise they all have problems finding Ba4+! The drawish eval is even more difficult.

So this is the output of the latest official Shredder:

[d]8/4p3/rr1p1p2/bkp1PPp1/1p1P2Pp/p1PK3P/PPB5/8 w - -

Code: Select all

Engine: Shredder WM Edition Bonn (1024 MB)
by Stefan Meyer-Kahlen

22/34  0:01   -0.29    1.Ba4+ Kxa4 2.b3+ Kb5 3.c4+ Kc6 
                       4.d5+ Kd7 5.e6+ Ke8 6.Ke4 Rb5 7.Kf3 Rb7 
                       8.Ke3 Ra8 9.Ke4 Bb6 10.Ke3 Kf8 
                       11.Ke4 Kg7 12.Ke3 Rc7 13.Ke4 Rcc8 
                       14.Ke3 (7.402.951) 5322 

23/35  0:02   -0.29    1.Ba4+ Kxa4 2.b3+ Kb5 3.c4+ Kc6 
                       4.d5+ Kd7 5.e6+ Ke8 6.Ke4 Rb5 7.Kf3 Rb7 
                       8.Ke3 Ra8 9.Ke4 Bb6 10.Ke3 Kf8 
                       11.Ke4 Kg7 12.Ke3 Rc7 13.Ke4 Rcc8 
                       14.Ke3 (11.708.703) 5390 

24/36  0:03   -0.29    1.Ba4+ Kxa4 2.b3+ Kb5 3.c4+ Kc6 
                       4.d5+ Kd7 5.e6+ Ke8 6.Ke4 Rb5 7.Kf3 Rb7 
                       8.Ke3 Ra8 9.Ke4 Bb6 10.Ke3 Kf8 
                       11.Ke4 Kg7 12.Ke3 Rc7 13.Ke4 Rcc8 
                       14.Ke3 (19.970.505) 5325 
best move: Bc2-a4 time: 0:04.469 min n/s: 5.289.217 CPU 394.6% n/s(1CPU): 1.340.399 nodes: 23.473.548 SB: 1

1s and -0.29 eval - not too bad

BYe
Ingo
That's actually more than just "not too bad." That looks really impressive. There were old versions of Crafty that would solve this instantly, as they had a locked pawn structure term in the evaluation. But it is not there in the new versions and it really has no chance.

As a human, the bishop check was my choice, as it is obvious. I tried Crafty on this a bit and it looks like there are several ways to draw it so long as white strives to lock things up as tightly as possible and doesn't get tricked into winning a pawn and losing the game...
Nimzovik
Posts: 1831
Joined: Sat Jan 06, 2007 11:08 pm

Re: Seperating the men from the toys - test position

Post by Nimzovik »

Yes....... All too true. This is "Pablo Chess". SOME day programs will handle stuff like this (and other positional motifs)....I hope in my lifetime.

Which engine handles such positional stuff the best today yet still maintains a high ELO in your opinion? My assertion was W chess......tho no facts to back up my claim.
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Seperating the men from the toys - test position

Post by bob »

Nimzovik wrote:Yes....... All too true. This is "Pablo Chess". SOME day programs will handle stuff like this (and other positional motifs)....I hope in my lifetime.

Which engine handles such positional stuff the best today yet still maintains a high ELO in your opinion? My assertion was W chess......tho no facts to back up my claim.
Shredder obviously "gets this" position. It's a trade-off. you can handle any type of position you want, the code isn't that hard to write. But it has to make you stronger and how just flashy on test positions...
Werewolf
Posts: 2035
Joined: Thu Sep 18, 2008 10:24 pm

Re: Seperating the men from the toys - test position

Post by Werewolf »

Shredder is the only one that gets this from my tests. Interestingly even Shredder 9 finds Ba4 instantly with a -0.4 evaluation!

Programs that failed were

Rybka 3
HIARCS 12
Fritz 10 & 11
Zappa M II
Junior 10
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: Seperating the men from the toys - test position

Post by bob »

Werewolf wrote:Shredder is the only one that gets this from my tests. Interestingly even Shredder 9 finds Ba4 instantly with a -0.4 evaluation!

Programs that failed were

Rybka 3
HIARCS 12
Fritz 10 & 11
Zappa M II
Junior 10
note that that doesn't seem to be the _only_ way to force a draw, just "a way"...
Cubeman
Posts: 644
Joined: Fri Feb 02, 2007 3:11 am
Location: New Zealand

Re: Seperating the men from the toys - test position

Post by Cubeman »

bob wrote:
Howard E wrote:in this case toys = chess programs
So easy for humans to spot.

[d] 8/4p3/rr1p1p2/bkp1PPp1/1p1P2Pp/p1PK3P/PPB5/8 w - -
I assume you are looking for 1. c4+ Kc6 2. d5+ K-any 3. d5 and then putting the B on b3 to lock things up? Crafty will play this instantly, but it has no clue it is a draw because it will take a ton of move to either reach the 50 move counter or cause black to repeat...

Also looks like c4, d5 and b3 will work too...
How can White play both 2.d5+ and 3.d5?
To me it looks like there is only 1 way to draw which is the line 1.Ba4+ etc.Any thing else will alow the pawn structure to be opened up with either axb2, dxe5, fxe5.