STS 1-15 epd and analysis tool

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

Moderators: hgm, Rebel, chrisw

Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: STS 1-15 epd and analysis tool

Post by Ferdy »

Ferdy wrote: During conversion I have encountered couple of epd issues. Here they are.
[1]
STS1, missing comma after Rg3=7

Code: Select all

1n6/4bk1r/1p2rp2/pP2pN1p/K1P1N2P/8/P5R1/3R4 w - - bm c5; id "Undermine.003"; c0 "c5=10, Rd3=7, Rdd2=7, Rg3=7Rd5=9";
I re-checked this epd and found out that it has space but no comma like the following.

Code: Select all

1n6/4bk1r/1p2rp2/pP2pN1p/K1P1N2P/8/P5R1/3R4 w - - bm c5; id "Undermine.003"; c0 "c5=10, Rd3=7, Rdd2=7, Rg3=7 Rd5=9";
Vinvin
Posts: 5228
Joined: Thu Mar 09, 2006 9:40 am
Full name: Vincent Lejeune

Re: STS 1-15 epd and analysis tool

Post by Vinvin »

Ferdy wrote:
Vinvin wrote:
Ferdy wrote:
Evert wrote:
Ferdy wrote:
Norm Pollock wrote:I noticed that there are 3 duplicate epd's

1q2r1k1/1b2bpp1/p2ppn1p/2p5/P3PP1B/2PB1RP1/2P1Q2P/2KR4 b - - c0 2; c1 line(s): 605 1405;
Right, one is from a/b/c pawn advancement, and the other is undermine, but have different move scoring. Could be dual purpose.

1r1r1bk1/1bq2p1p/pn2p1p1/2p1P3/5P2/P1NBB3/1P3QPP/R2R2K1 b - - c0 2; c1 line(s): 208 804;

r1r3k1/pb3p1p/1pqBp1p1/4P3/3b4/2P2P2/PR1N2PP/2RQ3K w - - c0 2; c1 line(s): 277 873;
Surely that makes the position useless then, because improving the score for one benchmark will decrease it for the other...?
I don't see why it is useless. If you are good in pawn advancement theme you will get the bm, and if you are good in undermining theme too you will also get the bm.
Because if you make an engine who play the very best moves in all positions, it won't get the maximum of points in all test, that's crazy ...
Pos 1 = 10, pos 2 = 10,
total points = 20, max points = 20, why it will not get max points?
OK, no problem so ! ;-)
Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: STS 1-15 epd and analysis tool

Post by Ferdy »

Vinvin wrote:
Ferdy wrote:
Vinvin wrote:
Ferdy wrote:
Evert wrote:
Ferdy wrote:
Norm Pollock wrote:I noticed that there are 3 duplicate epd's

1q2r1k1/1b2bpp1/p2ppn1p/2p5/P3PP1B/2PB1RP1/2P1Q2P/2KR4 b - - c0 2; c1 line(s): 605 1405;
Right, one is from a/b/c pawn advancement, and the other is undermine, but have different move scoring. Could be dual purpose.

1r1r1bk1/1bq2p1p/pn2p1p1/2p1P3/5P2/P1NBB3/1P3QPP/R2R2K1 b - - c0 2; c1 line(s): 208 804;

r1r3k1/pb3p1p/1pqBp1p1/4P3/3b4/2P2P2/PR1N2PP/2RQ3K w - - c0 2; c1 line(s): 277 873;
Surely that makes the position useless then, because improving the score for one benchmark will decrease it for the other...?
I don't see why it is useless. If you are good in pawn advancement theme you will get the bm, and if you are good in undermining theme too you will also get the bm.
Because if you make an engine who play the very best moves in all positions, it won't get the maximum of points in all test, that's crazy ...
Pos 1 = 10, pos 2 = 10,
total points = 20, max points = 20, why it will not get max points?
OK, no problem so ! ;-)
[d]1q2r1k1/1b2bpp1/p2ppn1p/2p5/P3PP1B/2PB1RP1/2P1Q2P/2KR4 b - -
There is somewhat an interesting point with this same position and same bm, in two different themes, but not about finding the bm but by not finding the bm, if the engine can not find the c4 bm, that would mean the engine will not get the best points in both test themes. But if the engine is able to find the bm it gets 20 points, 10 points for pawn advancement, and 10 points for undermining themes. So more points for engines that get the bm, implying that this position is important according to the designers of the test suite. As a consolation there are varied points on alternative moves that the engine may still gain. The alternative moves and corresponding points are not exactly similar for both test themes.
Carlos Ylich
Posts: 175
Joined: Wed Apr 28, 2010 9:31 pm
Location: Brazil

Re: STS 1-15 epd and analysis tool

Post by Carlos Ylich »

Hello,

I'm having problems with STS 3 an 4 in Arena 3.5 :roll:
Remember Sabra and Chatila
Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: STS 1-15 epd and analysis tool

Post by Ferdy »

Carlos Ylich wrote:Hello,

I'm having problems with STS 3 an 4 in Arena 3.5 :roll:
Can you describe the problems?
Carlos Ylich
Posts: 175
Joined: Wed Apr 28, 2010 9:31 pm
Location: Brazil

Re: STS 1-15 epd and analysis tool

Post by Carlos Ylich »

Hello,
My English is very bad, but here we go:
Stockfish 6 lock these two tests 3 and 4. I have to restart Arena after this.
Houdini and Komodo 9 work well.
Remember Sabra and Chatila
Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: STS 1-15 epd and analysis tool

Post by Ferdy »

Carlos Ylich wrote:Hello,
My English is very bad, but here we go:
Stockfish 6 lock these two tests 3 and 4. I have to restart Arena after this.
Houdini and Komodo 9 work well.
I tried sts3 on sf6 at 1s/pos and it worked just fine. I have not tried sts4 yet. This is under arena 3.5.
Carlos Ylich
Posts: 175
Joined: Wed Apr 28, 2010 9:31 pm
Location: Brazil

Re: STS 1-15 epd and analysis tool

Post by Carlos Ylich »

Ferdy wrote:
Carlos Ylich wrote:Hello,
My English is very bad, but here we go:
Stockfish 6 lock these two tests 3 and 4. I have to restart Arena after this.
Houdini and Komodo 9 work well.
I tried sts3 on sf6 at 1s/pos and it worked just fine. I have not tried sts4 yet. This is under arena 3.5.
I tried sts 3 and 4 at 10s/pos in position 30 crash. Stockfish 6 and 18052015
Remember Sabra and Chatila
Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: STS 1-15 epd and analysis tool

Post by Ferdy »

Carlos Ylich wrote:
Ferdy wrote:
Carlos Ylich wrote:Hello,
My English is very bad, but here we go:
Stockfish 6 lock these two tests 3 and 4. I have to restart Arena after this.
Houdini and Komodo 9 work well.
I tried sts3 on sf6 at 1s/pos and it worked just fine. I have not tried sts4 yet. This is under arena 3.5.
I tried sts 3 and 4 at 10s/pos in position 30 crash. Stockfish 6 and 18052015
If you enable the arena debug, that problem can be easily identified.

I tried 10s/pos now on sts3, no problem so far after 45th position.
Ferdy
Posts: 4833
Joined: Sun Aug 10, 2008 3:15 pm
Location: Philippines

Re: STS 1-15 epd and analysis tool

Post by Ferdy »

Carlos Ylich wrote:
Ferdy wrote:
Carlos Ylich wrote:Hello,
My English is very bad, but here we go:
Stockfish 6 lock these two tests 3 and 4. I have to restart Arena after this.
Houdini and Komodo 9 work well.
I tried sts3 on sf6 at 1s/pos and it worked just fine. I have not tried sts4 yet. This is under arena 3.5.
I tried sts 3 and 4 at 10s/pos in position 30 crash. Stockfish 6 and 18052015
STS3 test completed at 10s/pos on sf6 - no problem here.