Q for Dann and Swami

Discussion of chess software programming and technical issues.

Moderators: hgm, Harvey Williamson, bob

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
User avatar
opraus
Posts: 166
Joined: Wed Mar 08, 2006 8:49 pm
Location: S. New Jersey, USA
Contact:

Q for Dann and Swami

Post by opraus » Fri Oct 22, 2010 2:18 pm

Would it be _VERY_ easy to reproduce the STS series, but with the best moves replaced, by those found, after only, say, 0.01 second analysis ?

I wonder, if tuning a much weaker engine, to find those same moves in similar time, would improve the engine?

Thoughts? Has this been tried before?

Thanks.

Dann Corbit
Posts: 9894
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Q for Dann and Swami

Post by Dann Corbit » Fri Oct 22, 2010 5:59 pm

opraus wrote:Would it be _VERY_ easy to reproduce the STS series, but with the best moves replaced, by those found, after only, say, 0.01 second analysis ?

I wonder, if tuning a much weaker engine, to find those same moves in similar time, would improve the engine?

Thoughts? Has this been tried before?

Thanks.
Different engines would find different answers. (E.g. Rybka's 0.01 second aswers would not agree with Stockfish nor Zappa, etc.).

I don't think it is feasible.

User avatar
opraus
Posts: 166
Joined: Wed Mar 08, 2006 8:49 pm
Location: S. New Jersey, USA
Contact:

Re: Q for Dann and Swami

Post by opraus » Fri Oct 22, 2010 6:37 pm

Hi Dann,

Thanks for responding.

I am thinking it does not matter if they agree.

It seems (to me at least) that if I tune my engine to 'like' the same moves as Rybka, Zappa, Stockfish, it is likely to return similar moves at deeper searches. Seeing that the deeper searches are informed by the shallow ones.

None of the moves have to be 'right' - In fact, it seems one could use a random set of positions where we had no idea what the best move was.

I am just interest to see if my engines 'thinks like' theirs does.
Returning the same move as any one of them would be better than returning an all-together different move most of the time.

These hyper fast searches CANNOT be just noise, as I said earlier, the deeper search is informed by these shallow ones. Right?

...

Dann Corbit
Posts: 9894
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Q for Dann and Swami

Post by Dann Corbit » Fri Oct 22, 2010 6:48 pm

opraus wrote:Hi Dann,

Thanks for responding.

I am thinking it does not matter if they agree.

It seems (to me at least) that if I tune my engine to 'like' the same moves as Rybka, Zappa, Stockfish, it is likely to return similar moves at deeper searches. Seeing that the deeper searches are informed by the shallow ones.

None of the moves have to be 'right' - In fact, it seems one could use a random set of positions where we had no idea what the best move was.

I am just interest to see if my engines 'thinks like' theirs does.
Returning the same move as any one of them would be better than returning an all-together different move most of the time.

These hyper fast searches CANNOT be just noise, as I said earlier, the deeper search is informed by these shallow ones. Right?

...
The main problem is that these positions are not tactical but strategic in nature. Whether solved by search or eval, there is a right asnwer to these positions. A shallow search will definitely give a different answer: an inferior one. So which inferior answer do we choose?

User avatar
opraus
Posts: 166
Joined: Wed Mar 08, 2006 8:49 pm
Location: S. New Jersey, USA
Contact:

Re: Q for Dann and Swami

Post by opraus » Fri Oct 22, 2010 7:24 pm

Dann Corbit wrote: So which inferior answer do we choose?
The better ones that Rybka, Zappa, and Stockfish come up with, rather than the really stupid one that Xpdnt comes up with. :)

Thank you (and Swami), BTW for STS. They are very helpful indeed.

Dann Corbit
Posts: 9894
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Q for Dann and Swami

Post by Dann Corbit » Fri Oct 22, 2010 8:08 pm

opraus wrote:
Dann Corbit wrote: So which inferior answer do we choose?
The better ones that Rybka, Zappa, and Stockfish come up with, rather than the really stupid one that Xpdnt comes up with. :)

Thank you (and Swami), BTW for STS. They are very helpful indeed.
The alternatives are already listed and ranked. Are you looking for other answers besides these?

Antonio Torrecillas
Posts: 90
Joined: Sun Nov 02, 2008 3:43 pm
Location: Barcelona

Re: Q for Dann and Swami

Post by Antonio Torrecillas » Fri Oct 22, 2010 8:36 pm

An opposed approach, may be better for this purpose.
Rather than challenging the engine with dificult test position.

use a simple test suite, which will test key elements of the engine.

Once I constructed a test suite as follow:
I walked the file twic663.pgn to get all position reached.
on these set I discarded each position where fruit does not
give the same answer for depth=3 and depth=6
Then I ran the test at 10" with fruit, crafty, rybka 1.0,
spike 1.2 and other engines I do not remember now.
The final set of 6690 positions where these engines agree
was very useful to detect errors in quiesce, misevaluation etc.
at least for my weak engine.

This approach but with shorter time control will do your business.


if somebody is interested you can find twic663.epd at http://sites.google.com/site/barajandotrebejos/

regards,Antonio.

Mark
Posts: 215
Joined: Thu Mar 09, 2006 8:54 pm

Re: Q for Dann and Swami

Post by Mark » Sat Oct 23, 2010 2:22 am

Antonio Torrecillas wrote:An opposed approach, may be better for this purpose.
Rather than challenging the engine with dificult test position.

use a simple test suite, which will test key elements of the engine.

Once I constructed a test suite as follow:
I walked the file twic663.pgn to get all position reached.
on these set I discarded each position where fruit does not
give the same answer for depth=3 and depth=6
Then I ran the test at 10" with fruit, crafty, rybka 1.0,
spike 1.2 and other engines I do not remember now.
The final set of 6690 positions where these engines agree
was very useful to detect errors in quiesce, misevaluation etc.
at least for my weak engine.

This approach but with shorter time control will do your business.


if somebody is interested you can find twic663.epd at http://sites.google.com/site/barajandotrebejos/

regards,Antonio.
Thanks for the link. Is there any easy way to convert the format of the bm to algebraic notation? (like Kc4 instead of Kd3-c4)

Dann Corbit
Posts: 9894
Joined: Wed Mar 08, 2006 7:57 pm
Location: Redmond, WA USA
Contact:

Re: Q for Dann and Swami

Post by Dann Corbit » Sat Oct 23, 2010 3:55 am

Mark wrote:
Antonio Torrecillas wrote:An opposed approach, may be better for this purpose.
Rather than challenging the engine with dificult test position.

use a simple test suite, which will test key elements of the engine.

Once I constructed a test suite as follow:
I walked the file twic663.pgn to get all position reached.
on these set I discarded each position where fruit does not
give the same answer for depth=3 and depth=6
Then I ran the test at 10" with fruit, crafty, rybka 1.0,
spike 1.2 and other engines I do not remember now.
The final set of 6690 positions where these engines agree
was very useful to detect errors in quiesce, misevaluation etc.
at least for my weak engine.

This approach but with shorter time control will do your business.


if somebody is interested you can find twic663.epd at http://sites.google.com/site/barajandotrebejos/

regards,Antonio.
Thanks for the link. Is there any easy way to convert the format of the bm to algebraic notation? (like Kc4 instead of Kd3-c4)
I made a translation of it.
Here it is:
http://cap.connx.com/chess-engines/new- ... 3a.epd.bz2

User avatar
opraus
Posts: 166
Joined: Wed Mar 08, 2006 8:49 pm
Location: S. New Jersey, USA
Contact:

Re: Q for Dann and Swami

Post by opraus » Sat Oct 23, 2010 8:44 am

Thanks Antonio,

This may indeed be a better solution.

What percentage did your engine agree with these?

thanks!

Post Reply