Page 1 of 5

Re: Komodo Syzygy evals

Posted: Wed May 21, 2014 9:01 pm
by lkaufman
Quote from our Syzygy implementer Jeremy Bernstein:

"OK I do see the problem with the 6-man bases installed. This is the DTZ problem: the Syzygy bases are preferring a pawn move (resetting the DTZ clock) to a queen move, even if the queen move is mate! The solution to this would be to continue to search on the move list returned, but that might result in some performance drops. It would, in any case, need to be tested.

The "nonsense score" is a tablebase win, as explained in the documentation."

Re: Komodo Syzygy evals

Posted: Wed May 21, 2014 9:25 pm
by Hurnavich
I have sent you a private message to respect the forum rules

Stockfish and Houdini appear to have No problem with 6 man same situation

so must be the way K implements its use no?

Re: Komodo Syzygy evals

Posted: Wed May 21, 2014 9:50 pm
by lkaufman
Hurnavich wrote:I have sent you a private message to respect the forum rules

Stockfish and Houdini appear to have No problem with 6 man same situation

so must be the way K implements its use no?
Yes, as Jeremy explained there is a fix but it would have to be tested.

Re: Komodo Syzygy evals

Posted: Wed May 21, 2014 11:51 pm
by bob
Hurnavich wrote:I have sent you a private message to respect the forum rules

Stockfish and Houdini appear to have No problem with 6 man same situation

so must be the way K implements its use no?
What "forum rule" do you think Larry broke? He responded DIRECTLY to your question with an answer from the person that implemented the table base code for them. Seemed polite, accurate, and a direct response to your post... Which you seemed to miss as you asked the SAME question again after he had already answered it precisely.

Re: Komodo Syzygy evals

Posted: Thu May 22, 2014 12:40 am
by Hurnavich
What you going on about? I never meant any rules where broken I see I may of worded it wrongly my apologies.

Just in a previous post by someone the following was wrote


"There was a problem with the discount. It's working now.

Please PM Mark or Larry or me, in that order, with any technical web problems. We don't want to use the forum as the komodochess help desk, and surely the moderators feel the same way."

So I was just trying to say I would switch to private message to avoid making too many more posts.
And was trying to abide by the above thoughts

Re: Komodo Syzygy evals

Posted: Thu May 22, 2014 2:01 am
by shrapnel
As with any hastily released product.....BUGS :roll:
This is just one bug that Hurnavich has discovered....God knows how many more are there..will wait till all Komodo bugs are fixed before getting it.

Re: Komodo Syzygy evals

Posted: Thu May 22, 2014 9:08 am
by mjlef
The problem is the GUI. A few GUIs take the score Komodo gives them and for some reason, improperly converts them to a negative mate score to display. Scores above 29999 appear in these bad GUIs as negative mate score. Komodo still plays a move leading to a win. One GUI (Arena) does it backward, converting a mate in 1000 to a score!

Since these GUIs are not displaying the scores properly, it is a GUI problem, but we are looking at modifying Komodo to display a score around 250.00 to indicate an upcoming mate. But we want to test it in a bunch of GUIs first, since the GUIs implement things differently.

May GUIs have a log function where you can see the commends set to and from the engines to confirm this.

Re: Komodo Syzygy evals

Posted: Thu May 22, 2014 9:30 am
by Hurnavich
Hi,

will the following still be looked into?


Quote from our Syzygy implementer Jeremy Bernstein:

"OK I do see the problem with the 6-man bases installed. This is the DTZ problem: the Syzygy bases are preferring a pawn move (resetting the DTZ clock) to a queen move, even if the queen move is mate! The solution to this would be to continue to search on the move list returned, but that might result in some performance drops. It would, in any case, need to be tested.

Many thanks

Re: Komodo Syzygy evals

Posted: Thu May 22, 2014 9:44 am
by mjlef
We might have Komodo search some more, and pick the mating move, if found, over a TB winning move. Neither is "wrong". Like Ron said "different implementation". We appreciate feedback from customers.

Re: Komodo Syzygy evals

Posted: Thu May 22, 2014 9:57 am
by Hurnavich
I would like to see Komodo search some more, and pick the mating move, if found, over a TB winning move.

This would then work like Stockfish and Houdini and if modifying Komodo to display a score around 250.00 to indicate an upcoming mate. works then i will be a happy customer.

there are others at another forum i'm sure would agree.

I look forward to the above to test if implemented.