Fail high codes

Discussion of chess software programming and technical issues.

Moderator: Ras

Martin Brown
Posts: 46
Joined: Sun Oct 18, 2009 12:07 pm

Fail high codes

Post by Martin Brown »

In the commercial engines under complex positions near to mate or a major material promotion it is fairly common to get artificial scores back in the early stages of a search that failed high. Is there a standard interpretation of these magic code numbers. eg in Chessbase engines

279.99 +/- about 0.3

And then when mate seems inevitable but not resolved for depth
#150 and seq to about #159
#174 and seq to #178
and I think I have seen the odd #200 and #225

Is there a standard for this and does anyone know how to interpet them?

Thanks for any enlightenment.
Martin Brown
zamar
Posts: 613
Joined: Sun Jan 18, 2009 7:03 am

Re: Fail high codes

Post by zamar »

Well, there definitely isn't any standard for those strange values!
Joona Kiiski
Aleks Peshkov
Posts: 928
Joined: Sun Nov 19, 2006 9:16 pm
Location: Russia
Full name: Aleks Peshkov

Re: Fail high codes

Post by Aleks Peshkov »

EPD standard defines centipawn evaluation numbers and mate-distance values. I know at least Arena GUI does recognize it (in UCI mode).