Winboard Forum

New version is available at http://greko.chess.googlepages.com
Source code refactoring, bugfixes, small optimizations.

Hi Vladimir,

for rating purposes, should this be treated as a new engine or should it be combined with 5.7.1?

Regards, Graham.

I do not expect big changes in playing style/strength. So, the decision is completely up to you

Vladimir Medvedev wrote:I do not expect big changes in playing style/strength. So, the decision is completely up to you

Only reason I ask is that we have 63 games for GreKo 5.7.1 that will be "killed" from our list if we don't combine them with GreKo 5.8.
It makes sense to combine them if the only changes to the engine are cosmetic.
However, if you've made code changes that make the engine play differently and with strength change, then we should really treat them as different engines.
My gut feeling is that we should treat them as different engines.

In my fast test GreKo 5.8 scored 60/120 and got rating 2436. Very close to 5.7.1 (=2425, 61.5/120), which means nothing was seriously broken during refactoring.
I think, you can combine the results.

Vladimir Medvedev wrote:In my fast test GreKo 5.8 scored 60/120 and got rating 2436. Very close to 5.7.1 (=2425, 61.5/120), which means nothing was seriously broken during refactoring.
I think, you can combine the results.

Okay - thanks.