Author: Geoff
Date: 02:54:22 05/03/03
Go up one level in this thread
Hi Uri
>I guess that there is another reason.
>
>500->350 is not a big reduction and mobility is important factor to evaluate.
>
>Maybe there is a bug in your evaluation or your mobility scores are too high(I
>do not know th way that you evaluate mobility).
Bugs in my code !! no never, impossible ;-)
Seriously, I think the code is OK, maybe just the idea is not that sound ?
I look at the number of pseudo legalmoves for both sides, but ignoring pawns and
kings, then subtact one from the other and then scale the result by 0.1 pawns
per square.
>For me killer move helped significantly after history tables.
I have tested just retested my killer move mod, I took out the PV and the
history sorting and retested the 2 killer moves only ordering. For the start
position it reduced the nodes searched to 93% of that with no move ordering.
That was enough to make me think my code was working as I meant it, but only a
small improvement so it has made me think I have not understood something
properly in the way I implemented it?
>>Any advice as to what further mods I could do to gain the most improvement ?
>>Simple mods preferred first ;-)
>
>The most simple change is to increase the value of knight and bishop from 3
>pawns to 3.3 pawns to prevent tscp to play bad trades of BN vs RP.
Now that is what I call a simple mod, I will give this a try and see what effect
it has. TSCP works great as a Winboard engine under CM9000, so I will run a
tourney to see what difference it makes.
Thanks for the tips
Regards Geoff
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.