Mobility Evaluation ?

Discussion of chess software programming and technical issues.

Moderators: hgm, Rebel, chrisw

User avatar
Evert
Posts: 2929
Joined: Sat Jan 22, 2011 12:42 am
Location: NL

Re: Mobility Evaluation ?

Post by Evert »

Well, king mobility can be folded into king safety (where you factor in crowdedness and attacks near the king).
Similarly, pawn mobility can be folded into pawn structure. A typical pawn can move, or it cannot. If it cannot, then either it is rammed (in which case the opponent has a similar rammed pawn, which cancels) or it is blocked. If it's blocked and that is relevant (for passers, maybe for candidates), it's included. If it can move and it's on a closed file, then there is an enemy pawn that is likewise able to move. Where it's relevant is when the pawn can undermine a rammed pawn, or is backward (and must advance to no longer be backward). Those are scored as part of pawn structure.

So on the whole, pawn mobility is so poor to begin with that it only becomes relevant in relation to other pawns or in specific situations, where it can be handled in a different way (as part of pawn structure). Having said that, I think Komodo does something with raw pawn mobility, but I don't remember exactly what. A search may turn up something.