Sanny wrote:
I am now going to play a game against Rybka and see if the improvement
is useful or not and post the game below for you to see how good it
plays against Rybka.

ROFL. Rybka has already grabbed it by the goolies at move 13.
AFAICT Getclub offered its goolies to be grabbed two moves before

Here is the Game Rybka wins Easy Level in just 33 moves. Despite the
improvements suggested by you Rybka had the upper hand.

But Still GetClub handled the position for quite sometime.

Hardly - it was losing quite badly from move 13 still in the opening.

Do you see that the program now do not take out the Queen aimlessly.

What else can be done to improve GetClub?

I keep noticing that very often a key developing move like in this case
10. ... Be7 is never played by Getclub (or Be2 for white). It would have
at least have had a sporting chance with that line.

Perhaps you could tweak the evaluation to give a 20cp bonus to moves
which permit castling to be played next (and also a bonus for castling).
And also a bonus for moving minor pieces off their home squares.

AFAICT the least bad continuation after the disastrous

10. ... Nxe4

This could be a bug - a 6 ply search should reveal why this isn't a
very good idea. And most engines prefer Be7 at a 1 ply search.

11. a3 d5
12. a3xN c6??

What is it playing at here? c6 is absolutely awful. Must be a BUG!

12. ... Bxb4 is obviously the least bad move by far!
Even the simplest material evaluation function would spot that move!

Do you still see something wrong in GetClubs play other than Tactics?

See above. It will gift its N for a P when it could have had 2P and
freedom to castle into relative safety as compensation. This is still
badly losing, but a lot better than the daft line it played.

10. Bc2-b3{18} Nf6-e4{24}
11. a2-a3{16} d6-d5{22}
12. a3-b4{20} c7-c6{54}

Terminal decline sets in here. Rybka is already winning by miles.

Regards,
Martin Brown
** Posted from http://www.teranews.com **