On May 3, 1:56 am, Sanny wrote:

A week back Zebediah used to win in 20-30 moves but now the GetClubs
game was a bit improved now Zebediah takes 35 moves to win against the
Advance Level.

That was just one game; you can't draw sweeping
conclusions from the lengths of single games.


If you can spot any mistake in GetClubs game I can improve it further.

Please see the game and let me know which were the moves that were
wrong.

The GetClub program neglects normal piece development
in favor of marauding Queen moves. This is a hallmark of
most weak /human/ players as well.


Advance Level thinks for 30 min/ move So it is very difficult for me
to find mistakes in Advance level game. Anyone having Commercial
program please do the analysis and let me know which were the moves
which were wrong by GetClub that Zebediah managed to win in just 34
moves.

Rather than try to pinpoint specific moves, it may
help to simply point out that in this loss, just as in
many, many others, the program "forgot" to develop
its King-side pieces and castle. The result is that
much of the game if fought with just half its army.


Remember Advance level plays better than Rybka So you must analyze for
longer hours to give any good reasion.

Yes, the Advance level is probably around 3500--
3600 strength now. This is why the 3000-rated Rybka
program can only give it Knight or perhaps Rook
odds, and no more. With Queen odds I expect the
GetClub Advance level would /eventually/ prevail.



I do not understand how zebediah manages to win Advance in just 35-40
Moves?

Zeb is quite obviously using some other chess
program.


After analyzing the game with Computer have you found any mistake in
Advance Levels Game?

It doesn't require a computer to do that. One move
that was sub-optimal is 34. ...Qxc6. Generally
speaking, one should always avoid allowing one's
King to be captured for no reason.


Is it making Tactical Mistakes or it is doing Strategy Mistake. Please
Explain the mistake So that I really undersatand what went wrong.

Oh dear, that is a very tall order.

You need to coax the program into developing its
pieces and bringing its King to safety, instead of
chasing after spite-checks and wild tactical
skirmishes, the result of which is very unclear.

For instance, a while back you explained that the
reason it was losing via N-g5, Nxf7 suicide attacks
was that you gave a large bonus for spite-checks,
which of course was backfiring against decent
opponents (although it might work against weaker
ones).

I wouldn't worry too much about losing to Zeb's
chess program; for all we know, it could be one of
the strongest in the world, and even if the Rybka
program were "ported" to Java, you would still
lose on processing speed.


-- help bot