Computer Chess Club Archives


Search

Terms

Messages

Subject: CCT5 - Crafty perspective

Author: Robert Hyatt

Date: 17:44:21 01/20/03


Round 1.  Crafty vs Qalat

The game went 19 moves in book and when we dropped out, it was almost
perfectly equal.  Crafty's first search was 13 plies deep and the eval was
+.06...  It averaged searching 14 plies deep for the next 10 moves or so
and by move 28 the score was -.27, not a very good trend.  By move 32,
a 17 ply search pegged the score back to +.13 and things were moving in
the right direction.  At move 35, black let the game get away by playing
34. b5 which lets Crafty break things open and take advantage of the
openness...  Score was +.5 here after an 18 ply search.  At move 35,
crafty expected Bxb5 with the +.5 score, but black played axb5 and my
score went to +1.82 instantly and by the time it finished 18 plies the score
had stabilized at +1.25...  Another couple of inexact moves by black turned
this into a lost ending.  Altogether a bit of a worry as the win was more due
to the opponent making a couple of mistakes, rather than great play by Crafty.

Round 2.  Hossa vs Crafty

Crafty popped out of book at move 8 with a score of -.5 (-=good for black).
The eval climbed about .1 every move for the next 10 moves, and at move 18,
crafty uncorked a king-side attack that worked out well.  18. ... f5 saw the
score climb to -3.19 after 12 plies.  In the moves prior to this, it had slowly
built up a significant number of attackers and move 18 saw the beginning of
an explosion.  After white's 19. Rg1 move, fxe4 led to a score of -6.0
after 13 plies.  At move 26 Crafty announced a mate in 9 and Hossa resigned
in a hopeless position.  A nice king-side attack that was the start of a
trend for the next three games..

Round 3.  Crafty vs Ruffian

I had personally thought that Ruffian would be the program to beat in this
event, as it seems to be very tactical, although it doesn't seem to be a
"strategic" thinker.  The game went 13 moves in book, and then I watched a
repeat of the previous game, where Crafty slowly built up a "crowd" on the
kingside.  By move 18 the score was +.5 with the move Be3.  And things went
bad for black beyond that point.  By move 22 the score was +1.0, based only
on positional considerations on the king-side.  At move 29, the score had
climbed to +1.5, the average search depth for the previous 10 moves was 14
plies.  At move 31, Ruffian saw trouble and thought for 5 minutes but could
not find a better move than that played in the game.  After 17 plies, Crafty
was sitting at +2.08 as it played move 32. Qf6.  By move 38 the score was
approaching 3 with a fail-high on Rb7.  The rest was anti-climatic as the
kingside attack led to a simply won ending.  It was not as easy as it looked,
from my perspective.  :)

Round 4.  Crafty vs Yace

Another primary contender.  Due to a bit of botched seeding, Crafty ended up
seeded lower than it really should have been (seeding was on an old ICC
standard rating, rather than the current standard rating) crafty ended up with
white two games in a row.  Who am I to complain?   :)

Same opening, but yace chose a different path after a few moves and by the
time Crafty dropped out of book at move 13, the score was -.52.  By move 15
the score was back to "even" and the kingside build-up started again.  By move
21 the score was +.40.  For the next 20+ moves, nothing happened.  A lot of
shuffling, posturing and re-location of pieces by both sides.  Suddenly,
around move 40, things started moving up for crafty and at move 44, Nh4 had a
score of > 1.23 (a fail high on that beta value).

At move 48, the wildest event of the tournament happened.  Crafty was looking
at Bc7 and after depth=14, the score was +3.34.  It kibitzed this line, but as
we watched, it played Ne6+ instead.  I was afraid that a bug had just turned a
win into a loss.  I looked at the log file, no mention of Ne6 until it actually
played the move (ie no fail high).  It turned out to be ok.  After the game I
went back and sure enough, Ne6 was the best move at depth=14.  After looking
at the source, the move Ne6 apparently became best right at "time out".  And
while it saved the best move, after time runs out it doesn't print anything
else, which made it look strange.  It turns out Ne6 is only slightly better
than Bc7, but when you don't see any analysis, ugh.

The game was basically over at this point as crafty's score was +4 and
climbing.  Yace resigned at move 55.  Three good king-side attacks in a row.
Another is coming, but for the wrong side.  Read on.

Round 5  Ferret vs Crafty

Another tough opponent.  19 moves in book, first search score was +.24,
not bad.  At move 20, crafty played h6 to drive Ferret's bishop back to e3
it thought, but Ferret played Bh4 inviting g5 trapping and winning the bishop.
Ferret then traded a piece for the g/h pawns and started an attack.  Crafty's
score here was right at zero, which was alarming, because it was a piece for
two pawns up (+1.0 advantage) and the "bad trade" code added another 1.5 to
that since trading a piece for pawns is bad.  2.5 pawns of score, yet the
kingside safety was dragging that back down to zero.

It turned out the attack was sound, although it wasn't obvious to me that it
was, at the time.  And I can certainly forsee situations where it would fail
miserably.  But here it worked, Ferret played quite accurately, and by the
end of the game at move 57, it had played very well.

End of the first day.

four good games, plus one that was hard to fault other than the result.  Even
in the loss, Crafty played very accurately and made the win very difficult to
hit.

Round 6  Crafty vs Searcher

A near disaster for the first game of the second day.  The same d4 opening
led to a similar position, but things did not go very well here.  First score
out of book was -.42, which was typical for every 1. d4 game crafty played as
white. But it was able to pull that up quickly normally.  10 moves out of book,
the score hadn't changed, showing that searcher was playing very well and with
a reasonable amount of understanding of the position.

Finally by move 24, Crafty was back to a slightly + score, and this held until
it started dropping as it misjudged the queen/rook attacking in the center.  At
move 32, the score was -.68 after 16 plies.  at move 35, the score was -1.5, at
move 40 -2.0, -2.5 at move 50, -3 at move 60,  and at this point Crafty dug
in its heels and pulled the score back to -2.3 where it stayed for a long
while.  But it slowly traded pawns, and the score started swinting back.  By
move 80, it was -2.0 again, -1.5 by move 85, -1.0 by move 95,  and it finally
reported a draw score at move 102.  Of all the games it played, this was a
really nice effort as it showed a lot of understanding about king rook and pawn
endings, something I have worked on a lot over the years.

The game ended at move 115 with a repetition.

Round 7  quark vs Crafty

18 book moves, score 0 on leaving book, not a bad start as black.  A tactical
oversight saw the score drop to +.70 (good for white) at move 21, but this
pawn "sac" was quickly recovered positionally and by move 24 the score was -.10
and I was breathing easier.  By move 30 the score was -.7.  By move 65 the
score was nearly equal, and it looked to be headed to another draw.  However,
at move 65 quark sacrificed a bishop for crafty's last pawn, leaving it in a
KRB (crafty) vs KNPPP (quark).  The bishop/rook quickly ganged up on the
pawns and won them one by one, leading to a krb vs kn endgame table loss for
quark.  A good game by quark, and the bishop for pawn looked like a draw, but
with a lot of maneuvering, the two pieces were simply overwhelming.

Round 8  Pepito vs Crafty (two blacks in a row to offset two whites earlier)

I didn't know much about pepito, so I simply watched the game.  Out of book at
move 7 had a score of -.37 so this seemed promising.  Around move 25 the score
started dropping and pepito won a pawn. As in round 6, crafty dug in and by move
63 this was a drawscore game as well.  Nothing good or bad to say here, it was
just "a game".

Round 9  Crafty vs Tao

A Ruy Lopez that endedthat ended after 11 moves, and after Crafty played
Bh6, Bxh6, Qxh6 and then Qg7 the game instantly turned into a K, two rooks
and two minors for each side ending, with 3 pawns on each wing..  My score
was dead zero here and it really looked like a draw.

However, around move 24, black moved his rook off the e-file and that
gave crafty a window of opportunity to penetrate, and it did so with a
score of +.5 at move 24.  By move 33, the score was solidly at +.5 and
by move 50 it was clear that white was going to infiltrate and eat the
queenside, which it did...


All in all an interesting experience.

Again, as always, the book is critical, and I did _zero_ book preparation
which hurt in several games.  Using an automatically produced book can
work, I am convinced, but it needs to include recent games, so that old
lines with refutations are not played.

The next tournament will find me with a better book.  :)




This page took 0.17 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.