View Single Post
  #23  
Old October 15th 06, 10:51 AM posted to rec.games.chess.analysis,rec.games.chess.misc,rec.games.chess.computer
Bjoern
external usenet poster
 
Posts: 156
Default Why it is Strategically blind?

"Why it is Strategically blind?" - Well, I guess because it was
programmed that way.

Sanny wrote:
Strategically blind,
while the tactical skills were hardly impressive either.
.



Could You emphasis in detail what you mean by Strategically blind,
(What type of Strategy you are looking for?)


Any kind of strategy? Having a plan? Putting pieces on squares where
they are active and cooperate effectively with the rest of the pieces?

And what is the meaning of tactical skills?


Ability to calculate accurately. For a start not overlooking mates would
not be bad. Similarly not constantly loosing pieces.

If you could elaborate with Example that will be helpful in improving
the game.


Not sure whether this is helpful in improving the game. If there were
easy answers such as "always do X in situations resembling situation Y"
and if there weren't always caveats such as "unless you have already
done Z, or unless the opponent could reply with..." etc. As far as I can
tell your program has 2 problems:

1. It isn't very good tactically and overlooks simple tactical problems,
this could be fixed by programming more efficiently so that the program
searches to a greater depth in the same amount of time (however I guess
it is unlikely that your programmers will manage that, so forget that
for the time being).

2. Your positional evaluation function isn't good enough. It seems like
you have a couple of things sorted (like developing pieces), but not in
a way that reflects a true understanding of positions, as shown by how
the program handles the positions below.

Bjoern - Sanny's Chess Program (Master level)

1.d4 d5 2.c4 Nc6 3.cxd5 Qxd5 4.e3 e5 5.Nc3 Bb4 6.Bd2 Bxc3 7.bxc3 Nf6
8.f3 So far the opening by black was very much okay, while he took ages
for each move I wonder whether this was not the opening book after all,
because what follows is so much less accurate that one has to be
surprised by how quickly the computer manages to ruin the black position
following its perfectly reasonable first 7 moves. 8...Bd7 The computer
played this move without a plan, various moves such as 0-0, Qd6, exd4 of
e5-e4 all make sense, but probably the computer was programmed to make
sure it develops all its pieces?! 9.Ne2 Qd6 10.Ng3 Be6 As it turns out
8...Bd7 was a completely wasted move now. 11.Bd3 g5 Is black trying to
attack on the kingside? As it turns out, no. But unless that is the idea
11...g5 is just a weakening of black's position. 12.0-0 a6 In itself not
necessarily wrong, but in combination with 11...g5 this just looks
silly. 13.Nf5 Bxf5 14.Bxf5 exd4 releases the tension in the centre, not
necessarily totally wrong, but also played for no particular reason. Why
not just castle short? Oh, I forgot: the kingside is already weakened by
g7-g5! 15.cxd4 Ne7 16.Bc2 b5 17.Qe1 Nc6 18.Bc3!? 0-0 White is planning
to set up a big centre, but black just ignores that. Maybe Nd5 was worth
a try. In fact 18...0-0 is very dangerous, but I didn't spot that one
could punish black quite severly by playing 19.f4 in response (which is
so good because of the weakening move g7-g5). 19.Rd1 Rfe8 20.e4 Rad8 I
assume the computer has been following its programming: get all pieces
out and put the rooks on semi-open files. As this was done without any
regards to the requirements of the position white is at least better if
not winning. 21.Qf2 Qa3 22.Bb3 Rd7 Black now entangles himself and white
can now win material by force. What is the point of that move anyway?
Doubling rooks on the d-file? Surely that is totally ineffective as
white can just play d4-d5. 23.d5 Ne5 24.Qg3 Most of the defensive moves
by black over the next few moves are suboptimal, but white is winning
anyway. 24...Qd6 25.Qxg5+ Kh8 Not only is black a pawn down, the
computer has also placed his king on the same diagonal as the dangerous
bishop on c3. If the computer understood anything about chess he would
have resigned here or at most 5 moves later. 26.f4 Qb6+ 27.Bd4 Nxe4
28.Qf5 Qf6 29.Qxf6+ Nxf6 30.fxe5 c5 31.Ba1 c4 32.e6 fxe6 33.Bxf6+ Kg8
34.Bc2 exd5 35.Rf3 h6 36.Bg6 Re2 37.Rg3 Kf8 Black desperately wants to
be mated as quickly as possible. 38.Rf1 b4 39.Bh4+ Kg8 40.Bh5+ Kh7 Once
more preferring a quick mate to the "better" move Rg7. 41.Bxe2 a5 42.Bg4
Rb7 43.Bf5+ Kh8 44.Bf6+ Rg7 1-0



Bjoern - Sanny's Chess Program (Beginner level)

1.d4 d5 2.c4 Nc6 3.cxd5 Qxd5 4.e3 e5 5.Nc3 Bb4 6.Bd2 Bxc3 7.bxc3 Bf5
This has never been played before, I guess there is a reason for that.
8.f3 exd4 9.cxd4 Nf6 10.e4 Unless black is willing to sacrifice on e4,
the computers play up from move 7 onwards doesn't make any sense, surely
Nxe4 or maybe Bxe4 are pretty much forced? Following 10...Qe6 white has
a position that is so much better that the game is nearly over now.
10...Qe6 11.Bb5 a6 What is the point of that move? Forcing me to destroy
black's pawn structure surely makes no sense at all. 12.Bxc6+ bxc6
13.Ne2 Bg6 14.Bb4 That game is obviously over now, but the computer
doesn't know when to resign. Black's pawn structure is ruined, his
bishop on g6 is hardly more than a glorified pawn and his king is stuck
in the centre for now. To be able to castle black would have to do
something like sacrificing the a-pawn to distract the bishop from the
a3-f8 diagonal, but the computer is incapable of comprehending that this
is necessary and prefers to "keep" its pawn (for the time being).
14...Qc4 15.Qd2 Rd8 16.Rc1 Qb5 17.0-0 Nd7 18.Rfd1 f6 19.a3 Ne5 20.Nc3
Qb8 21.Na4 Bf7 For the first time in the game black has done something
sensible, he has put his bishop on a better diagonal, but otherwise he
has been maneuvering about aimlessly without any clear goal. Not sure
what more there is to say about it. 22.Qc3 Nc4 23.Nc5 Qb6 24.Nb3 Qb5
25.Na5 Nxa5 26.Bxa5 Rd7 The computer should just have castled and
accepted the loss of a pawn or so. 27.Bb4 a5 28.Bxa5 h6 What sort of
justification for a pawn sacrifice like 27...a5 is that? Previously
black always refused to give up a pawn when it might have helped with
saving the king from the centre, but now without any idea at all black
has sacrifice a pawn after all and now plays 28...h6? Unbelievably
stupid. 29.Bb4 Of course! Now black does some more idiotic shuffling
about of pieces just to make sure the bishop gets shut back in
(remember, the one good thing black did do all game, getting the bishop
on the better a2-g8 diagonal becomes worthless now). 29...Be6 30.Qxc6
Qb6 31.Qxb6 cxb6 32.d5 Bf7 33.Rc6 b5 34.Rb6 Rd8 35.Rc1 Obviously black
is "more lost" than ever, but the computer still playing on. 35...f5
Tactical suicide, but does it make much of a difference? 36.Rc7 fxe4
Overlooks a mate, but I didn't spot that either and just preferred to
win a rook. 37.Re7+ Kf8 38.Rd7+ Ke8 39.Rxd8+ Kxd8 40.Rb8+ Kd7 41.Rxh8
Black has suffered the consequences of not castling. It's hardly worth
it to comment on the following moves. 41...Bxd5 42.f4 g5 43.fxg5 hxg5
44.Rh5 e3 45.Rxg5 Kc6 46.h4 e2 47.h5 Bc4 48.h6 Bd3 49.Rg7 Kb6 50.g4 e1B
51.Bxe1 Kc5 52.g5 b4 53.Bxb4+ Kd5 54.g6 Ke6 55.Bc3 Kd5 56.h7 Kc4 57.Bb4
Bxg6 58.h8Q Bf5 59.Rc7+ Kd5 60.Qd8+ Ke4 61.Re7+ Kf3 62.Qf8 Kg4 63.Bd2
Bb1 64.Rg7+ Bg6 1-0
Ads
 

Mobile Phone - Problem Mortgage - Free Advertising - Mortgage Calculator - Personal Loans