On Apr 19, 9:54 am, Martin Brown <|||newspam...@[EMAIL PROTECTED]
>
wrote:
> > > White -- Black
> > > (easy) -- (chrisf)
>
> > > 1. e2-e4{0} e7-e5{2}
> > > 2. Ng1-f3{0} Nb8-c6{2}
> > > 3. Bf1-b5{0} a7-a6{2}
> > > 4. Bb5-a4{0} Ng8-f6{2}
> > > 5. Ke1-g1{48} Bf8-e7{4}
> > > 6. d2-d3{20} Ke8-g8{16}
>
> Although 7. BxN dxB 8. Nxe5 snaffles a pawn it isn't such a massive
> advantage. However, since it is a purely materialistic gain it is
> surprising that Getclub did not grab it with both hands.
After thinking about this a little more, I realized
that the Knight need not be snatched before
taking the pawn, since it is, after all, pinned to
the King. So, even 7. Nxe5 nets a free pawn. It
is a center pawn-- not some insignificant Rook-
pawn, ya know. (If there are any deeply-hidden
tactics here, just let me know and I'll fire up the
old 'puter to spot 'em.)
> > > 7. Nb1-c3{36} d7-d6{16}
>
> It is in the big 2006 database upto this point allowing for
> transpositions).
I hope that is not an endorsement of the hanging
of pawns. (A recent posting by Taylor Kingston
cited a game between two Class C players, as
taken from some random database.)
> > Look at the opening; after White plays d3, he is
> > threatening to win the e-pawn for free by Nxe5-- a
> > simple fact which both players somehow missed.
> > This is about as simple as tactics can get.
>
> Shredder doesn't think that grabbing the pawn is all that great.
Hmmph. If I were not already analyzing another
game, I would easily refute this hypocrite by
booting up my vastly superior program! (As it is,
I don't even have a chess board to look at.) For
the record, the Rybka program *never* gives full
credit for winning a pawn, because the opponent
invariably gains a half-open file-- or whatever. But
....a pawn is a pawn, in any country in the world.
And center pawns are the best-of-breed winners.
> > Funny thing is, after he allowed his Q and R
> > to get forked, he really had no choice but to
> > follow up-- or else he'd be a Rook down. The
>
> The fork may have been deliberate.
Deliberate? It was the GetClub program that
played the pawn-fork, not any "deliberating"
human chess player. As I recall, the first such
pawn-fork was not played by the GC program,
so for all we know its human opponent may
have figured it would be consistent (a huge
mistake with *this* program).
> Once the Q is into the kings
> castled position it is no holds barred. It was predictable that
> Getclub would take the knight making his king even more exposed. What
> is more surprising is that it still occassionally makes completely
> pointless pawn moves that are (or should be) way down the list of
> pluasible moves. 32. c4 is indicative of a problem in the defence and
> queiscence search logic.
Um, you talk as though the GetClub program
actually ranked every legal move in order by merit,
then selected the best one to play. The REALITY
is that it can't even determine which moves are
legal! It's a wonder it can play as well as it does.
> > GetClub program does not defend well, so if
> > you can get an attack on the King, as in this
> > game, you're set unless you run out of ammo.
>
> Its idea of King security is pretty feeble. I had forgotten from my
> analysis of last night. 20. Kh1 and then 22. Kg1 serve no useful
> purpose other than to enable black to build up his attack on the
> kingside unopposed. White would have been in much better shape had
> these moves been replaced with either 20. b4 to start a Qside counter
> attack or 20. Ne2 to prepare for the coming assault on the kingside.
> Random king moves is an inherent Getclub weakness.
Also consider the idiotic moves Re1, followed by
****fting wood back again with Rf1; such maneuvers
lose valuable time. I've had quite a few games in
which we are strolling along, I capture a pawn, and
then instead of recapturing -- like "normal" programs
and people do -- the GC program "develops" a
piece or attacks something. In sum, this is not
your everyday, run-of-the-mill chess program; there
is something rotten in Denmark.
I was amazed to find that, after his snide attitude
about the inconsistencies of his ratings system,
"Sanny" has now done an about-face and rigged
the ratings. Log on to GC's Web site now and
you will find that Advance level is "given" a 2400
rating, while Master level has already taken a
shellacking by Zeb, the computer operator. I
somehow missed that Normal level was at 1800,
and am playing the Advance level, to cash in on
the new "ratings bonanza" (if I can somehow win).
Inflation: it's not just an economic term anymore!
-- help bot
|
