GM and Rybka vs. Stockfish

Discussion of anything and everything relating to chess playing software and machines.

Moderators: bob, hgm, Harvey Williamson

Forum rules
This textbox is used to restore diagrams posted with the [d] tag before the upgrade.
rcmaddox
Posts: 173
Joined: Mon Oct 14, 2013 6:59 pm
Location: Winder, GA

GM and Rybka vs. Stockfish

Post by rcmaddox » Sat Aug 09, 2014 5:07 am


overlord
Posts: 198
Joined: Sun Jun 03, 2012 4:46 pm
Location: Trinec, Czech Republic

Re: GM and Rybka vs. Stockfish

Post by overlord » Sat Aug 09, 2014 5:18 am

I would win against Stockfish 5 with Rybka 4. People absolutly don´t understand that you don´t have to be GM...you have to know engine weaknessess and have experiences with analysis!!!

User avatar
AdminX
Posts: 5160
Joined: Mon Mar 13, 2006 1:34 pm
Location: Acworth, GA
Contact:

Re: GM and Rybka vs. Stockfish

Post by AdminX » Sat Aug 09, 2014 8:36 am

rcmaddox wrote:Interesting read at Chess.com:

http://www.chess.com/article/view/how-r ... -the-world
Thanks Robert, that was a very nice read.
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers

User avatar
pocopito
Posts: 236
Joined: Tue Jul 12, 2011 11:31 am
Contact:

Re: GM and Rybka vs. Stockfish

Post by pocopito » Sat Aug 09, 2014 10:09 am

It's important to note that Sotckfish with no opening book at all.
Two first meanings of the dutch word "leren":
1. leren [vc] (learn, larn, acquire) acquire or gain knowledge or skills.
2. leren [v] (teach, learn, instruct) impart skills or knowledge to.

User avatar
lucasart
Posts: 3041
Joined: Mon May 31, 2010 11:29 am
Full name: lucasart
Contact:

Re: GM and Rybka vs. Stockfish

Post by lucasart » Sat Aug 09, 2014 11:03 am

AdminX wrote:
rcmaddox wrote:Interesting read at Chess.com:

http://www.chess.com/article/view/how-r ... -the-world
Thanks Robert, that was a very nice read.
Indeed!
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.

Vinvin
Posts: 4378
Joined: Thu Mar 09, 2006 8:40 am
Full name: Vincent Lejeune

Re: GM and Rybka vs. Stockfish

Post by Vinvin » Sat Aug 09, 2014 12:07 pm

AdminX wrote:
rcmaddox wrote:Interesting read at Chess.com:

http://www.chess.com/article/view/how-r ... -the-world
Thanks Robert, that was a very nice read.
+1

User avatar
reflectionofpower
Posts: 1456
Joined: Fri Mar 01, 2013 4:28 pm
Location: USA

Re: GM and Rybka vs. Stockfish

Post by reflectionofpower » Sat Aug 09, 2014 12:39 pm

Of course you got throttled. Rybka 3 vs. Stockfish 5 but even Rybka 3 can still compensate for complex tactical situations and even with the GM Stockfish still won so the weakness is the human player AND the older Rybka.

I would also like to add that using Stockfish 5 is not necessarily the ultimate engine to use. It is tops in head to head with other engines but an engine that plays humans well is the ticket.
"hodie mihi, cras tibi"

Lonnie

Uri Blass
Posts: 8594
Joined: Wed Mar 08, 2006 11:37 pm
Location: Tel-Aviv Israel

Re: GM and Rybka vs. Stockfish

Post by Uri Blass » Sat Aug 09, 2014 1:23 pm

overlord wrote:I would win against Stockfish 5 with Rybka 4. People absolutly don´t understand that you don´t have to be GM...you have to know engine weaknessess and have experiences with analysis!!!
The GM used Rybka3 and not Rybka4 and I do not believe that you are going to win (at least if you do not know in advance the openings that stockfish is going to play)

Stockfish played with no book in the games so maybe it is possible to predict stockfish's openings based on knowing the time control and prepare something against it before the games but if stockfish play with a very small book not to get good positions but only to get you out of your preperations then I expect you not to win the match(and a small book can include uncommon lines like 1.e4 e5 2.Be2 for white or 1.e4 Nc6 for black).

User avatar
reflectionofpower
Posts: 1456
Joined: Fri Mar 01, 2013 4:28 pm
Location: USA

Re: GM and Rybka vs. Stockfish

Post by reflectionofpower » Sat Aug 09, 2014 1:29 pm

Uri Blass wrote:
overlord wrote:I would win against Stockfish 5 with Rybka 4. People absolutly don´t understand that you don´t have to be GM...you have to know engine weaknessess and have experiences with analysis!!!
The GM used Rybka3 and not Rybka4 and I do not believe that you are going to win (at least if you do not know in advance the openings that stockfish is going to play)
.
Kvicala Miroslav ... the name rang a bell or maybe it is a false one in my mind. I remember him posting games of him beating the top engines and he is 2000-2100 ELO strength and people did not believe him or something like that. Maybe he could clarify this.
"hodie mihi, cras tibi"

Lonnie

User avatar
AdminX
Posts: 5160
Joined: Mon Mar 13, 2006 1:34 pm
Location: Acworth, GA
Contact:

Re: GM and Rybka vs. Stockfish

Post by AdminX » Sat Aug 09, 2014 4:21 pm

[pgn]
[Event "Cyborg v. Machine "]
[Site "Hillsborough, CA "]
[Date "2014.07.19"]
[Round "1"]
[White "Naroditsky + Rybka 3"]
[Black "Stockfish 5 "]
[Result "0-1"]
[ECO "C05"]
[Plycount "84"]
[Sourcedate "2014.05.08"]

1.e4 e6 2.d4 d5 3.Nd2 c5 4.Ngf3 Nf6 5.e5 Nfd7 { By now, Stockfish had
spent around 10 minutes, but I was quite impressed that it managed to follow a
theoretical line. } 6.c3 Be7 7.Bd3 b6!? { An interesting variation that
become quite popular in recent years. Black delays the development of his
knight, and instead "threatens" to trade his French bishop with ...Ba6.
However, I actually faced this line against Spanish GM Julen Arizmendi in 2013,
so I was still moving quickly. } 8.Ne4!? { An enterprising continuation that
is supposed to give White a slight edge. The knight is obviously untouchable,
and depending on Black's reply, White will either penetrate to d6, practically
forcing the trade of Black's light-squared bishop, or retreat to g3, when the
threat of Nh5 could prove quite unpleasant. } 8...Ba6 { Practically a novelty,
although this move strikes me as very logical. Arizmendi essayed } ( 8...a5?! { , but this turns out to be cumbersome and unecessary. After } 9.Nd6+ Bxd6 10.exd6 Ba6 11.Bf4 Bxd3 12.Qxd3 Nc6 13.O-O c4 14.Qe2 b5 15.b3 { I was clearly
in the driver's seat. } ) ( 8...h6 { also does not seem to promise full equality.
For instance, } 9.Nd6+ Bxd6 10.exd6 O-O 11.Bf4 Ba6 12.Bxa6 Nxa6 13.O-O { and Black's 8th move has turned out to be a complete waste of time. } ) 9.Bxa6 Nxa6 10.Nd6+ { The root of my demise, although this move cannot be called an
error per se. It is important to note that Black is a tempo up compared to my
game against Arizmendi, and it will be a lot more difficult to hang on to the
d6 pawn. Instead, the prudent } ( 10.Ng3 O-O 11.O-O Rc8 12.Be3 { led to a
pleasant middlegame for White, although Black should have no problems
equalizing after } 12...Nab8 13.Nh5 Nc6 14.Qe2 Re8 { . Nevertheless, White's
position is very solid and the battle would have certainly have been less
one-sided. } ) 10...Bxd6 11.exd6 O-O 12.O-O Nf6 13.Bf4 { So far, I was
playing my moves very quickly, while Stockfish had consumed almost half of its
45 minutes. I was practically sure that it would go pawn-grabbing with the
dubious 13...Ne4, but its next move brought me back to reality. } 13...Rc8! { To be
completely frank, I cannot say for sure why this move is so strong. In my
opinion, it is more prophylaxis than anything else: Black does prepare the ...
Nb8-c6 maneuver, but the rook on c8 will also discourage White from playing
c3-c4. In fact, White must already tread carefully in order to keep the
balance. } ( 13...Ne4?! 14.Qa4 Nb8 15.c4! { and Black's position is already
on the verge of collapse. For instance, after } 15...g5 ( 15...Nxd6 16.dxc5 bxc5 17.cxd5 exd5 18.Rad1 Nc4 19.Qc2 Qf6 20.Rxd5! Qxf4 21.Ng5 f5 22.Ne6 { is
practically lost. } ) 16.Be3 Qxd6 17.cxd5 exd5 18.Nxg5 Nxg5 19.Bxg5 { Black's
position looks very suspect. } ) 14.Qe2?! { A very strange move that I still
cannot fully explain. Black's knight wants to leave a6, and this move clearly
plays into Black's hands. It looks like } ( 14.Qa4 { was the only way to
maintain equality, since the queen is much better placed on a4 than on e2: } 14...Nb8 15.Qxa7 Ne4 16.a4 ( 16.Qe7 Nc6 17.Qxd8 Rfxd8 18.a4 cxd4 19.Nxd4 Nxd6 20.Bxd6 Rxd6 { might be slightly better for Black, although White should not have
too many problems holding the draw after 21.Nb5. } ) 16...cxd4 17.Nxd4 Nxd6 18.a5 bxa5 19.Rfd1 Nc6 20.Nxc6 Rxc6 21.Qxa5 Qxa5 22.Rxa5 Nc4 23.Ra2 { and the
endgame is objectively about equal, although I would still prefer Black.
Nevertheless, had we reached this position, I would have actually had decent
chances of securing the draw. } ) 14...Nb8 { After Stockfish blitzed out this
natural move, I realized my (or should I say our?) mistake, and immediately
started to panic. Black's immediate threat is ...Ne4, but if he is allowed to
close up the queenside with ...c4 and win the d6, White will be in very
serious trouble. Instead of rectifying the situation by seeking immediate
central counterplay, though, I blindly followed Rybka's top suggestion: } 15.h3? { It is somehow ludicrous to call this the decisive mistake, but given the
circumstances, it cannot be called anything else. Losing a pawn for no
compensation against a 3400 engine is, in 99 cases out of a 100, an immediate
death sentence. White's position was already rather difficult, but with } ( 15.Rfe1 Nh5 ( 15...Ne4?! 16.c4 Nxd6 17.cxd5 exd5 18.dxc5 Rxc5 19.Qd3 { with
full compensation for the pawn. } ) 16.Be5 f6 17.d7 Nxd7 18.Bd6 Re8 19.h3 { I could have at least maintained a tenable position, with clear compensation
for the pawn. Of course, after } 19...Nb8 20.Bh2 g6 { Black is firmly in the
driver's seat, but at least it will not be as easy to win the game. } ) 15...c4! { The worst case scenario: Black has locked up the queenside, and White can
do nothing to save the d6 pawn. The game is indeed essentially over. } 16.Qc2 ( 16.Nd2 { did not help either, since Black's knight simply takes another route: } 16...Ne8 17.b3 cxb3 18.axb3 a5 19.Rac1 Nxd6 { with a similarly miserable position. } ) 16...Ne4 17.Rfe1 Nxd6 18.Ng5 { A somewhat desperado attempt to induce a
weakness, but White is up against a brick wall. } 18...g6 { The negligent } ( 18...Nf5? { would have allowed } 19.g4 h6 20.gxf5 hxg5 21.fxe6! { , when he is
right back in the game (the f4 bishop is untouchable due to 22.e7). After } 21...fxe6 22.Bxb8 Rxb8 23.Rxe6 g4 { White has the beautiful } 24.Rh6!! { , and Black
has nothing better than to allow a draw by perpetual. } ) 19.b4 { Perhaps the
computer's only weakness is its inability to recognize that locked positions
can sometimes be unwinnable (more about this later). I was still clinging to
the hope that I could lock up the queenside, overprotect the e5 square, and,
after stopping the g-pawn with h4, erect an impenetrable fortress. Of course,
this turns out to be nothing but fantasy. } 19...Qd7 20.a4 f6! { Another precise
move that would probably have taken a human 30 minutes to find and calculate
(if I recall correctly, it took Stockfish no more than two). The e6 pawn is
untouchable, so White's knight must ignominiously return to f3. } 21.Nf3 ( 21.Nxe6 Rfe8 22.Nc5 ( 22.Qe2 Nf5 { with the crushing ...Ng7 to follow. } ) 22...bxc5 23.dxc5 Nf5 24.Rxe8+ Rxe8 25.b5 d4 { and White's pawns are going
nowhere. } ) 21...Nc6 22.Re2 a6 { The computer's patience throughout the game
is also quite noteworthy. Before doing anything else, it ensures that White
will never have the ability to drive away the c6 knight with b4-b5. } 23.Rae1 Rce8 24.Bh6 Rf7 25.Bf4 { I saw nothing better than to move back and forth! } 25...Nf5 26.Qb2 ( 26.h4 { is almost always out of the question because it does more
harm than good. For example, after } 26...h5! ( 26...h6?? 27.Bxh6 Rh7 28.Bf4 { and the tables turn. } ) 27.Qb2 Nd8 28.a5 b5 29.g3 Rfe7 30.Kg2 Nf7 31.Qc2 Qc6 { Black will win by transferring his f7 knight to e4 (via d6) and following
up with the crushing ...e5. } ) 26...g5! { No hesitation whatsoever. The
alternative plan was to prepare ...e5, but Stockfish chooses to crush me on
the kingside first. } 27.Bh2 { I briefly had thoughts of "sacrificing" the
bishop with } ( 27.Bxg5 fxg5 28.Nxg5 Rf6 29.Nf3 { , but White's so-called
fortress is obviously nonexistant. } 29...b5 30.a5 Rg6 { followed by ...Qg7 is just
one way to penetrate. } ) 27...b5 { Once again, Stockfish combines decisive
action with patience. This move is probably not strictly necessary, but why
not prevent b4-b5 once and for all? } 28.a5? { Rybka was correctly suggesting
28.axb5, when White can at least try to develop some counterplay along the
open a-file, but by this point I had lost a good deal of faith in my assistant.
In 20/20 hindsight, this move indeed strikes me as ridiculously cooperative -
why deliberately close the queenside and allow Black to focus all of his
attention on the center and kingside?! } ( 28.axb5 axb5 29.Qc1 h5 30.Nd2 Rfe7 31.Qd1 { was the only way to put up any resistance, although I doubt that
Stockfish would have any real problems finishing us off after } 31...Kg7 32.Nf1 Kg6 { with a timely ...e5 to follow. } ) 28...Rg7 29.Qa1 h5 30.Nd2 { A sad
necessity - White's knight must relocate to f1 in order to halt the progress
of Black's g-pawn. Now, however, all hell breaks loose in the center. } 30...e5 { Of
course, } ( 30...g4 31.hxg4 Rxg4 32.Nf1 h4 { was perfectly acceptable as well,
but here White is still breathing after } 33.Qd1 Qg7 34.Ne3 Nxe3 35.Rxe3 Rxg2+ 36.Kh1 ) 31.dxe5 fxe5 32.Qc1 e4 { Charge of the light brigades! Black's
idea is not so much to threaten ...e3 as to vacate the e5 square for the
knight. } 33.Nf1 g4 34.hxg4 hxg4 35.Kh1 ( 35.Qd2 { was nominally more
resilient, but Black's after } 35...Qe6 36.Bf4 Rf8 37.Ng3 Rh7 { is obviously not in
question. } ) 35...g3! { A lovely sacrifice that puts an immediate end to
White's hopes of prolonging the battle. Black will not have forced mate, but
after the exchange of my bishop, the knight will have unrestricted access to
the monstrous d3 square. } 36.Bxg3 ( 36.Nxg3 Nxg3+ 37.Bxg3 Ne5! { lost
immediately, since } 38.Bxe5 Rxe5 39.Qf4 Rh5+ 40.Kg1 Qh3 { leads to immediate
checkmate. } ) 36...Nxg3+ 37.Nxg3 Ne5 38.Qe3 Rf8 39.Kg1 Nd3 40.Qd4 { I was
hoping that the computer would greedily capture the rook and allow me to set
up something resembling a fortress (although Black is clearly winning there as
well), but Stockfish has far bloodier thoughts on its mind. } 40...Qf7! ( 40...Nxe1 41.Rxe1 Qe6 42.Re2 Rg5 43.Nf1 Qf6 44.Qxf6 Rxf6 45.Ne3 { would have allowed
White to hold out for much longer. } ) 41.Qe3 Qe7 { Threatening to double on the
h-file with ...Rh7 and ...Qh4. Of course, White can do very little to stop the
rampage. } 42.Qb6 Nf4! { The decisive shot. With the g6 square secured, I
will not be able to stop ...Rh7 followed by ...Qh4. With mate clearly imminent
and Rybka slowly but steadily becoming more and more pessimistic about White's
saving chances, I decided to end the agony. }
0-1[/pgn]

[pgn]
[Event "Cyborg v. Machine "]
[Site "Hillsborough, CA "]
[Date "2014.07.19"]
[Round "2"]
[White "Stockfish 5"]
[Black "Naroditsky + Rybka 3 "]
[Result "1-0"]
[ECO "E46"]
[Plycount "133"]
[Sourcedate "2014.05.08"]

1.d4 Nf6 2.c4 e6 3.Nc3 Bb4 4.e3 O-O 5.Nge2 d5 6.a3 Bd6 { A passive, but
extremely solid continuation that I specially chose for the occasion. Black
gives his opponent a large space advantage and a clear queenside superiority,
but the resulting position will be nearly deadlocked and exceedingly hard to
win. } 7.c5 ( 7.Ng3 { leads to a more interesting position, although I doubt
that Black will have serious problems equalizing after } 7...c6 8.Bd3 ( 8.e4 dxe4 9.Ngxe4 Nxe4 10.Nxe4 Bc7 { does not strike me as particularly threatening. } ) 8...e5 9.dxe5 Bxe5 10.cxd5 cxd5 11.O-O Re8 { (Shengelia-Halkias, Bundesliga
2009). } ) 7...Be7 8.b4 { It is crucial to solidify the c5 pawn as soon as
possible, in order to take the sting out of the irksome ...b6. } 8...a5 9.Bb2 { Oddly enough, nearly everyone plays 8...b6 in this position (and only then, in
some cases, 9...a5), but frankly I do not quite understand why this is the
case. } ( 9.b5 { looks intimidating at first sight, but after } 9...c6! 10.b6 e5 { Black takes over the initiative, since White is not developed enough to meet
the sudden opening of the center. } 11.Ng3 { is met by the strong } 11...h5! { , when } 12.Nxh5 Nxh5 13.Qxh5 Bf6 { is very dangerous for White. } ) 9...b6 10.Nf4 c6 { Once again, Black's position is very passive, but it lacks any structural
defects and White will find it nearly impossible to break through on any flank.
At this point, I was already rejoicing at my successful choice of opening
variation: perhaps I would not lose all four games after all! } 11.Be2 { The
human move, of course, is } ( 11.Bd3 { , but there is no reason why the bishop
is better placed here than on e2. In Socko-Sargissian, Warsaw (blitz) 2012,
Black equalized without trouble after } 11...Nbd7 12.O-O Qc7 13.Nfe2 Ba6 14.f4 Rfb8 15.Rb1 axb4 16.axb4 Bxd3 17.Qxd3 bxc5 18.bxc5 Rb4 { and Black is
certainly not worse (the game was eventually ended in a draw). } ) 11...Ba6 12.Bxa6 Nxa6 ( 12...Rxa6 13.O-O Nbd7 { is also perfectly acceptable, but the rook
on a6 struck me as somewhat awkward. } ) 13.Nd3 Qc7 14.O-O { By this point, the
time disparity had grown rather large: I was playing nearly instantly, since
all of my moves "could not be wrong," while Stockfish was taking up loads of
time trying to sort through the subtleties of the position. } 14...Qb7 15.Na4 Nd7 16.bxa5 { At this point, Rybka was emphatically suggesting 16...bxc5 (at least
for some time), but I did not want to allow any kind of tactical possibilities. } 16...bxa5 { It turns out that my choice was actually the correct one! After } ( 16...bxc5?! 17.dxc5 Rab8 ( 17...Naxc5? 18.Ndxc5 Nxc5 19.Qg4 f6 20.Nxc5 Bxc5 21.Qxe6+ Kh8 22.Bc3 { and it turns out that Black cannot win the pawn
back. } ) 18.Bc3 Qc7 19.Qe2 e5 20.Rac1 Rfe8 21.Nxe5 Naxc5 22.Nxc5 Nxc5 23.Bd4 { Black is facing some serious difficulties due to the weak pawn on c6.
Furthermore, he is a pawn down and it is not clear how to immediately regain
the material. } ) 17.Rb1 Rfb8 18.Bc3 Qa7 19.Rxb8+ Rxb8 20.Qd2 { Of course, } ( 20.Bxa5 { is out of the question due to } 20...Naxc5! 21.Naxc5 Nxc5 22.Ne5 Rc8 23.Bb4 Ne4 24.f3 Bxb4 25.axb4 Nd6 { with an immediate draw. } ) 20...Bd8 { Now,
we enter the longest and most boring stage of the game. With the a5 pawn
untouchable for tactical reasons, I could see no way for White to make any
progress whatsoever. Furthermore, I still had about 30 minutes left on my
clock, will Stockfish was nearing single digits. His next 20 moves seemed to
confirm my suspicions: the game would end in a draw. But I urge you to watch
closely: White does not, in fact, stand still. } 21.g3 ( 21.Bxa5 { once again
leads to immediate liquidation after } 21...Naxc5 { once again leads to immediate
liquidation after } 22.Naxc5 Bxa5 23.Qc2 Nxc5 24.Qxc5 Qxc5 25.Nxc5 Rb2 ) 21...h6 22.Kg2 Rb5 { Not strictly necessary, but I wanted to err on the side
of caution. } 23.Ra1 { Typical computer chess. For some mysterious reasons,
other waiting such as Qc2 or Rc1 are branded as horrible by the engine. Of
course, if a computer starts making pointless moves, it is usually an
encouraging sign: he does not know what to do, and a draw is near! Little did
I know what was to befall me in another 30 moves. } 23...Qa8 24.Rc1 Rb3 25.Qc2 Rb8 26.Be1 Rb5 27.h3 Nab8 28.f3 { Ugh. So it won't be as easy as moving back and
forth after all! White clearly prepares a central expansion with e3-e4, but I
was still not particularly alarmed; after all, even if White advances his pawn
to e5, what will he have achieved? } 28...Rb7 29.e4 Qa6 { Beginning my own
regrouping. The game is beautifully placed on a6, and once the d7 knight
reaches g6, the other knight will find a good parking spot on d7. } ( 29...dxe4? { I briefly contemplated this move as well, but it looks like White gains
additional chances to break through after } 30.fxe4 Nf6 31.Ne5! Bc7 32.Nc4 Nbd7 33.e5 { and Black has done nothing but weaken his own position. } ) 30.e5 Nf8 31.Qd1 Ng6 32.Rb1 { After Jesse made this move on the board, I audibly
sighed with relief. After all, with the rooks off the board and all important
squares on the queenside protected, what could possibly go wrong for Black? } 32...Rxb1 33.Qxb1 Nd7 34.Nab2 Ne7 35.f4 ( 35.g4 { is also harmless, since after } 35...Ng6 36.f4 f6! 37.Qc2 fxe5 38.fxe5 Ndf8 { Black completely neutralizes
White's pawn advance. } ) 35...Nf5 36.Bf2 Qb7 37.Qc2 Nf8 38.a4 Qb8 39.Nc1 Qb4 40.Nbd3 Qb7 { Around this point, I began to grow somewhat fatigued and
irked by Stockfish's cat-and-mouse play. Although I still had plenty of time
remaining, I wanted to accelerate the process and started to make my moves
without too much. } 41.Qb3 Qa7 42.Ne2 { It was around this point that I began
to realize that my opponent was not running in place. Although g3-g4 is, for
the moment, not particularly dangerous due to ...Nh4+, it could certainly
become a future threat if White finds the correct regrouping. Nevertheless,
Rybka showed total equality, so I was still very confident. } 42...Ng6 43.Ndc1 Kf8 44.Qd3 h5 { Solidifying the knight on f5 and discouraging White from playing
g4. Of course, it is impossible to discourage the computer from doing anything! } 45.Kh2 Qb7 46.Be1 Kg8 47.Qb3 Qa7 48.Bd2 { This is the point. Now, g4
becomes a very real possibility, since the f4 pawn will be well-protected and .
..Nh4 will no longer be a convincing reply. } 48...Nh6 { An unfortunate necessity.
The negligent } ( 48...Qa8 { would have led to unpleasant consequences after } 49.g4 hxg4 50.hxg4 Nh6 51.Kh3 { , although even here, it looks like Black should
be able to hold the draw with } 51...Bc7! 52.Qf3 Qd8 53.Qf2 f5 54.exf6 Qxf6 ) 49.Qf3 Qa6 { An important resource. White's queen must return in order to defend
the pivotal c4 square. } 50.Kg1 ( 50.Qxh5? Qc4! { allowed Black to take
over the initiative, since } 51.g4 Qc2 52.Be3 Qe4 { is actually lost for White. } ) 50...Ne7 { Sending the other knight to f5! } ( 50...Qc4 { is well met by } 51.Qb3 ) 51.Qd3 { Unfortunately, the ending is lost for Black since he cannot
defend the a5 pawn, and 51...Qc4 is also ineffective on account of 52.Qb3. } 51...Qa7 ( 51...Qc4?! 52.Qb3 { and Black has nothing better than } 52...Qa6 { , since } ( 52...Kh7?! 53.Kf2 f6 54.exf6 gxf6 55.Qb8 Nf7 56.Bxa5! Bxa5 57.Qf8 { is a
course of events Black certainly does not desire. } ) 53.Qb8 Nc8 54.Kf2 { But
now, with Black's queen tied down, he has no choice but to allow g4. } 54...Kf8 55.Kf3 Nf5 56.g4 hxg4+ 57.hxg4 Nfe7 58.Qb1 { and the position becomes extremely
unpleasant for Black. } ) 52.Nb3 Nef5 53.Nec1 g6 54.Qf3 Qa6 55.Qe2 { The
immediate } ( 55.g4 { leads to a tenable position for Black after } 55...Nh4 56.Qf1 Qb7 57.Qe2 Kg7 { when I cannot see a way to make further progress. } ) 55...Qb7 56.Be1 { By this point, I was already quite exhausted from all of the
defending, and Rybka had changed its evaluation to a slight advantage for
White. Instead of patiently allowing it to analyze the position (not to
mention thinking independently), I blitzed out the move it briefly considered
best: } ( 56.g4? Ng3! 57.Qe1 Ne4 58.f5 Nxd2 59.Qxd2 Kh7 ) 56...h4?? { An
inexplicable howler, made all the more frustrating by the fact that it
nullifies all of my previous efforts! Indeed, had I actually given this
position any thought myself, I would never have made this move. If Black had
simply waited - } ( 56...Qa8 { , for example - the game would have probably
ended in a draw. It is true that White can play } 57.g4 { , but after } 57...Nh4 58.Kh2 ( 58.gxh5?! N6f5 59.hxg6 Nxg6 { is an instant fortress for Black. } ) 58...Kg7 59.Kg3 hxg4 60.hxg4 g5! { it looks like Black is doing quite all right,
even if he will be forced to give up the a5 pawn. For instance, } 61.fxg5 Bxg5 62.Bd2 ( 62.Bxa5?? N4f5+! 63.gxf5 Nxf5+ 64.Kf2 Bxc1 65.Nxc1 Qxa5 ) 62...Qd8 ( 62...Bxd2 63.Nxd2 Qd8 64.Qf2 Ng8 65.Qe3 { is less convincing. } ) 63.Bxa5 Qe7 64.Nd3 Ng6 { and Black has more than enough kingside counterplay to
secure the draw. } ) 57.g4 Ng3 { This was the idea, but White's 59th move shows
just how suicidal this "idea" turns out to be! } 58.Bxg3 hxg3 59.Qd2! { Simple, and insanely strong. With one move, White threatens f4-f5, attacks the
a5 pawn, takes the crucial b4 square under control, and prepares to win the g3
pawn with Kg2 (and Ne2, if necessary). With both of Black's minor pieces
prisoners in their own camp, I could do nothing but wait. } 59...Qa6 ( 59...Kh7 60.Kg2 Bh4 61.Qxa5 { was no better. } ) 60.f5 Kh7 61.f6! { Decisive. Black is
now essentially playing without two pieces, and Rybka's evaluation confirmed
my worst suspicions: the damage has been done. } 61...Qc4 62.Qd3 { A human might
have spent 45 minutes calculating the ramifications of this move, but
Stockfish took about five seconds. The queen trade is suicidal, but keeping
them on the board leads to immediate disaster as well. } 62...Qxa4 ( 62...Qxd3 63.Nxd3 Bc7 64.Kg2 Kg8 65.Kxg3 Kh7 ( 65...g5 66.Nd2 Kh7 67.Nf3 Kg6 68.h4 gxh4+ 69.Kxh4 { is similarly hopeless. } ) 66.Kf4 Kg8 67.Kg5 Kh7 68.h4 Ng8 69.h5 gxh5 70.Kxh5 { and Black will literally be pushed off the board after g5,
Nf4, and g6+. } ) 63.Qxg3 Qb4 64.Qh4 { Devastating. The threat of g5 can
actually be parried, but at too great a cost. } 64...g5 65.Qxg5 Qe1+ 66.Kg2 Qe4+ 67.Kf2 { Black runs out of effective checks, and he is forced to enter a
completely hopeless endgame with 67...Qg6. After analyzing the endgame for two
or three minutes, I flicked over my king. 2-0. } ( 67.Kf2 Qg6 68.Qxg6+ Kxg6 69.Nd3 Bc7 70.Nb2 { with Na4-b6 to follow. } )
1-0[/pgn]

[pgn]
[Event "Cyborg v. Machine "]
[Site "Hillsborough, CA"]
[Date "2014.07.19"]
[Round "?"]
[White "Naroditsky + Rybka 3 "]
[Black "Stockfish 5 "]
[Result "0-1"]
[ECO "C11"]
[Plycount "54"]
[Sourcedate "2014.05.08"]

1.e4 e6 2.d4 d5 3.Nc3? { The question mark is not misplaced! I was still
clinging to the hope that I could crush the machine with my theoretical
knowledge, but to choose the most tactically complex variation in all of the
French Defense is a decision of epic - uh, foolishness. } 3...Nf6 4.e5 { It was not
too late to enter relatively calm waters with 4.Bg5 or even 4.Bd3, but no such
thoughts crossed my mind. } 4...Nfd7 5.f4 c5 6.Nf3 Nc6 7.Be3 Be7 8.Qd2 O-O 9.dxc5 Nxc5 { Yet again, the computer has reproduced the theoretical continuation
move for move. I have not played 3.Nc3 in many years, and around this point, I
came to the realization that my knowledge of this position is somewhat scant.
In principle, White should have an edge, but the position is extremely sharp. } 10.O-O-O b6 { This is not currently in vogue (most French specialists have
switched to 10...a6 and 10...Qc7), but all of Stockfish's moves must obviously
be regarded with a great deal of respect. For now, Black solidifies his
position and shows that he is in no real hurry to begin the kingside pawn
storm. } 11.Kb1 { A logical reply. If Black is spending time on various
positional niceties, I figured that I could do the same. } ( 11.Bb5!? Bb7 12.Bxc6 Bxc6 13.Nd4 { is an interesting alternative that was successfully essayed
by a few strong GMs; after } 13...Qc7 ( 13...Bb7 14.f5! Qd7 15.f6 gxf6 16.exf6 Bxf6 17.Bh6 Rfc8 18.Qf4 { looks quite dangerous for Black. } ) ( 13...Bd7 14.Kb1 Qc7 15.Nf5! Bc6 16.Nxe7+ Qxe7 17.Rhf1 { is slightly better for White. } ) 14.f5!? Bb7 15.Bf4 Qd7 16.Bg5 exf5 17.Bxe7 Qxe7 18.Nxf5 Qxe5 19.Qg5 Kh8 20.Rhe1 Ne4 21.Nxd5 Qxb2+ ( 21...Nxg5 22.Rxe5 Rad8 23.c4 { leads to a very
pleasant endgame for White. } ) 22.Kxb2 Nxg5 23.Re7 { White has very decent
compensation for the pawn. } ) 11...Qc7 12.h4 { Probably not a mistake per se,
but at this point I was not sure how I was going to make any progress on the
kingside. } ( 12.Bb5 Bb7 13.h4 { was seen in Nakamura-Vitiugov, Reggio Emilia
2012. Black played the awkward 13...Na5, but after } 13...Ne4 14.Nxe4 dxe4 15.Nd4 Nb4! { Black is certainly not worse. } ) ( 12.g4 { might well have been a better
choice. White's idea is not so much to prepare f4-f5 but to simply complete
his development. For instance, } 12...Bb7 13.Bg2 Na5 14.Qe2 Rac8 15.Bd4 Ba6 16.Qf2 { and the position is very double-edged, but I do not think that White is
worse. } ) 12...a6 13.Qe1 { Rybka was emphatically in favor of this move, so I
decided to make it even though I did not fully understand its purpose at the
time. Perhaps White is waiting for ...b5, so that he can dislodge the annoying
c5 knight with Qf2. In any case, this is certainly not the culprit. } ( 13.Qf2 { deserved attention, but I did not like } 13...f6 14.exf6 Bxf6 { and White's queen
must return to e1. } ) 13...b5 14.Qf2 Na4! { An unpleasant move that
immediately induced a panicked response. To be fair, Rybka did recommend the
correct 15.Nxa4, but I was very reticent about opening the b-file. } 15.Ne2? { But this is an awful move that gives Black the initiative on a silver platter.
In fact, after } ( 15.Nxa4 bxa4 16.Nd4 { the position is still about equal,
since Black will find it very difficult to make use of the open file. A sample
line is } 16...Bd7 17.Nxc6 Qxc6 18.Bd4 Rfc8 19.h5 Bc5 20.Bxc5 Qxc5 21.Qxc5 Rxc5 22.Rd4 { with an approximately equal endgame. } ) 15...f6! { A rather obvious
move that completely changes the course of the game. With the inevitable
opening of the center, White's position becomes very precarious - to put it
mildly. } 16.Qg3 { Practically forced, since } ( 16.exf6? Bxf6 17.Ned4 Bxd4 18.Nxd4 e5! 19.Nxc6 Qxc6 { is clearly terrible. } ) 16...fxe5 17.fxe5 Nc5 { Power chess! The knight has done its job on a4, and now it relocates to a far
more juicy outpost on e4. White is already worse, but I decided to hugely
exacerbate the problem with another hasty, inexplicable move. } 18.h5? { The
decisive mistake, and this time I mean it! Obviously, h6 will never be a real
threat (Black will simply reply with ...g6), so in effect, this is a wasted
move. White's position would still have been tenable after } ( 18.Nf4 Ne4 19.Qh2 ( 19.Qh3 Rxf4 20.Bxf4 Nf2 ) 19...Bd8 20.Nd3 { , even though the
initiative is clearly on Black's side after 20...Na5. } ) 18...Ne4 19.Qh2 a5 { Simple and powerful, yet again. With White's pieces dreadfully passive and
uncoordinated, there is very little I can do to stop the queenside pawn
offensive. } 20.Ned4 a4! { The computer made this move very quickly, paying
absolutely no heed to the hanging b5 pawn. In fact, White's position is
already lost - the pawns are too far advanced, and the king is too unprotected.
Rybka, however, was still showing equality. } 21.Nxc6 { The immediate } ( 21.Nxb5 { loses quickly after } 21...Qa5! 22.a3 Rb8 23.h6 g6 { with ...Rxb5 to follow. } ) 21...Qxc6 22.Nd4 Qb7 { Only now did Rybka (and I) come to terms with the
reality. White simply cannot stop the crushing a3. } 23.Bd3 ( 23.Bxb5 a3 24.g4 axb2 25.Kxb2 Bd7 { is curtains. } ) 23...a3 24.Bxe4 { Desperation. I was hoping
for the naive 24...dxe4, which would have allowed me to undo all of the damage
with 25.b3, but the computer has no such frivolous thoughts in its mind. } ( 24.b3 Nc3+ ) 24...axb2! { Devastating. White can keep his extra piece with 25.
Bd3, but only if he wants to get mated along the a-file! } 25.Bxh7+ ( 25.Bd3 Qa6 26.Kxb2 Qa3+ 27.Kb1 Bb4 { and ...Bc3 cannot be stopped. } ) 25...Kxh7 26.Kxb2 Qa7 27.h6 g6 { In an over-the-board game I might have continued to resist,
but against a 3400 computer, there is simply no point in continuing the agony. } ( 27...g6 28.Rd3 Qxa2+ 29.Kc1 Qc4 30.Nb3 Ra2 { and the attack comes crashing
through. } )
0-1[/pgn]

[pgn]
[Event "Cyborg v. Machine "]
[Site "Hillsborough, CA "]
[Date "2014.07.19"]
[Round "4"]
[White "Stockfish 5"]
[Black "Naroditsky + Rybka 3 "]
[Result "1/2-1/2"]
[ECO "C05"]
[Plycount "142"]
[Sourcedate "2014.05.08"]

1.d4 Nf6 2.Nf3 g6 { Despite Rybka's vehement protests, I decided to stick
with the opening I know best. } 3.c4 Bg7 4.Nc3 O-O 5.e4 d6 6.Be2 e5 7.d5 { I was quite happy with Stockfish's choice, since computers generally have an
imperfect understanding of positions with a closed center. } 7...a5 8.O-O Na6 9.Be3 { In the Petrosian Variation, White usually situates his bishop on g5, but
this move is quite reasonable as well. } ( 9.Bg5 h6 10.Bh4 ( 10.Be3!? { was
recently essayed by strong Russian GM Boris Savchenko (Savchenko-Kokarev,
Khanty-Mansiysk 2013). After } 10...Ng4 11.Bd2 f5 12.g3 f4 13.Nh4 Qg5 14.Kh1 Rf7 15.Bf3 Bd7 16.Qe2 { the position was very complex, although White eventually
squeezed out a win. } ) 10...Bd7 11.Nd2 { leads to a well-known position that
is seen rather frequently at Grandmaster level. Black has a few ways to
continue, but I like } 11...Nc5 12.f3 b6 13.Re1 ( 13.Nb3 Na4 14.Nxa4 Bxa4 15.Qe1 Qe8 { does not promise much either. } ) 13...Qe8 14.Nb5 Rc8 15.a3 a4 16.Bf2 h5 { with a very double-edged position (Ivanchuk-Mamedov, World Teams 2013 - the
game was later drawn). } ) 9...Nc5?! { A somewhat impulsive move. It is usually
not a very good idea to allow the Be3-Nd2 regrouping in the King's Indian, and
this case is no different. Instead, } ( 9...Ng4 { strikes me as the most natural
and strongest reply. For instance, } 10.Bg5 f6 11.Bh4 ( 11.Bd2 f5 12.Ng5 Nf6 13.exf5 Bxf5 14.Be3 Qe7 { with approximate equality. } ) 11...Qe8 12.Ne1 f5 13.Nd3 Nf6 { and I do not think Black is any worse, although the position
remains very complex. } ) 10.Nd2 { Logical and strong. Now, ...Ng4 is no longer
possible, and ...f5 will always be met by f3. } 10...Ne8 { Around this point I was
already kicking myself for giving White a beautiful position and a clear edge.
However, I also remained hopeful, since Rybka's top choice allowed me to
completely block the position. } 11.Nb3? { As I mentioned at the outset,
computers can still be quite naive in some instances. Evidently, Stockfish did
not seriously consider my next move, which allows White to ruin Black's pawn
structure. } ( 11.a3! { was the right way to play, when Black has no good way
of stopping b2-b4. } 11...f5 { is probably the lesser of evils, but after } ( 11...a4 12.Bxc5 dxc5 13.Nxa4 Qe7 14.Nc3 { gives Black very little for the pawn. } ) 12.exf5 ( 12.f3 a4! 13.Bxc5 dxc5 14.Nxa4 Bh6 15.Nc3 Qh4 { and Black suddenly
develops a fearsome initiative. } ) 12...gxf5 ( 12...Bxf5 13.Bxc5 dxc5 14.Nde4 { is positional suicide. } ) 13.Qc2 a4 14.Rae1 Bd7 15.f3 { and White will
increase the pressure with Qb1 followed by Bd1-c2. Black has his share of
chances as well, but this is the kind of position in which Stockfish truly
thrives. } ) 11...b6 12.Nxc5 { It was already too late to undo the damage with } ( 12.Nd2 { , when } 12...f5 13.f3 Nf6 14.Bg5 h6 15.Bh4 Qe8 { gives Black a great
position. } ) 12...dxc5! { This is the move Stockfish evidently underestimated.
Black's pawn structure is irreparably damaged, but White will have absolutely
no way to make any significant progress on the queenside! He can try to
prepare a3-b4, but Black will quickly counter with his own pressure on the
other flank (...f5-f4, g5, etc.). This is an instructive moment, and a clear
sign that computers are still not totally infallible. } 13.Bd3 ( 13.f4 { I was
slightly worried about this move during the game, but it turns out that Black
is more than okay after } 13...Nd6 14.Bd3 Ba6! 15.Qe2 Qe7 16.f5 gxf5 17.exf5 e4 { with a strong initiative. } ) 13...Nd6 14.f3 Qe7 15.b3 f5 { Despite the
computer's optimistic evaluation, Black is already not worse. } 16.Bc1 { A
clear sign that the computer has no clue what to do. Practically speaking, it
was probably better to leave the bishop on e3 (or d2) in order to make a3-b4
possible. } 16...f4 17.Bb2 g5 18.h3 Bd7 19.Nb5 Nxb5 20.cxb5 Qd6 { The queen is
not usually a skilled blockader, but it is perfectly comfortable on d6, since
White has absolutely no way through. } 21.a4 { White will probably have to play
this move sooner or later, since I already had the possibility of doubling
rooks on the a-file and breaking through with ...a4. } 21...h5 22.Be2 { Overjoyed at
the prospect of not getting swept, I made my next move instantly: } 22...h4? { Inexplicable cowardice. This move indeed leads to a totally blockaded position
(although I had to suffer for another 50 moves!), but I simply cannot
understand why I didn't play for the win by slowly preparing the ...g4 break.
After all, I can always lock the position with ...h4! } ( 22...Rf6 23.Rc1 Rg6 { and given time (of which he has plenty), Black will double on the g-file, move
his bishop to h6 or f6, and break through! However, White obviously does not
stand still either; after } 24.Kf2 Kf7 25.Rh1 Rg8 26.Qd3 Bf6 27.Rh2 { White's position is rock-solid, and even } 27...g4 28.hxg4 hxg4 29.Ke1 { does not
achieve the desired effect. Nevertheless, Black is firmly in the driver's seat
after } 29...gxf3 ( 29...g3 30.Rh7+ Bg7 31.Kd2 Rh6 32.Rh1 Rxh1 33.Rxh1 { is
drawn as well. } ) 30.Bxf3 Ke7 { In any case, as kibitzers on Chess.com pointed
out, I should have at least given this a try! } ) 23.Qd2 Bf6 { Now, neither side
can make any progress, but Stockfish still evaluated the position as clearly
better for White (probably due to Black's bad pawn structure), and continues
to move back and forth...for another 50 moves! } 24.Rac1 Bg7 25.Bd3 Rfe8 26.Bb1 Bf6 27.Qc3 Bg7 28.Rfd1 Kf7 29.Rd3 Kg8 30.Bc2 Kf7 31.Qe1 Kg8 32.Rdd1 Kf7 33.Qc3 Kg8 34.Bd3 Kf7 35.Re1 Kg8 36.Rcd1 Bf6 37.Rf1 Bg7 38.Rc1 Bf6 39.Rfd1 Bg7 40.Re1 Bf6 41.Kh1 Bg7 42.Rf1 Bf6 43.Ra1 Bg7 44.Kg1 Kf7 45.Rac1 Kg8 46.Ba1 Kf7 47.Rfd1 Kg8 48.Qd2 Kf7 49.Bb2 Kg8 50.Ba3 Bf8 { Not
strictly necessary, but I wanted to make sure that the computer would not get
any thoughts of sacrificing the bishop on c5. } 51.Qc3 Ra7 52.Bb2 Bg7 53.Bb1 Raa8 54.Bc2 Ra7 55.Qd2 Raa8 56.Kf2 Bf6 57.Ba1 Kg7 58.Bd3 Kg8 59.Qe2 Kg7 60.Bc3 Kg8 61.Be1 Kg7 62.Kg1 Kg8 63.Kh2 Kg7 64.Bf2 Be7 65.Bc4 Bf6 66.Kg1 Be7 67.Bd3 Kg8 68.Rc3 Bf8 69.Bb1 Be7 70.Ba2 Kh7 71.Bb1 Kg8 { And finally,
the 50 move rule comes to the rescue. Draw! }
1/2-1/2[/pgn]
"Good decisions come from experience, and experience comes from bad decisions."
__________________________________________________________________
Ted Summers

Post Reply