Members · Prefs · Laboratory · Collections · Openings · Endgames · Sacrifices · History · Search Kibitzing · Kibitzer's Café · Chessforums · Tournament Index · Players · Kibitzing

AlphaZero (Computer) vs Stockfish (Computer)
AlphaZero - Stockfish Match (2018), London ENG, Jan-18
English Opening: Anglo-Indian Defense. Nimzo-English Opening (A17)  ·  1-0



explore this opening
find similar games 219 more AlphaZero/Stockfish games
PGN: download | view | print Help: general | java-troubleshooting

TIP: If you find a mistake in the database, use the correction form. There is a link at the bottom that reads "Spot an error? Please suggest your correction..." Avoid posting corrections in the kibitzing area.

PGN Viewer:  What is this?
For help with this chess viewer, please see the Olga Chess Viewer Quickstart Guide.


Kibitzer's Corner
Premium Chessgames Member
  keypusher: At best this is A0's second-most-remarkable-67-move-victory in the Nimzo-English (or whatever this opening is called), after AlphaZero vs Stockfish, 2018. But it's pretty striking.

After 17.Rab1:

click for larger view

SF plays 17....Nd7. The Bible says, "Suffer not a witch to live" (Exodus 22:18) and I think the same rule should apply to A0's bishop on the a1-h8 diagonal (see the linked game above for another example of the evil it does). A0 of course responds with 18.Ba1!.

But SF knows of A0's wiles, and here

click for larger view

it plays 24....Be4, compelling 25.Nxe5 fxe5 and the long diagonal seems secure.

But A0 never rests.

click for larger view

34.Bd3+! Kh8 (34....e4 35.Bxe4+ Bxe4 36.Qxe4+ Qg6 37.Qxg6+ Rxg6 38.Rd5 +-) 35.Rdd1 Nxa5 36.c5! dxc5 37.Bc3! and now the pawn at e5 cannot be defended, e.g. 37....Re6 38.f4. So the bishop gets its diagonal after 37....b5 38.Bxe5.

click for larger view

A0 responds to SF's doubled rooks with the elegant 42.Bf5! Kf8 (42....Rxd1 43.Be6) 43.Rxd6 cxd6 44.Ba1 (natch) Qe7 (44....d5 is recommended in the computer annotated score on this page. My SF 10 evaluates the position after 45.Bd3 at +1.37 at 40 ply) 45.Bd3 Nc4 46.a4! and the pawn is off to the races.

Position after 46....Na3

click for larger view

Even at 40 ply (35 minutes) my SF10 agrees with CG Stockfish that 47.axb5 (+1.94) is stronger than 47.Qf4+ (+1.51). In fact 47.Qf4+ is SF10's third choice (47.e4 is +1.60). It's almost a cliche to say it at this point, but it does feel like A0's superiority over SF is strategic rather than tactical.

Incidentally, in the game, if 49....bxa4, then 50.Bg6 Qe7 51.Rc1! Rc7 52.Rc3 Nb5 53.Rc4 is winning.

After 59....Nd5

click for larger view

You can see that SF is going to get steamrolled in the center after 60.f4.

Premium Chessgames Member
  keypusher: <Even at 40 ply (35 minutes) my SF10 agrees with CG Stockfish that 47.axb5 (+1.94) is stronger than 47.Qf4+ (+1.51). In fact 47.Qf4+ is SF10's third choice (47.e4 is +1.60). >

I got curious, so I let SF10 run overnight in this position. At 50 ply/676 minutes, 47.axb5 was still the clear champ, +2.52. There was a new #2, 47.f4 (+2.26). 47.Qf4 was +1.65 and 47.e4 was +1.43.

Premium Chessgames Member
  maxi: <keypusher>, your "Suffer not a witch to live" bothers me, since the *witch* in this case is a piece of the cloth and respect is due.

But, never mind, there is something else that bothers me. SF plays 16...Na4 and has a chance of exchanging the accursed soul, yet blows it and does not execute the exchange. Possibly SF had not smelled the danger and simply wanted to take its N from b6 to c5.

Evidently Jehovah does not look favorably upon mercy.

Premium Chessgames Member
  maxi: But A0 does retreat the piece.
Premium Chessgames Member
  AylerKupp: <AlphaZero (Computer) vs. Stockfish (Computer) (part 1 of 2)>

<<maxi> Possibly SF had not smelled the danger and simply wanted to take its N from b6 to c5.>

Possibly so. I had Stockfish analyze the position after 17.Rab1 for a while (I forgot how long) and 17...Nxb2 was not among its top 3 moves.

So that begs the question, how might have the game continued after 17...Nxb2 18.Rxb2? I had Stockfish analyze the position for several hours and this is what it came up with at d=50:

1. [+0.44]: 18...Reb8 19.Rdb1 Ra7 20.Bd1 Bg6 21.Bc2 Nd7 22.Qd2 Qd8 23.a4 Rba8 24.Rxb7 Rxa5 25.R7b2 Nb6 26.Ra2 Bf5 27.Rb5 R5a6 28.a5 c6 29.Rb3 c5 30.Rba3 Nc8 31.Ne1 Be6 32.Bd1 Rb8 33.Ra1 Ne7 34.Bf3 Qc7 35.Qc3 f5 36.Nc2 Kh7 37.d4 Bf7 38.Be2 e4 39.Ne1 (I thought that 39.Ra4 with the plan of Nd2-a3-b5 looked promising and although not realizable, at low ply (d=27) Stockfish evaluated White as winning with evaluations > [+3.00] after either 39...Rba8, 39...Be8, or 39...Rc8) 39...Qb7 40.d5 Kg8 41.g3 Qb4 42.Qc2 Be8 43.Ng2 Ng6 44.Bh5

click for larger view

White is a pawn up and has an easier game, although Stockfish must consider that Black has some compensation since its evaluation is < [+1.00]. Indeed, restarting the analysis from this position Stockfish evaluates the resulting position at [only [+0.76]], d= 39 after 44...Ne5 45.Bxe8 Rxe8 46.Ra4 Qb7 47.Rb1 Qf7 48.Raa1 Rea8 49.Rb5 Qf6 50.Ra4 g5 51.Ne1 Qd8 52.Qa2 h5 53.Qb1 Kg7 54.f4 exf3 55.Qxf5 f2+ 56.Qxf2 R6a7 57.Qe2 Qf6 58.a6 Rxa6 59.Rxa6 Rxa6 60.Qxh5 Ra7 61.Rb8 Rf7 62.Qh8+ Kg6 63.Rg8+ Kf5 64.Qxf6+ Rxf6 65.Re8 Nxc4 66.Kg2 Ne5 67.Nf3 Nxf3 68.Kxf3 g4+ 69.hxg4+ Kg5+ 70.Ke4

2. [+0.63]: 18...Ra7 19.Rdb1 Rb8 20.Bd1 (this is just a move transposition of line 1 above, although the game later takes a different turn) 20...Bg6 21.Bc2 Nd7 22.Qd2 Qd8 23.a4 Rba8 24.Rxb7 Rxa5 25.R7b2 Nb6 26.Ra2 Bf5 27.Rba1 (here White deviates, from line 1's 27.Rb5) 27...c5 (but it's also just a move transposition with a tempo gain for both sides since Black plays ...c5 in one move rather than two by ...c6 and ...c5) 28.Rb1 R5a6 29.Qd1 Be6 30.Rba1 f5 31.Nd2 Kh7 32.Rb1 d5 (but this is a true deviation since in line 1 Black never got to play ...d5) 33.cxd5 Nxd5 34.Ra3 Qe7 35.Nc4 Bf7 36.e4 fxe4 37.dxe4 Nf4 38.Nb6 c4 39.Re3 Rb8 40.Nd5 Qd8 41.Qf1 Rg6 42.Rxb8 Qxb8 43.Rg3 Rxg3 44.fxg3

click for larger view

White has a better game. It is a pawn up (although doubled) and threatens to win a second pawn by 45.Qxc4. The pawn also falls after 44...Nxd5 45.exd5+ Bg6 46.Bxg6+ Kxg6 47.Qxc4, and White's king has a safe haven against Black's queen checks on h2. I think it's a resignable position and I'm surprised that Stockfish only evaluates it at [+0.63]. Horizon effect?

Premium Chessgames Member
  AylerKupp: <AlphaZero (Computer) vs. Stockfish (Computer) (part 2 of 2)>

No, restarting the analysis from the position above I clearly missed 44...Qa7+ with the possibility of 45.Kh2 Nd6 46.Bxd6 cxd6 47.Qxd6 Qxa4 preventing the loss of a second pawn. Indeed, restarting the analysis from this position Stockfish evaluates the resulting position at "only" [+0.91], d=44 after 48.Qb1 Qc6 49.Ne7 Qd6 50.Nf5 Qf6 51.Qd3 Be8 52.Ne3 Bg6 53.Nc4 Bf7 54.Nd6 Be6 55.Qa3 Qe7 56.Qc5 Bf7 57.Qc6 Qf6 58.Qc7 Bh5 59.Nf5 Bg6 60.Qc2 Bf7 61.Qc5 Be8 62.Ne7 Bf7 63.Qc7 Be8 64.Qd8 Bf7 65.Qf8 Ba2 66.Qa8 Be6 67.Qa3 Bc4 68.Nd5 Bxd5 69.exd5 and the evaluation remained the same d=34.

And after 44...Qa7+ Stockfish prefers 45.Kh2 Nd3 46.Qf5+ Bg6 47.Qf8 Ne1 48.Ne7 Nf3+ 49.Qxf3 Qxe7 50.Qc3 Qc5 51.a5 Qd4 52.Qb4 Qa1 53.Qb5 Qa2 54.Bb1 Qb3 55.Qxb3 cxb3 56.Kg1 Kg8 57.Kf2 Kf8 58.a6 Be8 59.Ke3 Ke7 60.Kd2 Kd6 61.Kc3 Kc7 62.Kxb3 Bb5 63.a7 Kb7 64.Kb4 Bf1 65.Kc5 Kxa7 66.Kd6 Kb7 67.h4 h5 68.Kxe5, evaluating it at [+2.16], d=31, also winning.

3. [+0.96]: 18...c6 19.Rdb1 Ra7 20.Rb6 e4 21.Nd2 Bxe2 22.Qxe2 d5 23.a4 dxc4 24.Nxc4 exd3 25.Qxd3 Ne4 26.Rd1 Qh4 27.Rb2 Nc5 28.Qc2 Rd8 29.Rxd8+ Qxd8 30.Qf5 Qd5 31.Qxd5 cxd5 32.Rb5 dxc4 33.Rxc5 c3 34.g4 c2 35.Rc8+ Kh7 36.Rxc2 Rxa5 37.Rb2 (a strange move, a pawn up and with pawns on both sides of the board I would have thought that White should avoid pawn exchanges with 37.Rc4) 37...Rxa4 38.Rxb7 Kg6 39.Kh2 Ra2 40.Kg3 Ra5 41.Rb6+ f6 42.Rb7 h5 43.f4 Ra3 44.Kf3 hxg4+ 45.hxg4 Kh6 46.Rd7 g6 47.Ke4 Ra5 48.Rd3 f5+ 49.Kf3 fxg4+ 50.Kxg4.

click for larger view

And this is a tablebase draw.

But of course we have no idea how AlphaZero would have played after 17...Nxb2 18.Rxb2 and (presumably) 18...Reb8. It's even conceivable that with 17.Rab1 AlphaZero "set a trap" for Stockfish to allow it (and maybe encourage it) to respond with 17...Nxb2. You never know how these neural network-based engines using MCTS will play!

Oct-30-19  onigorom: This computer stuff is hilarious. These machines understand chess 'literally', there is nothing outside of chess. In other words, they almost play like Ivanchuk.
NOTE: You need to pick a username and password to post a reply. Getting your account takes less than a minute, is totally anonymous, and 100% free—plus, it entitles you to features otherwise unavailable. Pick your username now and join the chessgames community!
If you already have an account, you should login now.
Please observe our posting guidelines:
  1. No obscene, racist, sexist, profane, raunchy, or disgusting language.
  2. No spamming, advertising, duplicate or nonsense posts.
  3. No malicious personal attacks, including cyber stalking, systematic antagonism, or gratuitous name-calling of any member Iincludinfgall Admin and Owners or any of their family, friends, associates, or business interests. If you think someone is an idiot, then provide evidence that their reasoning is invalid and/or idiotic, instead of just calling them an idiot. It's a subtle but important distinction, even in political discussions.
  4. Nothing in violation of United States law.
  5. No malicious posting of or linking to personal, private, and/or negative information (aka "doxing" or "doxxing") about any member, (including all Admin and Owners) or any of their family, friends, associates, or business interests. This includes all media: text, images, video, audio, or otherwise. Such actions will result in severe sanctions for any violators.
  6. NO TROLLING. Admin and Owners know it when they see it, and sanctions for any trolls will be significant.
  7. Any off-topic posts which distract from the primary topic of discussion are subject to removal.
  8. The use of "sock puppet" accounts to circumvent disciplinary action taken by Moderators is expressly prohibited.
  9. The use of "sock puppet" accounts in an attempt to undermine any side of a debate—or to create a false impression of consensus or support—is prohibited.
  10. All decisions with respect to deleting posts, and any subsequent discipline, are final, and occur at the sole discretion of the Moderators, Admin, and Owners.
  11. Please try to maintain a semblance of civility at all times.
Blow the Whistle See something that violates our rules? Blow the whistle and inform a Moderator.

NOTE: Keep all discussion on the topic of this page. This forum is for this specific game and nothing else. If you want to discuss chess in general, or this site, visit the Kibitzer's Café.

Messages posted by Chessgames members do not necessarily represent the views of, its employees, or sponsors. All Moderator actions taken are at the sole discretion of the Admin and Owners—who will strive to act fairly and consistently at all times.

This game is type: CLASSICAL. Please report incorrect or missing information by submitting a correction slip to help us improve the quality of our content.

<This page contains Editor Notes. Click here to read them.>

home | about | login | logout | F.A.Q. | your profile | preferences | Premium Membership | Kibitzer's Café | Biographer's Bistro | new kibitzing | chessforums | Tournament Index | Player Directory | Notable Games | World Chess Championships | Opening Explorer | Guess the Move | Game Collections | ChessBookie Game | Chessgames Challenge | Store | privacy notice | contact us

Copyright 2001-2019, Chessgames Services LLC