<Sally Simpson: ... Black's a piece up, White has the initiative and is coasting along in home prep. It's a KID and Black is playing on the Queenside.

Bollocks to that.>

Yep, if Black is defending on the queenside in the KID and not attacking on the kingside then he is losing.

After 14.Nb5 we reach the position in your first diagram:

click for larger view

As you mentioned, "Black's a piece up, White has the initiative and is coasting along in home prep. It's a KID and Black is playing on the Queenside." You suggested that instead of playing like a kid (14...Ra5) that Black should try 14...Nexd5. Before we look at that, let's see what Komodo 9.2 thinks of the situation .

At d=25 Komodo evaluates the position as effectively even with evals of [-0.15] after 14...Ra8 and [0.00] after 14...Ra6. Komodo evaluates 14...Nxe2+ as slightly favorable for White after 15.Qxe2 Ra4. It's clear that White has a lot of compensation even though he's down a knight for a pawn. Nakamura's 14...Ra5 is not among Komodo's top 3 moves.

After 14...Nxd5 15.cxd5 (I agree that 15.cxd5 is better than 14.exd5 for White, probably <much> better for the reasons you stated), Komodo evaluates the resulting position at [+0.39], d=24 after 15...Nxe2+ 16.Qxe2 Ba6 (perhaps you overlooked this possibility) 17.Rfb1 (What else? If 17.bxc7 then either the simple 17...Qxc7 or 17...Qd7 provide a small advantage for White, [+0.27] and [+0.38] respectively at d=21) 17...Qd7 18.Qc4 cxb6. And Komodo does not consider 15...Rh3 to be among Black's top 3 moves, but perhaps I should have let Komodo run longer, as I usually do.

But lets see what Komodo says after 15...Rh3 (shades of Rotlewi vs Rubinstein, 1907!) and the following position:

click for larger view

Komodo thinks it's best to just ignore the rook and evaluates the position at [+2.05], d=24 (probably winning) after 16.bxc7. It considers 16.gxh3 to be Black's 3rd best move, evaluating the position at [+1.29], d=20 after 16...Qg5+ 17.Bg4 Nxh3+ 18.Kh1 Bxg4 as you suggested, reaching the following position:

click for larger view

Certainly a terrifying position to have when you're facing Nakamura! But Komodo has no nerves and instead of your 19.Qe1 suggests 19.f3 with a possible continuation being 19...Bd7 20.Nxc7 Qe3 21.Nc4 Nf2+ 22.Rxf2 Qxf2 23.Qf1 Qc2 24.Nxd6 Bh6 25.Re1 (preliminary to setting up a 2nd rank defense, more important than the Pa2) 25...Qxa2 26.b7 Bf4 27.Qe2 Qb3 28.Na6 Qb6 29.b8Q Rxb8 30.Nxb8 Qxb8 31.Nc4 and White's exchange advantage and protected passed d-pawn should win.

click for larger view

And, BTW, at d=24 Komodo elevates 16.gxh3 to second best move and evaluates it only slightly worse than 16.bxc7 at [+2.01], also probably winning.

Back to Komodo's top line after 16.bxc7. It continues 16...Qh4 17.g3 Qh6 18.Bf3 Rxh2 19.gxf4 Rh4 20.f5 gxf5 21.Re1 Qg6+ 22.Bg2 Bh6 23.exf5 (A computer move; Komodo apparently doesn't care about opening up even more lines in the vicinity of the White king, something that I doubt anyone of us would do!) 23...Bxf5 24.Nf1 Rb4 (and this is also not a move that anyone of us would play, removing the attacking rook from the k-side) 25.a4 Bg4 26.Re2 Bf4 27.Ng3 Bxg3 28.fxg3 Bxe2 29.Qxe2 Qxg3 30.Qf3 Qxf3 (if the Black queen moves, what is Black to do about 31.Na7 and 32.c8=Q?) 31.Bxf3 Rf4 32.Rc1 (in your face, Nakamura!) 32...Rxf3 33.c8Q Rxc8 34.Rxc8+ Kg7 35.Nxd6 and the position is clearly lost for Black as White's passed pawns are more advanced.

click for larger view

You ran an engine over it!

Although there is a certain amount one-sided imagine in it (especially the last bit - I did see f3) and I was just flicking the Black pieces onto good looking squares, the statement.

"He has been Laskered!" can hold water.

The position on the board after Rh3 is the one you place in the player's mind. Not one the board. Upset them, let them see ghosts.

As you say.

"Komodo apparently doesn't care about opening up even more lines in the vicinity of the White king, something that I doubt anyone of us would do!"

It's a human I'm playing and humans are.....human!

You said in a Nxe2+ and Ba6 variation which was OK for Black.

"perhaps you overlooked this possibility."

I never looked at anything, the ten minutes I quoted was how long it took me to type that post up. I just followed my gut and pointed thing at the King. It's how I've always played. I knew there would be tricks and traps on the board.

Would I have given up my f4 Knight for the e2 Bishop. I doubt it, perhaps, but I also doubt we would have gone there, my opponent would have seen that - got a fright - and started wondering what else he may be missing. 'Laskered!"

Stop slandering Lasker. :-)

See my posts here

Humans playing humans, as we've discussed before, is different than humans playing computers or computers playing computers. None is inherently better than the other, just like rapid and blitz are different than playing at classical time controls or at correspondence time controls (with or without computer assistance). Not better, just different. And different strategies are better in some situations than others; optimum play results from using optimum strategies adapted to the different circumstances.

And, yes, the statement "He has been Laskered" can certainly hold water. The fact that engines can't be "Laskered" as far as I know doesn't change it a bit. And of course making things difficult for an opponent, whether it's carbon or silicon, is standard procedure. Often humans when playing computers try to steer the game into well-known situations where computers at the current time are not as strong; closed positions, fortresses, endgames where long-term planning is required that would push the winning moves beyond the engine's horizon (but below the tablebase threshold), etc. All's fair in love and war, and chess is a war.

I didn't know what possibilities you were able to see in the 10 minutes you took to write up the post, so the fact that you might not have seen 12...Ba6 in response to 16.Qxa2 is not a criticism, just an observation. Whether you saw it or not is not important. As you pointed out, there would be tricks and traps on the board.

My favorite all-time player is Tal, for obvious reasons. He might not have been the best player ever (although he might reasonably be considered to be in the top 20 and certainly in the top 50) but he's in the running for the most entertaining, both over the board and beyond the board. He knew that many of his sacrifices were not sound and said so, and many were indeed proven to be unsound afterwards, and without computer assistance. But, so what? It was a practical approach to playing chess, and he knew that there would be tricks and traps on the board that his opponent, possibly after being Laskered and with the clock running, would probably be unlikely to see. And if his opponent saw the refutations and beat him, all that proved is that he couldn't be right all the time and he knew that ("there are two types of sacrifices: correct ones, and mine"). It certainly looked good when it worked (most of the time during his prime years, and often enough beyond them) and I'm sure that he also enjoyed being himself.

I stumbled into this link while I was looking for Tal's exact quote on unsound sacrifices: http://www.chess.com/article/view/t.... I think that you will enjoy it, particularly the first game which, in a not too dissimilar position as in this game, Tal played 11...Nxd5, sacrificing his queen. And the author was amazed to find that Houdini evaluated that 11...Nxd5 was also Black's best move in that position!

I've read most of your posts on Lasker and I'm not slagging off Lasker I am paying him my highest compliment.

I often slip into names mode.

You've been Morph'd - Hammered because you never developed.

You've been Nimzo'd - Baffled by mysterious Rooks moves and Zugzwanged.

You've been Nak'd - If your name is Carlsen you have just won from a lost postion anybody else and you have just been scudded.

---

Hi AylerKupp,

I read the article.

'Tal's Sacrifices Explained'

And good luck with that GM Daniel Naroditsky.

"Feeding Tal's sacrifices into the computer is usually a thankless operation which results in disappointment rather than enlightenment."

Then don't do it.

Bobotsov could have played on as White in one of the games explained....

click for larger view

...Just to see if Tal would have under promoted to a Bishop for the mate.

You have a computer, what does it think is the best mate after Ka2.

Serious question, which one of the three does it play?

---

Hi everyone else,

Sorry for a hikacking the thread yakking about me, Lasker and Tal - back to Nakamura.

Nakamura recently said he does not care if he losses.

That's OK when he's playing solo in tournament's but would you as a team captain, and this is a team event, be happy about hearing that?

1. The sacrifice is sound, and then I admire Tal's conception and calculating ability.

2. The sacrifice is unsound but Tal gets away with it. To be able to do that consistently at the highest levels is a testament to Tal's genius and judgment, and I it does not diminish my enjoyment of the sacrifice or the game.

3. The sacrifice is unsound and Tal gets punished for it. Then I have admiration for <both> players, Tal for conceiving the sacrifice in the first place and his opponent for managing to find a refutation under pressure. And this is the result that usually pleases me the most.

As far as which of the mates in this position the computer selects to be "best" I had assumed that it would be purely random depending on the order that the moves appear in their search tree. And, chess engines being non-deterministic, I would have thought that even with the same engine it might change each time you run it.

click for larger view

Alas, it was not to be. I tried 15 engines and all but one (Spark 1.0, who chose 31.Ka2 Ra1#) chose 31.Ka2 b1Q# as their best move. Frustratingly, even though I have the default for each engine to display its top 3 best moves, each of the engines only displayed the one move.

Then I tried Ziggurat 0.22. Ziggurat 0.22 has the distinction of being the <worst> out of 1,674 engines in the CCRL 40/40 engine tournament, rated at 1744. And it did not disappoint, it indicated a mate in <two> after 31.Ka2 and then failed to display the mate that it selected. There might be hope for us humans yet.

Speaking of indecisive chess engines, I wanted to remind you of this post, Robert James Fischer (kibitz #50476), which listed a link to this post, Caruana vs Anand, 2013 (kibitz #280), referencing the game between COKO III and GENIE in 1971.

click for larger view

That above position reminded me of another short story I had published in CHESS (B.H.Wood must have been hard up for material in those days.)

It was about one computer that controlled everything on the planet...and I mean everything...from Banking to traffic lights...from electric toasters to satellites.

Some lad played it at chess (it always won.) but a position arose where the computer could mate in one with either a Bishop or a Knight but it could not decide which way to do it. (both pieces in it's binary code being worth 3 pts.)

It devoted more and more of it's resources to figuring out which way to mate it eventually shut down the whole planet.

That makes sense to me, I wonder why I didn't think of it (pleaaaase, nobody answer that!) But that then makes me wonder why Spark 1.0 didn't choose it in favor of 31...Ra1# which does not involve a promotion. Even 31...Bb1# would have given it a material advantage over 31...Ra1#.

This seems reinforced when I changed the position to

click for larger view

And indicated that it was Black's move. Now, with MPV=3, the engines listed 3 mates in one in the following order: 31...b1Q+, 31...b1B+, 31...Ra1+. All but Toga II 3.0 which listed the order as 31...b1Q+, 31...Ra1+, and 31...b1B+.

Except for Gull 3.0 and Rybka 4.1. They listed their top 3 moves as being 31...b1Q+, 31...Ra1+, and ... 31...h5 (Gull) and 31...h6 (Rybka), which did not immediately lead to mate, so the horizon effect took over and White played 32.Qxb2 (!). They apparently missed 31...bB1+ as one of its top 3 moves! And Rybka was not even consistent; on the 2nd run it listed 31...Bc3 as its 3rd best move, back to 31...h6 as its 3rd best move on the 3rd run, 31...h6 as its 3rd best move on the 4th run, and back to 31...Bc3 as its 3rd best move on the 5th run. I had neither the heart nor the patience to run more than 5 analyses.

And, good old Ziggurat. It does not support MPV so it listed 31...b1B+ as its mate of choice. But at least it found the mate!.

Indeed there is.

<....And I suspect that there are about 30 people on this site that will disagree with you.>

That's an incredibly silly statement. If anything, it's the best, and easiest way to meet 9. b4.