Comments on: Compulsion

By: Steven

Steven — Fri, 16 Nov 2007 13:50:22 +0000

Me neither!

Update: with the help of computer analysis, it seems that d4 Qxe4 is in fact merely a draw.

By: ejh

ejh — Fri, 16 Nov 2007 13:31:10 +0000

IMO would be a dubious OTB decision

Well if I saw Qh6 and its consequences I’d go for that instead, but somehow I don’t think I ever would…

By: Steven

Steven — Fri, 16 Nov 2007 10:30:43 +0000

Well, like I said, it’s a practical mess and IMO would be a dubious OTB decision. ;) (I would like to be more convinced by the inevitability of your second diagram too.)

By: ejh

ejh — Fri, 16 Nov 2007 07:59:07 +0000

“some tedious checkfest where his only advantage is king position and doubled pawns vs singleton on d-file.”

I don’t think that happens: I think White can win the Black d-pawn in return for one of his own and should then win the resulting ending with the advanced passed pawn. More on the first diagram.

By: Steven

Steven — Fri, 09 Nov 2007 15:19:59 +0000

No, I had already looked at d4 Qxe4 and discounted it, I can’t believe that white wants to give up his centre and go into some tedious checkfest where his only advantage is king position and doubled pawns vs singleton on d-file. I would say black has pretty decent practical chances to draw that.

Qh6 is nice computer move, though. ;) FWIW, my engine abandons d4 very quickly and also settles on Qh6.

By: ejh

ejh — Fri, 09 Nov 2007 13:54:53 +0000

Well, he can play 1.d4, after which 1…Qxe4 2.Qxf7 and although I don’t see a clear win yet, Black’s not got perpetual and I think White ought to have enough (though a stronger player, knowing more about queen endings than I do, might have a different opinion and would certainly have a better-informed one). That was what I saw personally after I started wondering about the initial diagram.

However, what the computer then saw, after I consulted it, is 1.Qh6!! Now if 1…Kd8? 2.Qf8+ is easy, while if 1…f5 2.Qg6+ (not 2.Qe6??? which loses) followed by 3.exf5 and the pawn will promote. 1…f6 is better but 2.Qg6+ Kd8 3.Qg7 Qe5 4.Kf7 wins the f-pawn and although Black can have the a-pawn, White will win the d-pawn while being sheltered from checks. Finally if 1…Kd7 2.Qf8! and the point to grasp is that after White’s king takes the f-pawn (after, say, 2…Qg5+) Black won’t have a perpetual because White will put his king on the eighth rank and then interpose with the queen.

I don’t recommend trying to follow this without a board.

By: Steven

Steven — Fri, 09 Nov 2007 09:33:07 +0000

A very fair account. Out of interest, if it is white’s move in the first diagram, how does he plan to win, avoiding perpetuals?

By: ejh

ejh — Fri, 09 Nov 2007 08:23:39 +0000

The posting is here.

As a matter of interest, even in the first of the two diagrams, the position would be won for White were it his turn to move. I still think it’s zugzwang, because Black is obliged to abandon his defensive post not by threats but by the fact that it is his turn to move: but it’s not, perhaps, a perfect zugzwang.

By: Steven

Steven — Sat, 03 Nov 2007 10:47:12 +0000

Aha, thanks. To be fair Bennett’s prose itself describes the positions correctly: “You’re in zugzwang, Reuven,” comes after 50 a4 when it is actually zugzwang, and after Kg7 the narrator just says “the game was mine”. Evidently something went drastically wrong in the captioning stages (I didn’t follow the version that ran in the Observer, so don’t know if it was like that originally: you’d think if it had been, some reader would have pointed out the mistake). Look forward to your blog post on the matter (do provide a link here).

For the benefit of other interested readers, here are the two positions in question, at left after 50 a4, at right after 52 Kg7. Both are described as positions of zugzwang in the captions on p.261, but as ejh points out, while the first one is zugzwang (black is forced to abandon the double defence of f7 and f8; and the alternative …f6 doesn’t help after Qxe7 Kxe7 Kg7; but if black could pass the move then white couldn’t make progress), the second position is not really zugzwang but just normally lost:

The full game score is:

[Event “SUI-chT”]
[Site “Switzerland”]
[Date “2000.??.??”]
[Round “0″]
[White “King,Daniel J”]
[Black “Sokolov,Andrei”]
[Result “1-0″]
[Eco “B25″]
1.e4 c5 2.Nc3 Nc6 3.g3 g6 4.Bg2 Bg7 5.d3 d6 6.Nge2 e5 7.h4 h5 8.Nd5 Nce7
9.Nec3 Nxd5 10.Nxd5 Be6 11.c4 Bxd5 12.cxd5 Bh6 13.b4 Bxc1 14.Rxc1 b6 15.Bh3 Nh6 16.Qd2 Kf8
17.0-0 Kg7 18.f4 exf4 19.Rxf4 Re8 20.Qb2+ Re5 21.bxc5 bxc5 22.Rxc5 g5 23.hxg5 Qxg5 24.Rc2 Kh7
25.Rg2 Rg8 26.Qf2 Qe7 27.Rf6 Kg7 28.Rf4 Kh7 29.Bf5+ Nxf5 30.Rxf5 Rxf5 31.Qxf5+ Kh6 32.Qf4+ Rg5
33.g4 hxg4 34.Rxg4 Kh5 35.Rg2 Rxg2+ 36.Kxg2 Qc7 37.Qf5+ Kh6 38.Qf6+ Kh7 39.Kg3 Kg8 40.Kh4 Qb6
41.Kh5 Kf8 42.Kh6 Ke8 43.Kh7 Qc5 44.Qg7 Ke7 45.Qg5+ Ke8 46.Kg8 Qc7 47.Qh6 Qe7 48.Qg7 a6
49.a3 a5 50.a4 Kd8 51.Qf8+ Qe8 52.Kg7 1-0

By: ejh

ejh — Sat, 03 Nov 2007 09:49:11 +0000

This database does the job. It does incidentally point up another error, albeit a minor one: the listing “King-Sokolov” (enter those two names in the White and Black fields) isn’t quite adequate because there were two active Sokolovs at the time. An initial (A, in this instance) should have been appended.

The database shows that the real and fictional games finished at the same point. There would be several ways to win if it were White’s move: taking the pawn on f7 does the job, as does the plan you propose, or simply exchanging and playing d4, since if …Ke7 then e5.

It’s absolutely not a position of Zugzwang. The previous diagram is an example of zugzwang, but this one is not. It’s just a won position, it has no special features.

Were the diagrams incorrectly rendered when the piece was running in the Guardian? Whether they were or not, that they came to be printed in the state they were beggars belief.