Monday, February 03, 2014

A Theoretical ISE

Black to play
Emms - Williams, 4NCL 1999

How good are your exchange sacrifices?  A chance to find out today.

… Rxf3 is a standard sacrifice, yes, but here it's more than that. Not just a thematic sacrifice, but a theoretical one. Not just the sort of thing that you do in this kind of position, but exactly the recommended move in this precise position. Recommended by the GingerGM in The Killer French, that is.

Anyhoo, the point is Black's going to sacrifice the exchange. What I tried to do was take a look at the starting position as if I had reached it in the game. I set up a board and analysed without moving the pieces. I knew that taking on f3 was the theoretical move, but could I convince myself to do it?

Feel free to give it a go and compare your thoughts to mine.

Black to play

First things first: I know that White doesn't have to accept the gift of Black's rook and can go for 17 Bxh5 instead of gxf3, but I'm ignoring that for now. I'll come back to it in a couple of weeks.

So, 16 … Rxf3, 17 gxf3 Bxh2+, 18 Kh1 brings us to the diagram. I made a mental note not to forget about the possibility of the king coming to g2 and ploughed on. I was pretty sure the recommended move at this point was … Nf4 so I concentrated on that.

18 … Nf4
19 Nxf4 Bxf4

I found it hard to calculate beyond this point, there being no more exchanges. I thought I remembered Williams' analysis stopping here and concluding that Black had compensation.  Although it seemed a bit cackhanded to me, I had hopes for 19 … Qxf4 at first based on 20 Bg3 Bxg3, 21 fxg3 Qxg3, 22 Rg1 Qh3 checkmate but then I realised it was all nonsense, not the least of the reasons being that the White queen defends along the second rank after the pawn recaptures on g3.

18 … Nf4
19 Kxh2 Nxg6+

My first thought was that White was forced to go for
20 Kh3 e5+
21 Kg3 e4+
dropping the bishop

but after checking some other lines it eventually dawned on me that White had the infinitely better

20 Bg3 Qf7

As before, I found it difficult to analyse further than this and I wasn't convinced I much liked this position.

18 … Nf4
19 Kxh2 Nxe2

So I tried taking on e2 instead. Now if

20 Kh1 Nexd4

and I've bagged another pawn, so

20 Bg3 Nxg3
21 fxg3

I wasn't convinced I wanted to be undoubling White's pawns, so I looked for something else and realised that after 20 Bg3 I had the possibility of trading queens with

20 … Nexd4
21 Bxc7 Nxc2

You will have noticed - although I did not - that there's a problem with this idea.

18 … Nf4
19 Bg3

By now the variations were starting to swirl around in my head and after

19 … Bxg3
20 Nxg3 Nxg6
21 Qxg6

I stopped analysing without noticing that d4 was hanging. Also, it didn't occur to me to think about 20 fxg3.

18 … Nxf4
19 Ng3

Emms' choice in the game against Williams, although I very nearly forgot about it. I didn't fancy the look of

19 … Nxg6
20 Qxg6 Bxg3
21 Bxg3 Qf7

although it did eventually dawn on me that … Qf7 was impossible anyway and I tried … Qe7 instead.

Still, I knew that Williams had chosen … Qd6 on move 19. A TN at the time I believe.


19 … Qd6

and missing Emms' 20 Rad1 I only looked at,

20 Kxh2 Nxd4
21 Qb1 Nxf3+
22 Kh1 Nxg6
23 Qxg6 Nxh4

which I felt worked out rather nicely for me.

And that's where I left it. You'll have noticed that I completely forgot to go back to check 18 Kg2 right at the beginning!

So there you have it. Hardly a neat tree of analysis. Even though I was familiar with the position and recalled some of the analysis from the DVD, and even though I wasn't under tournament conditions   so I could write the variations down in tables as I went along, it still all sprawled completely beyond my control.

Leaving aside any errors of calculation (Line B1 in particular), my ability to play … Rxf3 in a real game would have rested on being able to evaluate Black's compo at the end of Lines A and B properly. That's something I find difficult starting from here

let alone three or four moves in advance.

Still it's only February. Plenty more ISEs to go yet.

2014 Exchange Sac count: 7
TISE Index


Anonymous said...

Another way of looking at this would be to treat .. Rxf3 as a positional idea. You need to know that it sets off lots of complex tactics, but based on the Emms-Williams game, these fizzle out to an ending where White will have an advantage of the exchange for a pawn, but perhaps not enough to play for a win. Would anyone want to take on the ending against a grinder though?

Many chess engines appear to consider the worth of a rook to be about four and half pawns rather than five. As a consequence they look favourably on exchange sacrifices that win a pawn and gain initiative or structure.


AngusF said...

Jonathan set this as a puzzle for me a few weeks ago and I never got back to him about it.

I looked at it for a few minutes and couldn't believe that Black would get play. Then I switched off.

One thought I recall: what's wrong with 16... Rxf3 17 Bxh5?

Jonathan B said...

@Angus: "... what's wrong with 16... Rxf3 17 Bxh5?"

- nothing at all. Last I heard it was the main continuation, in fact.

We'll be coming back to it in a few weeks.

@Roger: fizzle out to an ending where White will have an advantage of the exchange for a pawn, but perhaps not enough to play for a win

I was planning to be at the Chess Shop last weekend anyway, so while I was there I bought the book that John Cox referenced last week. I haven't had the opportunity to study it at all yet, but I did have a nose through the introduction whether the author writes something along the lines of exchanges sacs in the opening becoming more popular because the advantage of the rook is felt most in the endgame which is still a long way a way.

In terms of judging whether an exchange sacrifice is good or not - this kind or any other - perhaps you just have to play a few and work it out through trial and error. I don't know.