GM Joel on Judging a Sacrifice Print E-mail
By GM Joel Benjamin   
July 30, 2009
Hi, Joel,

My name is Victor Lendermon and I'm writing to you about a position from the January 09 issue of Chess Life. I've analyzed this, skittled it, and had a long distance "lowly" master friend look into it as well, but without any definitive answer.


The move Bf2 was played in the position above with the notation that White should play Nd3!? and be happy. Why can't White play Nxe6 and be even happier? I've only explored two main lines here and I'll just give you the initial moves without my thoughts.
21.Nxe6  fxe6
22.Bxe6  R5-g7
23.Bxg8  Qxg8
21.Nxe6  Bc8
22.Nxg5  Bxh3

ARRRGH!!!! Any insight would be much appreciated!
In both lines, it seems like White has ample compensation, much better pawn structure, and nice attacking chances. The Black sac on e4 needs watching, but I must be missing something.
Once again, if I've sent this to you in error, please accept my apologies. I really enjoy analyzing positions like this one...especially if there seems to be a tactical solution.

Thanks in advance, Victor.


The game in question is Antipov-Adelberg from last November’s World Youth Championship. (Login and read GM John Fedorowicz's article, Bronze for Shankland and Yang at this link.) David Adelberg made a nice showing for the USA, winning this game from round eight quite nicely.

The implication of the comment you cited is that White should be happy to repeat because a draw would be a good result from this position.  Like you, I’m not so sure.  White is a pawn ahead (Black sacrificed the g-pawn earlier) and has his king fairly well protected, while Black’s king is not entirely safe.

I consulted a specialist (Rybka 3), who claimed White is quite a bit better. [I know that you don’t need a grandmaster to turn on a chess engine, but Rybka + GM Joel gives you a better answer than just one of us]  Rybka recommended 21.a4 b4 22.a5 and gave White more than a pawn plus…until it noticed 22…h4! with the idea 23.Qxh4 Nxe4 24.fxe4 Rh8.  [Rybka’s second move, the more defensive 21. Ng2, might be the best option.]
This is the kind of position where computers set a trap for us.  I’ve seen it manifested in a lot of amateur analysis, where the recommended line is the same one I found on my computer for 30 seconds or so…until the computer realizes it doesn’t work.

I think that White is objectively better, if you want to assume perfect play.  In a practical game situation though, Black should be happy.  It’s easier to make a mistake defending than attacking.

All this preamble is important because your move 21.Nxe6 doesn’t win but has no obvious refutation either.  To decide if White should play it, we need to compare it to the alternatives.

Your second line with 21...Bc8 is unplayable for Black. 

Position after 21...Bc8

White indeed has good compensation for the queen at the end of your line, but even better is 22.Bxg5! Rxg5 23.Qh4 (Here 23.Nxg5 Bxh3 24.Nxh3 probably wins as well) and White wins.

At the end of the first line, my computer likes White a lot, but cuts its score in half after a few minutes thought.  This is because the dynamic factors favor Black.  White is ahead in material, but Black has all the attacking chances.

There are other options for Black, of course.  22…h4 is interesting, with these variations:

Position after 22...h4

A)    23.Bxg5 Rxg5 24.Qxh4 Nh5 25.Qh3 Qf8 26.Bg4 Nf6 27.Bf5?! d5 28.e5? Rh5 winning for Black
B)    23.Qxh4 Nxe4 24.fxe4 Rxg3+ 25.Qxg3 Rxg3+ 26.hxg3 Qg6 27.Bh3 Qxg3+ 28.Bg2 Bxe4 29.Ba7+ Kxa7 30.Rxe4 Bd8 with definite advantage to Black
C)    23.g4 (best) Nh7 24.Bxg8 Rxg8 25.Qg2 Ng5 with compensation for Black

I have to conclude that 21.Nxe6!? probably speeds up Black’s attack.  This is not surprising, since it creates a situation where Black now has more attackers than defenders, while there are too many pieces on the board for White’s pawns to be an immediate factor.

Ask Joel a question on anything from openings to endgames at [email protected] .