Both the attacking (going towards the enemy king) and forcing (deliver a check, restrict your opponent's replies) nature of this move make it an obvious candidate move for players with even minimal experience.

click for larger view

Now, after 22 Bxe3 fxe3 white has 23 Nd5, providing support for his passed pawn.

click for larger view

<pgp> The move also jumped out at me because, first of all, you have a clue that the position holds a 'game changer'. The recapture on c5 does not seem to feel right - it is too boring. Otoh, the Knight fork looks exciting because you can sense the strength of the pawn on g3 as well as the follow-up with the Queen on h4. So I think the sacrifice seems obvious - the only question will be: Is it sound? With the benefit of hindsight (and Rybka ;-)), we know it is.

click for larger view

Great question. I'm no chess wizard either, not even close. I saw the move very quickly probably because my coach teaches me to look at checks, captures, and threats (usually in that order). But even if you haven't been formally trained to do so, many of these chess problems have a forcing move as the solution. In that case players become self-trained to look for crazy looking moves usually when confronted with a puzzle, whereas over the board they may not even consider such moves.

I think after seeing puzzles for a while you begin to see similar patterns where pieces are sacrificed or there are loose pieces or the king seems vulnerable to attack and the puzzle illustrates how to take advantage of that. Our brains are mostly geared visually (so I've heard), though I'm no expert on that either.

I saw that 19.hxg3 fxg3 looked bad for white with ...Qh4 looming. So, I looked at declining the sac with 19.Kg1 Nxg1 20.Nxb7 (getting a pawn for the knight that I'm about to lose anyway) Bxb7 21.Kxf1.

Somehow, I mentally concluded that white ended up with 2 pieces for a rook, which didn't sound so good for black, so I dismissed it.

At the end of your A.1.c), I believe black's Q is hanging. Instead, black has 26... Bf5#, a line I intended to include in my post, but somehow omitted.

<WhenHarryMetSally>

A master will quickly choose 18... Ng3 instead of 18... fxg3, because (1) it's more forcing and (2) after the exchange,the pawn on g3 controls the king's flight square f2, but the knight on g3 does not. Of course, you need to visualize the basic mating net and analyze the only feasible defense (Re1) a bit further before sacrificing a 2nd piece.

<Patriot> <[snip]Because once you clearly establish a best move then it's a complete waste of time trying to figure out everything that "might" happen.>

I agree with your essential point that a number of the solutions present more detail than is necessary to establish the best move. For instance, in this game, if I visualize and assess correctly up to move 22... Bh3, I see that I will get my material back with interest and I think that would be enough for me to decide to go for it OTB. However, we also know that our experience and biases can lead us to a bad assessment of what is actually the best move. A chess game is like a minefield - the more you can see, the better off you are! In any case, exploring the detail in these puzzles, as you note, is great for extending one's board vision.

<NewLine:> <Oh come on! This isn't a Medium puzzle, and hardly forcing (I mean, for regular not-rybka humans)..>

Well, if you find the first move and the knight is accepted, the continuation is pretty much forced up through black's move 22. If you've seen that much and you realize that black is winning, you can credit yourself with solving it, even if you didn't succeed in guessing how the entire game played out. Bear in mind that many of the kibitzers post solutions based on the diagram alone, without consulting engines or moving pieces. This can be done with pattern recognition, experience, and practice.

If you think this was difficult, try out a few weekend puzzles!

<Jimfromprovidence> (& <agb2002> )

Chess is sometimes a pictorial game. We deal in images, patterns, visual cues. In some cases, we spot a combination because we have seen something like it before. Sometimes, the arrangement of pieces suggests a continuation - eg a queen and king on the same diagonal cries out for a bishop to pin or skewer.

In today's puzzle, we need to use one of Jeremy Silman's maxims, that you should attack in the direction that your pawns point. If you look at the starting position, nearly all of black's pieces and pawns are pointing at the white kingside. It's like the starting position has a massive flashing neon sign pointing towards g3 and h4 - both squares near the white king and both squares attacked by two black pieces.

With clues like that, who could fail to be attracted by a move like Nxg3+??

The other clue for those with a little more chess knowledge is that the opening was almost certainly a kings indian, which means that white attacks on the queenside and black attacks on the kingside.

Black is a knight down, but can immediately restore material parity by capturing on c5.

Obviously there is more in the position than this.

Straight away I see the line

18......Nxg3+ 19. hxg3 fxg3 followed by 20..... Qh4+ and 21.....Qh2 mate.

What defence can white put up against this?

On move 19 the only other legal move is 19. Kg1, whereupon 19.... Nxf1 20. Bxf1 dxc5 and black emerges the exchange and a pawn up.

On move 20, white has no reasonable move to prevent Qh4+ / Qh2 mate except 20. Kg1. After this move black continues 20....Qh4, and white must move the rook (Re1) to avoid immediate mate on h2.

Now black continues 21.....Qh2+ and white 22. Kf1

Now what? Clearly it must be 22.... Bh3 threatening 22.... Qxg2 mate.

Only 23. Ke2 stops this (23. Re2. Qh1 mate)

Then 23.... Qxg2+

I am now visualizing the position 5 moves ahead, so I hope I don't make mistakes.

24. Kd3/Ke3 Rxf3+ and black can play ...dxc5 with a won position (3 pawns up, one being a passed pawn on the sixth rank surrounded by major pieces).

This must be the solution.

They are excellent training if you want to develop your tactical chess ability. Not so much when it comes to your positional chess ability.

As you pointed out, almost all of these positions involve a combination to bring about the desired result.

Yes, I do see that this is very good practice for visualizing combinations, etc. It might be more beneficial, though, if there were some positions in which the tempting sac does *not* work, and you have to find the refutation. But it would be a lot less fun to settle for the quiet move that, as Dzechiel says, pays off many moves later.

Niebult vs Tal, 1954 (18...?)

Black to play and win.

Material: Down a B, with the possibility of the immediate recapture 18…dxc5 (but this is Tal!). The White Kh1 has 1 legal move. Black has 18…Nxg3+ or

18…fxg3 (threatening 19.Qh4 20.Qxh2#)

White apparently must delay rescue of Nc5, while Black opens the White K-side. The Black Kg8 is immune to check.

Candidates (18...?): fxg3, Nxg3+

[18.fxg3 Ne4 19.Qh4 h3 holds]

In fact, a Pg3 is better than Ng3, because it controls the (now) weak dark squares around Kh1.

18…Nxg3+

(1) White can accept the sacrifice of Ng3:

19.hxg3 fxg3 (threatening 20…Qh4+ 21…Qh2#)

20.Kg1 (now, or later by transposition) Qh4

(1.1) 22.Re1 Qh2+ 23.Kf1 Bh3 24.Re2 [Bxh3 Qf2#] [Qe2 Qh1#]

24…Rxf3+ 25.Rf2 Qxg2+ 26.Ke1 Qxf2#

(1.2) 22.Bh3

Black has at least 22…Qxh3 23…dxc5 (if nothing better), winning a P, with the initiative.

(2) White can decline the sacrifice of Ng3:

19.Kg1

At the end of your A.1.c), I believe black's Q is hanging. Instead, black has 26... Bf5#,..>

<18....?>

Target: 3:30;000
Taken: 1:50;058
Eagle: -1m <> -2m

Material: White up, ♘

Candidates: <[Nxg3†]>

-ML-
First and foremost, I missed Black was down a Knight. That changes the combo slightly, as it actually has to gain back more than just the piece sacced on move 18, also the piece that was sacced earlier. The keymove is obvious whatsoever:

<18....Nxg3†!> without this demolition, Black is losing. Interesting is that Kg1, avoiding the sac, yields a comfortable position for Black with a Rook against two Knights, where the Rook is definitely better (the f8♖ vs the White ♘s), nonetheless, it might be best, what if White does accept?

/A\
<19.hxg3 hxg3> many options are open to White to avoid 20....Qh4† 21.Kg1 Qh2‡, but not all are as sufficient. Blocking the h-file with the Bishop is one:

/AA\
<20.Bh3> aside putting a Bishop in completely, Black can still get his Queen into play with some terrible consequences for White:

<20....Qh4 21.Kg1> White cannot protect the Bishop.

<21....Bxh3 > and the only one defending is White.

/AB\
<20.Kg1> this is a lot more simple to defend, and in truth the only true defense.

<20....Qh4 21.Re1> the Rook must move, Bh3 leads to the same position as above.

<21....Qh2† 22.Kf1 Bh3!> now the threat is simple: 23.Bxh3?? Qf2‡, so White has to defend the Bishop. 21.Qe2 leads to 21....Qh1‡, so White may only play:

<23.Qd2> but the pin remains:

<23....Qh1† 24.Ke2 Qxg2† 25.Kd1 Rxf3 > and I guess White is losing there quickly. So, playing 19.Kg1 might be the safest;