15 Jan 22
In the game Euwe-Averbakh (Zurich Candidates, 1953), the following position was reached after 25...Ne7xf5.
At this point, White played 26. Rae1.
In his book on the tournament (translated by Oscar D. Freedman), Bronstein writes, "Stronger is 26. Re5 first, tying Black to the defense of his d-pawn and considerably impeding his drive toward the endgame."
On 26. Re5, Black might consider 26...Qg6, being that the attack on White's queen defers the capture of Black's d-pawn for at least one move. And if White exchanges queens, 27...hxg6 would attack White's knight. (See next chess movie.)
After 26. Re5 Qg6, White could instead protect the queen by 27. h4 (for example). The decision of whether to play 27. Qxg6 or to protect the queen will to some extent depend on whether White stands to benefit from the doubling of Black's kingside pawns.
For guidance on this question, let's fast-forward to the position reached after 31...K(h8)-g8. (Next diagram.)
Here Bronstein writes, "It is necessary for White to transfer his knight to d2 or b1, keeping his g-pawn on g3. He can accomplish this plan at the very moment that Black completes his preparations for the breakthrough on the Queenside: 32. Nf4 Kf7 (else 33. Ne6) 33. g3 Nd6 34. Ng2 Nb5 35. Ne3 Ke6 36. Nf1 Nc8 37. Nd2." (See next chess movie for this sequence.)
(White's knight maneuver is intended to bolster the a3 and c3 squares and thereby inhibit the sacrifice ...Nxa3 followed by ...Nb5 and the advance of the queenside pawns, which occurred in the game.)
But according to Averbakh, Black should meet 32. Nf4 Kf7 33. g3 by 33...g5 34. Ng2 Ng6, and answer 35. Ne3 by exchanging knights, moving the king to f5, and preparing a kingside breakthrough. (See next chess movie for this sequence.)
Being that 37. g4 (the only way to control the f5-square) would doom the bishop after 37...Nf4, 38...Nd3, and 39...b2, I will take the liberty of depicting Black's king as already at f5 (and many of the remaining diagrams will assume that White "passes," being that the bishop must restrain the b-pawn and White's king is confined by the pawn structure and Black's king).
Not having personally read Averbakh's suggestion verbatim, I was left to ponder how Black might achieve a winning kingside breakthrough. One possibility is for Black to advance the h-pawn to h4 to bite against White's g-pawn, which I've noted must continue to guard the f4-square. (See next diagram.)
Having created tension against White's g-pawn, how does Black amplify it? Manueuvering the knight to h5 would allow a fork by g4+. Even if that fork would absent, White could hold the g-pawn by playing Kf2.
It seems that Black needs another target: either a pawn or a square. The e4-square qualifies, being the stepping stone to a potential infiltration by Black's king.
The desire to weaken the e4-square suggests ...g4, but doing so immediately would permit the immediate liquidation of both of Black's kingside pawns (by fxg4+ and gxh4). So, Black should preface ...g4 by ...h3, not only releasing the tension but creating latent threats owing to being two squares from promotion. (See next diagram.)
Now, it seems Black is ready for ...g4. (See next diagram.)
If White maintains the pawn tension, Black can maneuver the knight to attack the f-pawn a second time (such as from g5) and therefore force White to resolve the tension. If White plays f4, then Black's knight can deploy to h5 to threaten ...Nxg3 (because hxg3 would lose to ...h2). White's king would then have to abandon the e3-square, allowing Black's king into e4. (See next diagram.)
To complete his preparations, Black would maneuver the knight to f5, where it hits g3 while blockading the f-pawn. Then the harvest of White's queenside could proceed. (See next diagram.)
Returning to the position where Black has played ...g4,
if White exchanges, Black's king naturally would recapture. (See next diagram.)
Black would prefer that g4 be occupied by the knight, where it would attack the h-pawn and free Black's king from blockading the g-pawn. This can be accomplished by maneuvering the knight to f6 (for example), withdrawing the king to f5, and putting the knight on g4. I'll assume that White's king has gone to defend the h-pawn in anticipation of this plan. (See next diagram.)
Let's consider a slightly different ending, where Black's kingside pawns are doubled and it's White's move.
Would this alter the outcome?
It seems that White could have set up a kingside fortress, starting with h3 (to be able to capture on g4 without losing control of the e4-square). If Black maneuvers the knight to f5, White's king can go to f2 so as to control both e3 and g3. If Black's king "tempos," then White's bishop can tempo. (See next diagram.)
Revisiting the decision that White would have faced after 26. Re5 Qg6,
one factor in favor of exchanging queens immediately is that the dislocation of Black's h-pawn would give White better chances to achieve a fortress if the game should reduce to a single minor piece ending.
To be continued...
(A list of the threads I've initiated at this forum is available at http://www.davidlevinchess.com/chess/RHP_my_threads.htm .)