How to improve my early endgame / late middlegame?

Question

When I search for books on the middlegame versus the endgame, I usually find discussions about positions that have almost nothing to do with each other: the former being rich with strategies with upwards of five pawns on each side, and the latter being more precise and cut-and-dry, rarely with more than one or two pawns if there's a valuable piece on each side. Discussions of the latter sometimes include vague advice along the lines of "try to reduce to a winning cut-and-dry endgame" to extend the complexity a couple pieces/pawns more. Yet when I reach a point in the game like this:

6k1/5p1p/6p1/2Np1b2/1r1P4/7P/5PP1/3R2K1 w - - 5 28

I feel like I'm playing on nothing more than shallow tactical threats sometimes colored by old adages like Tarrasch's rule, the principle of two weaknesses, bishop and pawn colors, open files, etc. Can anyone recommend me sources of good theory/practice for this stage of the game?

Answer 1

It sounds like you are looking for strategic guidelines in non-theoretical endgames. Then a modern classic is Endgame Strategy by Mikhail Shereshevsky. Strong players tend to like this book, which should tell us something, but I'm not sure exactly what (causation? correlation?).

Another decent choice is Practical Endgame Play - beyond the basics by Glenn Flear. It focuses on so-called NQEs (Not Quite Endgames) where each side has a couple of pieces each, not counting kings and pawns. The chapter on Rook and Bishop against Rook and Knight contains 50(!) fragments from real games with this material balance, the one in your example. At least it should give you plenty of study material.

Answer 2

There exist all kinds of position from that stage of the game, so it is difficult to give general advise. But it certainly holds, that if you cannot gain any advantage out of the middlegame, it is not likely that you will win an equal endgame unless your opponent cooperates. That's why books focus on middlegames with many pieces or endgames with few pieces.

Studying endgames first is not a bad idea as you will learn what kind of positions can be won and which positions are draw. So you know what to aim for in the middlegame. It is usually a good idea to transfer into the endgame if:

• you have decisive material advantage (which is often more easily won with few pieces on the board)
• if your opponent has weaknesses in the pawn structure which can be attacked
• if you would end up with more active pieces

Regarding your example: If there is no immediate tactics, it looks very drawish to me. The only way to potentially win this for either side is by winning the d pawn. Starting from this you can think of plans for white and for black. My line of thought would be:

White plan: Obviously white needs to free his rook from defending the d4 pawn. The only way to try to do this is by defending it with the king from e5 (too dangerous), e3 or c3. This might not be easy because black can check with the rook from c3 and c2. So it could be worth trying to get the black bishop from the b1-h7 diagonal which would give white also the square d3 for the king or perhaps another piece to defend against checks.

But even when white frees the rook, it is not clear to me how to proceed. Black can just centralize the king to f6 or d6, put the bishop on e6 and a pawn on h5 and wait.

Black plan: Since black has a light-squared bishop, the only way to win the d4 pawn is with the help of the black king. So black wants to walk the king to c4 or c3 basically. This is a serious threat and white needs to think of how to prevent it, either with active counter-play on the king side or cutting-off the black king or defending passively.

Knowing these plans you can come up with concrete moves. Also it would be worth keeping in mind what endgames can result from this position. E.g. what about if the rooks are exchanged, or if knight/bishop are exchanged or if you end up with an endgame (e.g. after a piece had to be sacrificed against a pawn and the f-g-h pawns were exchanged) R+B vs R or alternatively R+N vs R. This will help you in coming up with moves as you will know what to aim for or what to avoid.

I don't know what this position resulted from, but if it was a kind of good knight vs. bad bishop position, white should have tried not to exchange pieces.