In the above recent game, the pure Bishops of Opposite Colors endgame began in this position after 50...Bxf5

click for larger view

In this Nimzo vs Capa game, the pure Bishops of Opposite Colors endgame began in this position after 29.. Bxg4

click for larger view

I am astounded to read Capa's notes: <The ending is now won by force.>

He knew this ending was won even before he got into it, but how!? What makes it different from the hundreds of drawish BOOC endings that we see in books and live on the internet? Visually it looks like a draw. Nimzo certainly thought it was drawish and could not accept that he had lost.

After serious analyse it seems to me that <30.a4?> was the decisive mistake. With <30.b8> Nimzo would have had very good drawing chances.

Here is mine :
30. b8 forcing black pawn to advance 30...a6

(30... a5 31. d6 c4 32. c7 a4 33. e1 e6 34. d2 d5 35. b6 e4 36. d4 f5 37. g7 h5 38. h6 g4 39. e2 d5 40. e3 The c2 pawn is not useful, White has to block Black. 40...xc2 41. a3 e5 42. f4+ f5 43. c7 h4 44. d8 g3 45. h3 d3 46. xh4 g2 47. f2 =)

31. a7 forcing another pawn to advance 31...c4 32. e1 f5 33. d2 b5 34. e3 e6 35. b6 d5 36. e3 xc2 37. a3 f5 38. d8 and it seems Black can't win.

click for larger view

30.Bb8 a6 31.Ba7 (You say <forcing another pawn to advance>). I say, no, 31...b5 32.Bxc5 Bf5, with an extra Pawn.

If we follow your second line, 30.Bb8 a6 31.Ba7 c4 32.Ke1 Bf5 33.Kd2, I prefer 33...Ke6, instead of the weaker 33...b5. After that, Black’s King moves to e4 after ...Kd5. If Black tries to prevent this with 34. Ke3, then ...Bxc2. If, and, there are other lines, 34. Be3 Kd5 35. Bb6 Ke4, Black wins.

33...Kg6 could also be played, with the idea of ...Kg6-h5-g4-h3, advance the Pawns, and Black may be forced to give up the Bishop for the extra Pawn.

Did you read what I've written in my comment : <The c2 pawn is not useful, White has to block Black.> ? 2 times you end your variations with "and Bxc2 wins a pawn", winning c2 doesn't mean anything!

You said <30.Bb8 a5 31.Bd6 b5 32. Bxc5 Bf5 and ...Bxc2, with an extra Pawn.>

The white King goes to e3 and stays here. The white Bishop goes to the b8-h2 diagonal. If black King goes e6-f5-g4-h3, white simply answers Kf2 and it's a draw.

You said <30.Bb8 a6 31.Ba7 (You say forcing another pawn to advance). I say, no, 31...b5 32.Bxc5 Bf5, with an extra Pawn.>

Same position as above, with pawn on a6 instead of a5, it doesn't change anything. It's a draw.

You said <If we follow your second line, 30.Bb8 a6 31.Ba7 c4 32.Ke1 Bf5 33.Kd2, I prefer 33...Ke6, instead of the weaker 33...b5. After that, Black’s King moves to e4 after ...Kd5. If Black tries to prevent this with 34. Ke3, then ...Bxc2. If, and, there are other lines, 34. Be3 Kd5 35. Bb6 Ke4, Black wins. 33...Kg6 could also be played, with the idea of ...Kg6-h5-g4-h3, advance the Pawns, and Black may be forced to give up the Bishop for the extra Pawn.>

The positions after ...Bxc5 and ...Bxc2 are not blockades. There is plenty of mobility and fight here, although drawish.

<The white Bishop goes to the b8-h2 diagonal. If black King goes e6-f5-g4-h3, white simply answers Kf2 and it's a draw.>

You are right. This is what happens when I analyze blindfold. I simply forgot about the Bishop being on the dark squares. Inexcusable.

<If you look at the line I suggested with 33...b5, Black win the c2 pawn but it's not useful to win the game. It's the same with your 33...Ke6 or 33...Kg6, I play Ke3, give you the c2 pawn, and draw.>

I said that 33...b5 is weaker than 33...Ke6 and I meant it. Black has much better lines at his disposal.

As an example, by not playing ...b5, I give myself some wiggle room on the Queenside. A sample: 30. Bb8 a6 31.Ba7 c4 32. Ke1 <Let me ask you something. Why play 32.Ke1 Bf5 33.Kd2 b5 (or my ...Ke6) 34.Ke3 Kd5, when you could have played 32.Kf2 and 33.Ke3, which is faster?> 32...Ke6 33.Kd2 Kd5 34.Ke3 Bxc2, and now the idea is to infiltrate on the Queenside with the King via c6-b5-a4. It is a harder struggle on the Queenside, but I have gained some tempi by not having moved ...b5 earlier.

I really liked studying this endgame over the weekend. It is a very uneasy affair, but I liked Black's chances on the Queenside. I think I am going to work on this some more. I did like how three Pawns blockade against five in your line, but your line is not forced (...b5 is not forced), and therefore, we cannot consider it an absolute draw. There is till a lot of play here.

Nowadays it would be the choice of all GM. It's well known that Capa was not searching for best moves in the opening.

<TheFocus> : I agree that Kf2-e3 is faster but there is no need to be faster because it's already a draw with this. I agree that not playing b5 is better but I think it's a draw anyway.

A couple of wrong ways to play it with white are:

a) 4.... Nxe4 5. Bxf7+ Kxf7 and black will shortly have a dominating central position.

b) 4.... Nxe4 5. Nxe4 d5 6. Bb5 dxe4 7. Nxe5 Qd5 and black is better

Instead, white can play 4... Nxe4 5. Bd3 dxe4 6. Bd3 dxe4 7. Bxe4 and work diligently towards his draw. It's true that black might win anyway, but the simplified position offers less prospects of this win then the complex position after 4.... Bc5.

He knew this ending was won even before he got into it, but how!?>

just because cap a says it doesn't make it true! remember, he had no tablebases or progs to analyze games with.

the fact is capa was simply wrong when he said the ending is now won by force. after move 29) the game is a classic (boc) draw, capa just didn't know it, and nimzo wasn't good enough to prove capa wrong.