Riddle solved: Flohr was winning!

by Karsten Müller
7/28/2020 – Coming from taking first place at the Soviet Chess Championship, Mikhail Botvinnik faced Salomon Flohr in a 12-game match played in Moscow and Leningrad. In one of the games, Flohr showed the chess world how to force a win in the 2B vs 2N endgame. But did he really, or did Botvinnik simply misplay it? We asked our readers to help settle the historical debate. And they helped endgame expert GM Karsten Müller to settle the question. | Pictured: Flohr and Botvinnik in 1935

ChessBase 15 - Mega package ChessBase 15 - Mega package

Find the right combination! ChessBase 15 program + new Mega Database 2020 with 8 million games and more than 80,000 master analyses. Plus ChessBase Magazine (DVD + magazine) and CB Premium membership for 1 year!


Mikhail Botvinnik, Salo FlohrA young Mikhail Botvinnik was playing one of the top players at the time, Salo Flohr from Czechoslovakia. The latter was considered the favourite. Flohr won the first game with black, and outplayed Botvinnik in game six with the white pieces. In his book “How to Play Chess Endings”, Eugène Znosko-Borovsky says that the game contained an extremely instructive ending, which well demonstrates the concentrated power of the two bishops, and the helplessness of the two knights against them — they move in various directions on the board, without any coordination. 

We described the background of the match in our previous article. But was the position winning for white all along? We asked our readers to send their analyses, so we could settle the question. 

It turned out the handling of the ending was indeed extremely instructive. 

This time no mistake was made. Flohr’s bishops were always winning. But the analyses now are much deeper than in the old sources.

But here the general intuition that the bishops must be winning proved correct. They are indeed a powerful endgame weapon!

This endgame is indeed a masterpiece for the conversion of the advantage of the pair of bishops. At first Charles Sullivan had some doubts, but then Zoran Petronijevic proved that the bishops always win. And then Albert Silver announced that he has created a new neural network engine, which is especially trained in endgames. I of course immediately asked Albert to check this riddle and here comes his answer, which confirms Zoran's analysis:

Hi Karsten,

I ran this through the new neural network and giving it a lot of time and a lot of resources, and the eval has only been rising for White after 41. f6 (I have not analyzed to see if f6 was a mistake, so if I should, please tell me). This is still an early net as you know, but so far its endgame skills have proven exceptional, so hopefully it is not very wrong if indeed it is. ;-)

Although it flip-flops between Ke2 and Kf3 for a bit at first, it decisively selects Kf3 after 1:02 with a modest evaluation of +1.04. After over a full hour of analysis, and 65 billion nodes, the evaluation climbs to over +3, leading me to believe that after Botvinnik’s 41...f6, the game is winning for Flohr. As I mentioned before, if this is correct, but 41...f6 was the losing move, I cannot say. The mainline after 41...f6 changes a few times, so I am sending the lines it displayed as it attempted different answers for Black.


So I asked Albert to check if the position is lost anyway, and indeed it is:

I see no evidence that any other move 41 by Botvinnik would have saved it. The obvious alternative 41.Nef6 is no better.


Salo Flohr, Mikhail Botvinnik

Flohr and Botvinnik

Finally as usual comes Zoran Petronijevic’s solution:


In our replay board above there are a large number of functions you can use to really understand the game and the analysis. Recently we published a comprehensive tutorial plus video instructions which tells you about all the powerful features and buttons that make the ChessBase's replay one of the best replay experiences around.

One big advantage is that you can start an engine (fan icon) that will help you to analyse. You can get multiple lines of analysis by clicking the + button to the right of the engine analysis window. The "!" key, incidentally, shows you the threat in any position, which is incredibly useful in the case of unclear moves.

There is one more thing you can do. It is a lot of fun, but also a serious challenge: Click on the rook icon below the notation window. This will allow you the play the above position against Fritz, at your level of playing strength (e.g. "Club Player"), right here on the news page. Note that your analysis, in which you can delete, move or promote lines, is stored in the notation as new variations. In the end you will find the game with your analysis in the cloud. So nothing is ever lost.


Karsten Müller, born 1970, has a world-wide reputation as one of the greatest endgame experts. He has, together with Frank Lamprecht, written a book on the subject: “Fundamental Chess Endgames” in addition to other contributions such as his column on the website ChessCafe as well as in ChessBase Magazine. Müller's ChessBase-DVDs about endgames in Fritztrainer-Format are bestsellers. The PhD in mathematics lives in Hamburg, where he has also been hunting down points for the HSK in the Bundesliga for many years.
Discussion and Feedback Join the public discussion or submit your feedback to the editors


Rules for reader comments


Not registered yet? Register

CharlesSullivan CharlesSullivan 7/30/2020 06:58
Stockfish-11, despite deep and long searches, has not been able to find a win at Black's 32nd move after 32...Ne8; for example: 33.Bf1 Kc8 34.h4 h6 35.Kb3 Nc7 36.Bh3 Kd8 37.g4 Nf8 38.g5 hxg5 39.hxg5 Ng6 and so forth. The evaluation is +1.69, but it seems that White does not break through. However, I agree with Zoran (see Karsten's comment on 7/29/2020) that "it is impossible to give a proof" -- for now, the riddle remains open: after move 30, where did Black make his final losing mistake?
Karsten Müller Karsten Müller 7/29/2020 04:27
Zoran Petronijevic on this deep problem:
Dear Karsten,

I didn't analyze this game so deep, but i think that both players made some inaccurate moves. I agree that 31...Na4 leads to an even play. I should also add that this move was suggested by Botvinnik in his comments to this game.
I think that 29...Na4 was the best move for Black. Instead of 30.g3, White has 30.a4. Is he winning after that? Hard to say, but he has clear advantage. Also, is Black lost after 32.a4? I think he is, but it is impossible to give a proof - at the moment engines aren't able for that.

Also very interesting seems 28.a4, preventing Black's aggression on the Queen side. I am not sure that this is enough for a win, however.

Best wishes,

Karsten Müller Karsten Müller 7/28/2020 10:49
The riddle, as I had put it, is solved. But of course one very deep questions remains: was Black already lost at the 24th move or did he make a mistake later. What do you think? One reaction is:
Charles Sullivan 7/26/2020 02:24
Black did make subtle errors before the 41st move. Stockfish-11 suggests that Black still had good drawing chances after 31...Na4 32.b5 a5 33.b6+ Nxb6 34.Bd2 a4 35.Ba5 Nfd7 36.Kc3 Kb8 37.Kb4 Na8 38.Kxa4 Nc5+ 39.Kb5 Nb3 40.Bc3 Nc7+ 41.Kb4 Nc5 42.Bd2 N7a6+ 43.Kc3 Nc7 44.h4 h6 45.Kb4 N7a6+ 46.Kc4 Kc8 47.Ba5 Nd7 48.Bf1 b6 49.Bd2 Kb7 50.Bh3 Nf6 (+0.97).
Earlier still, LeelaChessZero thinks Black should gambit a pawn with 15...b5 16.cxb5 Nb3 17.b6 Qb8 18.Rd1 Nd7 19.Qd3 Nxb6 20.Qxd6 Nc4 21.Bxc4 Bxc4 22.Na4 Bb5 23.Nc3 Bc4 24.Kf2 Be6 25.Qxb8 Raxb8 26.Rd3 h5 (+0.43).