The truth of this short note is borne out when you go through Keres' analysis of the Queen ending, and then use even the most powerful engine to try to keep an advantage.

Garry speculates that Rubinstein "moved his bishop from h7, so as not to have to calculate the 'sharp' variations with g7-g6."

Could be, but there were a number of other tactics he would have to see. For example,

16 Ne4 Be7 17 h4 Rac8 18 Ng5 (Garry gives the more positional 18 Nc5, but a direct attack appears stronger) Bb4+ 19 Kf1 c5 20 d5 (this may have been what Rubinstein missed-this pawn sacrifice leading to an exchange sacrifice opens the path for the bishop to f5)

click for larger view

20...Bxd5 21 Rd1 Rcd8 22 Rxd5 exd5 23 Bf5 Qe7 24 Be6! g6 25 Bxf7 Qf6 26 h5 Qxg5 27 hxg6+ Kg7 28 Rh7+ Kf6 29 Bxe8 Rxe8 30 Rf7+ Ke6

click for larger view

I looked at it briefly, but didn't understand if it was purely defensive, or it could be analysed to a win.

Kasparov calls 20 Qb3 a5!? "interesting" in OMGP without fully endorsing it, and I currently have Komodo looking at it...

click for larger view

Here are two attempts by Black after 20 Qb3. I like to analyse with Komodo because it plays for fortress and positional draw more than Stockfish or Houdini.

At 27 ply it still thinks its best chance is Kasparov's 20...a5, retaining the idea from the game of using the pawn majority to create a passed pawn.

At 28 ply (and 11.5 hours) it switches to a tactical idea with 20...Bc8, allowing White to win a second pawn with the surprising 23 Nxb5, but getting rid of all the Qside pawns, and going for a positional draw two pawns down with the two bishops, a safe King, and a more active Rook.

It would be nice to see what Stockfish makes of 20 Qb3, as it would reveal some sharper attacking methods to break these ideas.

20...a6-a5 21 Rf1-d1 Bb7-c8 22 Nc3-e4 a5-a4 23 Qb3-d3 Rc5xc1 24 Rd1xc1 Bc8-f5 25 g2-g4 Bf6xb2 26 Rc1xc6 Ra8-d8 27 Qd3-c2 Bf5-d7 28 Rc6-c5 Bb2-e5 29 Rc5-d5 Kh8-g8 30 Qc2-c5 Rd8-e8 31 Ne4-d6 Be5xd6 32 Rd5xd6 Re8-c8 33 Qc5-d4 Bd7-e6 27/-0.79

20...Bb7-c8 21 Qb3-a3 Bc8-e6 22 b2-b4 Rc5-c4 23 Nc3xb5 a6-a5 24 b4xa5 Qe7xa3 25 Nb5xa3 Rc4xc1 26 Rf1xc1 Ra8xa5 27 Na3-b1 Ra5xa2 28 Bf3xc6 Ra2-a1 29 Bc6-e4 Bf6-b2 30 Rc1-d1 g7-g5 31 g2-g3 Kh8-g7 32 Kg1-g2 g5-g4 33 Rd1-d6 Bb2-e5 34 Rd6-b6 Ra1-a4 35 Be4-c6 Ra4-c4 36 Rb6-b5 Be5xg3 Bc6-d5 Be6xd5+37 Kg2xg3 Bd5-e4 38 Nb1-d2 28/-0.69

The fact that Tarrasch, not a wild player, saw 20 Ne4 Be7 21 h4! shows that the seeds of further combinational ideas were in the air even then.

Kasparov says White also could have got an advantage from 16 Bd3 e5 17 Ne4 exd4 18 Nxf6 gxf6 19 Qe2 Qd5 20 0-0 c5 21 e4.

The complications are daunting, and many more players, including Kasparov, who gives the note, would lower their sights and opt for a bind after 16 Ne4 Be7 17 h4 Rac8 with 18 Nc5 playing against the bishop. That looks good enough too, and does not need the god-like vision that Komodo's line suggests.

But I think we could add Tal to the list of players though who would take up the banner of Kingside attack with 19 Ng5. In a way it is more natural to go to g5, since h4 has been played.

click for larger view

"White has a fine position in this Queen ending and at first glance one can hardly believe that Black will be able to avert loss. White has in the first place a sound pawn more, but this fact is not decisive in itself. As is well known, in Queen endings the most vital thing is not to have a pawn more but to have a passed pawn, this being of much greater importance. A strong passed pawn in a Queen ending is usually enough to compensate for a material disadvantage. However, in the position after 27. Qd6 Black, on the contrary, has no such advantage.

In the second place it should be observed that White's Queen is much more actively posted than Black's, and it not only threatens the pawn on c5 but it also controls the more important central squares.

Finally, in the third place, White's King is much more securely situated than Black's, this being no small advantage. So it can be seen that Capablanca's plight is not at all enviable.

...In general, Black must not behave too passively and thus give White the occasion to advance his King-side pawns without having to worry about counterplay on the Queen's side. In that event White's win would be merely a question of time. Black must therefore find an active plan and hence Capablanca played with absolute correctness 27...b4!

I believe Alekhine wrote of this after the match.

<....In general, Gazza postulates that Rubinstein contributed more to the progress of the game than Capablanca.>

Rubinstein is one of those players who everyone gives lip service to as a great player who didn't become world champion, but his actual games just don't seem to have that much visibility. That's not to say that they shouldn't, they just don't.

From aged thirteen or so, I owned a copy of the Dover edition of Rubinstein's games and cannot imagine how this could have happened.

While I do not consider him amongst the very greatest players who never won the title, anyone could learn from Rubinstein's games.

<...."Contributing to the progress of the game" is a pretty nebulous term....>

Especially in those early days, game collections on the greats had a tendency to very often lapse into paeans to that player and compendia of how their opponents could not put a foot right.

There's a section in one Irving Chernev book about "The Incredible Genius of Petrosian", or some such, in which Irving credits Tigran with anything and everything, sort of like a non-nasty Lt.Surena post, if such a thing ever existed.

Read most of Chernev's books, but do not recall that one.

<....sort of like a non-nasty Lt.Surena post, if such a thing ever existed....>

I would say the odds are agin it; at least I remember no such animal.

<....In one game Petrosian played the Winawer, and of 3... Bb4, Chernev wrote "Knights before Bishops, say the Experts!" It's a <book> move, fercryinoutloud!>