Karpova: A conjecture:
Time control: 2 h for the first 30 moves, 1 h for additional 15 moves
<Die Bedenkzeit beträgt 2 Stunden für die ersten 30 Züge, sodann 1 Stunde für je weitere 15 Züge.> (Page 109 of the 1911 'Wiener Schachzeitung') This game was played in Round 22 (of 24) on September 19, i. e. alomst 1 month after the start of the tournament (August 20). Now after 35.Rxa5
click for larger viewRubinstein blunders with 35...Ke6 (this is the actual mistake, 35...Rb2 was better) as now 36.Nd4+ forks King and Bishop. Now there's hardly an alternative to 36...Kxe5 as 36...Kf7 still loses the Bishop and leaves White with the strong passed e-pawn and drives further away the King (while the white King can come easily to avoid the d-pawn push that happened in the game, e. g. 36...Kf7 37.Nxc2 Rb2 38.Rc5 and Black cannot win time by driving the white rook to c3). The closer they got to move 30 (first time control), the more White's advantage increased, so I conjecture that at least Rubinstein had been in very bad time trouble already back then. So after move 30, now that many liquidations into a real endgame were about to occur, they might have used up again a lot of their time to assess the new situation so that they got into time trouble pretty soon again. With Kostic missing 36.Nc6+, when he was also in time trouble already he may have had to move almost instantly as he still had many moves to make to reach the next time control. Also, he may have been concentrated on the obvious threat and purpose of 35.Nd4+, i. e. taking the bishop that was en prise - neither having the time to look for alternatives and doing something natural, underestimating the strength of Black's counterplay (after all, White is ahead a piece with 3 pawns for each side which should usually be enough). Another phenomenon might be that in time trouble, you have to save as much time as possible and moving the knight closer to your own position so as to be able to faster make the remaining moves (to get the point across, imagine the basic K+R vs K mating pattern and you are about to checkmate the K on the 1st rank, so to drive the K in opposition to your own K, you move the R to the 3rd file and not to the 8th so as to save time. Another situation in time trouble when you move your piece closer to the other pieces that you'll have to move so that your hand is already in the right position - important when the opponent responds instantly also). Though there's no fool-proof evidence for this conjecture, at least it makes sense and explains the blunders. After all, these are typical time trouble blunders which you could not expect from these players otherwise. A better explanation than <very very funny> and <this is 1911> though. Especially considering that even more grotesque blunders can also be seen nowadays among the elite, especially when in time trouble. | 
Kibitzing
