Greetings,

Smyslov_Fan wrote on 01/05/09 at 04:32:42:

DraganGlas' observation that the players play the check is just the opposite of what I thought was going on.

It seems to me that the players with the queens are often almost afraid to play the check and look for the overly subtle moves to win.

In the famous Gelfand-Svidler game, snake-bit Svidler lost another half point in the endgame, but this was as much due to time pressure as anything. (I remember watching the game live on ICC, and the spectators with the database kept screaming the moves without paying attention to Svidler's time.)

I wonder how many of the other endgames were played in time pressure. I also wonder how the great players of the past would have done against such defensive acumen.

I've gone through them again - to check if my first impression was correct or not.

Not having studied this type of ending before, I now see that the "diagonal waiting move", to which I referred, is actually the queen attacking the file on which the rook stands from the fifth rank.

It's almost similar to the square of the pawn, only in this case it involves the queen on the fifth rank, where her king stands within the square on the fourth rank. It's clearly a pattern with the intention to set up the fork of king and rook. I've marked the move which sets up this pattern in bold.

Also, by including the number of moves to win (and lose), I've shown how the queen's side keeps shooting themselves in the foot with the one-pace-forward/two-paces-back errors.

Girkassa's ending: For the most part, you played the best move or one which shortened the loss by only one move - except for:

80..., Rd3 (loses in

**25** instead of the "best", 80..., Kc4, which loses in

**29**)

91..., Ke2 (loses in 20 instead of the "best", 91..., Kd2, which loses in 22)

91..., Rd3 (loses in 23 instead of the "best", 91..., Rh4+, which loses in 25)

Where your opponent made mistakes:

83 Qe2+ (wins in 25 instead of the "best", 83 Qe3, which wins in 23)

84 Qe4+ (wins in 26 instead of the "best", 84 Qe5/f1, which wins in 24)

91 Qb1+ (wins in

**23** instead of the "best", 91

**Qc5**, which wins in

**19**)

93 Kf4 (wins in 21 instead of the "best", 93

**Qc5**, which wins in 19)

95 Qd4+ (wins in 21 instead of the "best", 95

**Qd5**, which wins in 19)

96 Qg1 (wins in

**26** instead of the "best", 96 Qe4+, which wins in

**20**)

Most of the errors were queen checks, instead of waiting queen moves - in the final case, the chosen move was the fifth choice(!) instead of the check - or

**Qd5** again.

Morozevich - Jakovenko Jakovenko, likewise, made fewer errors in comparison with Morozevich:

72..., Rd5 (loses in

**23** instead of 72..., Kb4/Rc4/Rd3, which loses in

**27**)

73..., Rd4 (loses in 24 instead of 73..., Rc5/Rd3, which loses in 25)

77..., Kb5 (loses in 19 instead of 75..., Kb6, which loses in 20)

88..., Re4 (loses in 24 instead of 88..., Rd3/f4, which loses in 25)

92..., Kd3 (loses in

**20** instead of 92..., Ke3/Re4, which loses in

**24**)

96..., Rg4 (loses in

**20** instead of 96..., Rh4, which loses in

**24**)

Morozevich's errors:

73 Ke6 (wins in

**26** instead of 73 Qc3, which wins in

**23**)

81 Qc3+ (wins in 18 instead of 81 Ke7/Qf6, which wins in 16)

82 Kd6 (wins in

**26** instead of 82 Qe5+, which wins in

**17**)

84 Qc6+ (wins in 26 instead of 84 Qe3/e5+, which wins in 24)

85 Qd6+ (wins in

**28** instead of 85 Qb6+, which wins in

**25**)

87 Qa3+ (wins in 27 instead of 87 Qf6/g3+, which wins in 26)

92 Qf6 (wins in

**25** instead of 92 Qf2+, which wins in

**21**)

93 Qd6+ (wins in

**28** instead of 93 Qf1+, which wins in

**20**)

104 Qd4+ (wins in 14 instead of 104

**Qe5**, which wins in 13)

105 Qd3 (wins in 15 instead of 105

**Qe5**, which wins in 13)

106 Qe3+ (wins in 15 instead of 106

**Qd5**/b1+/d4+, which wins in 14)

108 Qd2+ (wins in 15 instead of 108 Qd1, which wins in 13)

110 Qg3+ (wins in 14 instead of 110

**Qd5**, which wins in 13)

111 Kf3 (

**DRAW** instead of 111

**Qe5**, which wins in 13)

Topalov didn't do too badly - he made fewer mistakes than Svidler, which is telling. The latter's endgame with Gelfand featured more errors by Svidler than Gelfand, as with the first two endgames above.

I did do the same for his first game against Topalov - but there were so many errors, it's pretty clear he's not as accomplished at these endgames.

He really needs to learn how to play this endgame better!

I'm surprised that he hasn't done what he did - and still does, apparently - with his opening preparation: use Bookup or some similar CBT aid to practice endgames. Or if he has, it doesn't seem to be working as well!

Kindest regards,

Dragan Glas