White to play
Speelman-Sokolov, Brussels 1988
After last Monday's interlude, it's back to rook endings this week. We have two items on our agenda today. First, find out how Speelman-Sokolov finished; second, have a look at how tricky rook endings can be.
Speelman - Sokolov isn't in this
If you were with us a fortnight back, you'll perhaps remember we left Spess having a "long think" in the position at the head of today's blog. In his article in CHESS (July, 1988) he writes,
On move 44 I wanted to play my rook behind his c-pawn, but somewhat misanalysed and so did something else constructive - improve my king's position.
Speelman felt he could have improved his play later on (48 Ra3! and 62 Kf8!), but he went on to win the game anyway. Sokolov sealed 67 ... Rg2 but resigned without resumption.
Time to move on to the trickiness of rook endings. 44 Rh8 was the move Speelman wanted to play. Let's have a look at why he might have rejected it.
Well, Black's got to get going or he'll simply be lost so what about,
44 ... c4, 45 Rc8 Rc5, 46 Rxc5 Kxc5
Now you're first thought might be - at least my first thought was - to try pushing the h-pawn. However,
47 h5 c3, 48 h6 c2, 49 h7 c1=Q+, 50 Kg2 Qa1
just loses. So instead,
47 Kf1! when 47 ... c3, 48 Ke2 c2, 49 Kd2
wins for White, but better would be
47 Kf1 Kb4!, 48 Ke2 Kc3 (to stop the White king getting to d2) 49 Kd1 Kb2, 50 h5 c3, 51 h6 c2+
when Black wins.
However, what about 47 Kf1 Kb4
48 Ke2 Kc3, 49 h5! Now, 49 ... Kb2, 50 h6 c3, 51 h7 c2, 52 h8=Q is check so it's +- again.
OK, but Black can improve too.
48 Ke2 Kb3!, 49 h5 c3, 50 h6 c2, 51 h7 c1=Q, 52 h8=Q
52 ... Qc2+, 53 Kf1 Qd1+, 54 Kg2 Qf3+, 55 Kg1 Qd1+, 56 Kg2 and there are no more checksbut
52 ... Qc4+
is just an instant draw.
To summarise, from our starting position after 44 Rh8 c4, 45 Rc8 the main line is,
45 ... Rc5, 46 Rxc5 Kxc5, 47 Kf1 Kb4, 48 Ke2 Kb3, 49 h5 c3, 50 h6 c2, 51 h7 c1=Q, 52 h8=Q Qc4+
when White has two extra pawns, but no way to win - as shown by Speelman in that old CHESS magazine.
How difficult are rook endings?
Think about this: a full consideration of White's options at move 44 is impossible unless you can calculate (or at least intuit) all of that lot in advance
Now think about this: after 44 Rh8 c4, White's 45 Rc8 isn't even the right move.
Rook and pawn Index