The Elo rating system is a brilliant statistical application that enables predictions to be made not just of who is more likely to win a game between any two rated players, but what the score is likely to be after a long series of games. Other sports have their ranking systems, but the late Prof Elo gave chess a percentage expectancy curve on which you could look up the difference between two players ratings and estimate the score between them.

A 200-point difference, for example, predicts that the higher rated player will score roughly 75 per cent - a figure that really puts into perspective Garry Kasparov's extraordinary result against the Israeli national team last week. The average rating of his four opponents was just over 2600, and Kasparov's rating is just over 2800, so he would expect to score a little over 75 per cent against them in normal play. Yet playing them all simultaneously, he ended with six wins and two draws: winning 3-1 on the first day and 4-0 on the second.

It looks as though the sheer bravado of playing four games at once more than compensates, in its intimidating effect, for the extra pressures of time and concentration it creates. His opponents certainly gave him a good deal to think about, as this hugely complicated game from the match confirms.

White: G. Kasparov

Black: A. Huzman

1 e4 c5 22 Be3 Qa6

2 Nf3 d6 23 Qb4 Rxc6

3 d4 cxd4 24 dxc6 Qxc6

4 Nxd4 Nf6 25 Qb3 0-0

5 Nc3 a6 26 g5 Qxf3

6 Be3 e5 27 gxh6 Rb8

7 Nb3 Be6 28 Rhf1 Qxd1+

8 f3 Nbd7 29 Rxd1 Rxb3

9 g4 h6 30 cxb3 gxh6

10 Qd2 Be7 31 a4 f5

11 0-0-0 b5 32 Bxh6 f4

12 h4 Nb6 33 a5 Kh7

13 Qf2 Nfd7 34 Bg5 Bxg5

14 Kb1 Rc8 35 hxg5 f3

15 Nd5 Bxd5 36 a6 e4

16 exd5 Nc4 37 a7 Nb6

17 Bc1 Qb6 38 Kxb2 e3

18 Qe1 a5 39 Kc2 e2

19 Bxc4 bxc4 40 Rh1+ Kg6

20 Nxa5 c3 41 Kd2 resigns

21 Nc6 cxb2