Lessons From Mariánské Lázně: Rules Of Thumb Are Not Sacrosanct

ENGINES play chess by calculating multiple lines, even those lines humans can easily see are hopelessly bad.

Some form of pruning is usually used, but basically an engine calculates, calculates, calculates

Position in my round-eight game after I played 6.cxd4

In the above position, for example, an engine will consider 6...Bxd4?? and 6...Nxd4?? before quickly dismissing them.

Humans also calculate, but rely on rules-of-thumb to tell them which moves are likely to be most promising.

For example, when White establishes a classical pawn-centre, ie one with white pawns on e4 and d4, most humans will give strong attention to the reply ...d5.

This occurs in its simplest form in the French (1.e4 e6 2.d4) and Caro-Kann (1.e4 c6 2.d4), where Black usually continues 2...d5.

Another common example is in a line that could be classified as a Hyper-Accelerated Dragon or a Sicilian Alapin, ie 1.e4 c5 2.Nf3 g6 3.c3 Bg7 4.d4 cxd4 5.cxd4, and now the normal continuation is 5...d5, after which White's most-popular replies are 6.e5 and 6.exd5.

Note that the latter is usually met by 6...Nf6 with ...Nxd5 to come.

In the above diagram my opponent, presumably using his rule-of-thumb knowledge, played 6...d5?

However he failed to properly take into account that his sixth move leaves the black queen's knight in a pin, and that after 7.cxd5 Black cannot play 7...Nf6??, while 7...Qxd5 8.Nc3 gives White a winning advantage.

As I pointed out in my notes to the game, 6...d5? has been played by a 2696, which only goes to show how easy it is to place too much reliance on rules-of-thumb.