004. The Rorschach Knight

Black: M. Camejo de Almeida - 14th CC Olympiad (Preliminaries), 2000

This was the first time the Correspondence Chess Olympiad had been conducted by the astonishing new medium known as “email”. I was on board three (for England) and made a lot of draws. My opponent in the game below was playing for Portugal.

It is also my personal record for the longest (by a long way) I've ever followed “theory”. The position after 30...Qg5 was (with a slight detour on moves 26-28) from V.Anand-A.Beliavsky, Madrid 1998. At this point I deviated (from Anand's 31 Neg6) with 31 Bxc5, but this too was following theory – specifically, a line given in MegaBase (by either Anand or Wedberg) through to 38...Ke7, assessed as “with compensation” (for Black, who is a pawn down), which seemed reasonable in view of Black's active bishop and king. Nevertheless, I managed to grind out a win.

My analysis of this endgame was quite comprehensive. Whether that analysis is correct or not is another question, but my notes do indicate several things (which, again, may or may not be correct):
— 40...Kc5 “?!”. Rather than going towards the queenside, albeit temporarily, centralizing the king at once with 40...Ke5 seemed better.
— 46...g5 “!?”. In other words, not necessarily bad. All the same, I might have preferred 46...Ke5 again.
— 50...gxh4 “?”. Here I reckoned that 50...Bf7 would have held. Whereas after 50...gxh4, White has the strong plan Kf4, Ng6, Kg3, Kh4, Kh5 and Kxh6.
— 61 g5+ “+-”. With the added comment: “since Black cannot prevent the pawn reaching g7”. Which is correct, since the Lomonosov tablebases now tell me it's mate in 32.

But perhaps the most intriguing aspect of the game comes in the sideline 70...Ka4 71 Kg6 Bg8 72 Kf6 Bh7 73 Ke7 Bg8 74 Kf8 Bh7 75 Ne8 Kxa3 76 Nf6 (and wins). White's lone knight has clearly done a lot of work – in fact it has made 24 moves so far. For this type of situation, ChessBase offers a Special Annotation: “Piece Path”, which maps all of a designated piece's moves on a small insert. Running that function on this knight produces the following picture:



Well, isn't that nice. The symmetry created by the knight's peregrinations is almost a Rorschach inkblot test.

So, what do you see? :)