George Duke wrote on Fri, May 10, 2013 04:20 PM UTC:
Reti versus Capablanca 1924 31 Moves.
Http://www.chessgames.com/perl/chessgame?gid=1102101.
Everything before is a black box. '20 Bxd4...' is the technical
brilliancy on account of leaving c-Pawn en prise -- completing White's
previous 19 Q-d2 set-up. Hence analytic post-mortem parfait commences on
Capa's Move 20 response. Then the Locus of Suspicion Theorem points to in
all likelihood correction being at that Move 20, 21, or 22 in descending
order.
Where does Black go wrong? Without even looking ahead, we know it is Move
20 or 21. And correction has to carry strong enough justification for
Black winning.
Okay seeing in reality, Capa takes the bait 20 ...Qxc4? Frankly taking
that free Pawn puts Queen in awkward spot without support. Sucker's bait.
Obviously the grab '...Qxc4' essays no plan, and found already by plain
abstraction right where expected where Black strays. So what should at the
juncture J. R. Capablanca do instead?