Imperfections in Chess.coms Opening Database

Sort:
Avatar of folderal

This position arises out of the Neo-Catalan:

 
If you look at the database, it will show 5.d4 as the most frequently played move.  Which is  wrong. The position where the pawn is on d4 on move 5 is the most frequently occurring, but not in this move sequence:  1. Nf3 Nf6 2. c4 e6 3. g3 d5 4. Bg2 dxc4

The difference is that in the normal Catalan, the move d4 would be played before c4, but in the Neo-Catalan, c4 is played first.  This means that the capture 4...dxc4  now allows 5.d4 to be answered by 5...cxd3 e.p., leaving White a pawn down and with a ruined pawn structure.

So the transposition engine is saying "Ah, I've seen this position before!", but what is neglected is the move sequence that gave rise to that position.  5.d4  is a losing move, though the database will tell you it's the most frequently played.
Avatar of notmtwain

I think you may be on to something.