Evaluation of gambits

Apr 25, 2015, 2:22 AM |

I evaluated ten gambits with chess engines and I found that six of them were "unsound" to use Roman's terminology. The other four did not show an obvious weakness. I should add that I have no answer to the usual defense of gambits that they will surprise and confuse club level players who have not studied the lines. There are many traps that the gambiteer's opponent can fall into. Even in the Elephant gambit one little mistake by white hands black an early checkmate (Matojelic has a great youtube video on this). And in the Latvian, which is by design a sacrifice-o-rama, there are lines in which Black sacrifices two pawns and a bishop to put White into deep doodoo. My results are summarized in the image below. An interesting statistic is that in the ten gambits tested, gambits by white did better than gambits by black. I am not sure what that means, if it means anything at all. My analysis and comments are available  online. The link to the blue download button is: