Ask Houdini

A while back I bought Houdini 3, and have been enjoying its assessments, having placed it next to Rybka 3 as an analysis partner.

For fun, I set up the original Jerome Gambit sacrifice (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7) and asked Houdini the best follow-up for White. Not surprising, after 3 hours of "infinite analysis" it chose the classical 5.Nxe5+ as its top line, but only about a quarter of a pawn behind that move were 5.c3 and 5.d3.

Checking The Database, I noticed there were 1040 games with 5.c3, and White scored 44% with it. By comparison, there were 842 games with 5.d3, and White scored 31%. (There were 4165 games with 5.Nxe5+, with White scoring 54%)

I again challenged my silicon assistant, how to follow up after 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7 5.Nxe5+ Nxe5, and after 3 hours of "infinite analysis" it chose 6.Qh5+ (heading for the "boring defense"), but only by a bit less than a half-pawn over 6.d4.

There were 2,886 games with 6.Qh5+, with White scoring 55%, versus 1,075 games with 6.d4, with White scoring 50%.

Allowing for all the cautions associated with statistics, and there are many, I think it's time for me to take a closer look at 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7 5.c3, and maybe give a second glance at 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ Kxf7 5.Nxe5+ Nxe5 6.d4.