Friday, October 4, 2024

Jerome Gambit: Wisdom of the Stockfish - Huh??


The previous blog post looked at the different move choices that show up early in the Jerome Gambit (1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+), and their popularity.

It is an interesting, but shorter, (and somewhat depressing) excursion, to see how a computer engine - in this case, Stockfish 16.1 - evaluates things.

The question is: Is it helpful?

(I use a computer to evaluate my play in a game, post mortem, looking for my errors. I am far more likely to choose opening lines based on popularity - see the previous post.)

For the following, Stockfish's assessments are based on looks that are 30 ply deep.  

Be warned: the computer sees White's chances in the Jerome Gambit as being rated from bad to worse.

Starting with 1.e4 e5 2.Nf3 Nc6 3.Bc4 Bc5 4.Bxf7+ the computer has no interest in any response by Black other than 4...Kxf7, which it evaluates as -2.67 for White. It evaluates 4...Kf8 as +2.11 for White, and 4...Ke7 as +2.54 for White. (Not surprisingly, the strongest response, as we have seen, is the most popular among players.)

How should White respond after 4...Kxf7?

The computer prefers 5.Nxe5+, which it assesses as -2.87 for White, as this is better than 5.O-O (-3.18), 5.d4 (-3.39) 5.Nc3 (-3.40) an 5.c3 (-3.54) - but not by a lot. (Again, Stockfish's top choice is the move that is also the most popular. It is interesting, though, that the second highest-rated move is neither the second most popular nor the second most successful; perhaps 5.O-O needs to be explored further.)

As might be expected, 5...Nxe5 (-3.02 for White) is Black's strongest response to White's capture by the Knight - and there is only one rival move, 5...Kf8 (-1.56). Other moves are evaluated as good for White: 5...Ke8 (+.55), 5...Ke6 (+.98), 5...Kf6 (+1.59) and 5...Ke7 (+2.83).

White does best to follow the Knight capture with 6.Qh5+ (-3.01 for White), as the computer considers it the least worse move, followed by 6.d4 (-3.35), 6.O-O (-3.91), and 6.c3 (-6.75).

In turn, Black's best response to the Queen check, 6...Ke6, is evaluated as three pawns better for him (-3.00), with 6...Kf8, at -2.71 not too far behind, followed by 6...g6 (-2.24)...

And so it goes.

I hope that is enough computer "wisdom" for Readers.

Until artificial intelligence can inject surprise and being intimidated and casual dismissal and just-for-fun (and a host of other reactions), computers like Stockfish will only evaluate part of the Jerome Gambit.

No comments: