Bishop exchange in Italian (Giuco Piano)

While analyzing my yesterday's game



I was surprised that the engine evaluated 8. ... Be6 as a mistake and suggested 8. ... Bxe3 as best move instead.

In book Understanding the Chess Openings by Sam Collins, the author has this comment on a position almost the same as here after 8. Be3 (except for the h6 and h3 moves):

> I've selected a very boring position from the Giuoco Piano to talk about captures for a moment. In the diagram position, captures are bad, and here's why:
> 1) If the black bishop takes on e3, then after fxe3 White has stregthened his centre as well as now having half-open f-file to work on. Also he's covered the d4 square.
> 2) If White plays Bxc5, then after ...dxc5 black has increased his control over d4 with his pawn and the d-file with his major pieces.

Both points seem to make sense (even if it's obvious that there are also some drawbacks). And yet, Stockfish evaluates the position after 8. ... Be6 as significantly worse than both before the move and after 8. ... Bxe3 (1.1 vs -0.1). Moreover, it also suggests 9. Bxc5 as the best reply to 8. Be6, directly contradicting both statements quoted above.

What should I take from this? Is it an example of a "computer chess vs. human chess" difference or is the reasoning in the book wrong (or at least outdated)?

Wow that was Really Cool , a Game Style like 1900+

@mkubecek said in #1:
> In book Understanding the Chess Openings by Sam Collins, the author has this comment on a position almost the same as here after 8. Be3 (except for the h6 and h3 moves):
> [...]

Perhaps in the specific case of the Black's move ...Bxe3 the difference might be significant, because White's move h3 has weakened the g3 square and the dark squares near White's king are then severely weakened after the recapture fxe3.

But I don't know. It's interesting.

@aVague Thank you. :-) It was far from perfect, apparently, but I was afraid that with my poor time management I would flag sooner or later so I went into the "give your opponent some opportunities to blunder" mode and thankfully it worked.

@mkubecek said in #4:
> @aVague Thank you. :-) It was far from perfect, apparently, but I was afraid that with my poor time management I would flag sooner or later so I went into the "give your opponent some opportunities to blunder" mode and thankfully it worked.
I ve used that before very often , but uncounsinacly, it's working good, but in the End this reflex game should be out worked to clean fast vision of Desk and to Use Your Real Chances!

@Brian-E said in #3:
> Perhaps in the specific case of the Black's move ...Bxe3 the difference might be significant, because White's move h3 has weakened the g3 square and the dark squares near White's king are then severely weakened after the recapture fxe3.
Interesting idea, thank you. I tried to run Stockfish on the position with pawns back on h2 and h7 (i.e. exactly the same as in the book) to check this hypothesis and the results are the same so this is likely not the reason.

The difference lies more in the fact that in this specific position, after Be6, white has the concrete reply of Bxc5 and Bb5 after which you will end up with tripled pawns on the C file as you must defend your e5 pawn.

Long ago, when the earth was green, I learned the same general principle. That the doubled e pawns and semi open f file are strong for white. Perhaps the additional weakening of g3 after h3 is a mitigating factor, perhaps not. I was also taught that the dark square control black gets after ...dxc5 is, while not brilliant, at least workable for black. Certainly though that Qside structure works better for white after ...B/e6xBc4. I think the main point of reference is a high level game Larsen played back in the 60s, BICBW.

A matter of machine chess being different from human intuition? Perhaps. I like to think that human intuition is shorthand for hours and years of study and experience. For what generally seems to work. John Watson had a lot to say about rule independence in his Secrets of Modern Chess Strategy. For modern players, and especially for the machines, general rules may (or not) be generally true, but what counts is what works or doesn't in specific positions. Food for thought here.

Thank you both.

Also worth noting here is that concretely, after Bxe3 fxe3 black has Na5, forcing white to lose their light-squared bishop which limits their attacking prospects.