1

I don't know anything about the Semi-Slav.

The 5.e3 Semi-Slav with ...Be7 looks like the QGD with ...Be7, except that Black has lost a tempo playing the move ...c6 (Black will play ...c5 in two moves rather than one) and White has blocked his Queen's Bishop inside the Pawn chain (in the QGD White develops it to f4 or g5). One plus and one minus. I thought that the plus and the minus would cancel each other out, but apparently not.

...Bd6 is played 100 times more commonly than ...Be7, therefore ...Be7 is bad, but why?

Stockfish says that ...Be7 is almost as good as ...Bd6, but computer evaluations are not trustworthy in the opening.

Fate
  • 4,786
  • 2
  • 34
  • 68

1 Answers1

2

I'm not sure which database has 5...Bd6 100 times more common than 5...Be7, maybe that is through a different move order? The main move would be to play 5...Nbd7 first. That being said, the main point of putting the Bishop on d6 is to support an eventual Black e5. For example

    [FEN ""] 

1. d4 d5 2. c4 c6 3. Nc3 Nf6 4. Nf3 e6 5. e3 Bd6 6. Bd3 Nbd7 7. O-O O-O 8. e4 dxc4 9. Bxc4 e5!

This move order is probably not recommendable for Black, however, in the 5...Nbd7 lines too there's always a point of Bd6 allowing Black to reply to a White e4 with e5.

but computer evaluations are not trustworthy in the opening.

I would like to challenge this statement, that is simply not true and hasn't been for a while. If you give the computer enough thinking time, its recommendations and evaluations are the most accurate there are. They might not necessarily mirror what is easy to play for a human, however, they are as close to objective truth as we have.

koedem
  • 3,285
  • 9
  • 22
  • 1
    Engine evaluations still aren't reliable in openings - for example just look at how computers misevaluate king indian defence. – Matija Sirk Jun 12 '23 at 07:02
  • @MatijaSirk I disagree on that one. The reason why e.g. the starting position of the KID is evaluated as +0.40 is that it *is* objectively not as solid as other defenses. When playing correspondence where it's very hard to win, I've gotten a couple of wins against the KID. It's just that as mentioned that may not mirror how well humans do with it. – koedem Jun 12 '23 at 08:18
  • But more so, I should have mentioned this earlier, even if you disagree with how Stockfish rates certain structures: if you compare similar moves in the same structure in particular, the evaluation will usually be what one would expect even more so. – koedem Jun 12 '23 at 08:27
  • I haven't lost with KID in correspondence yet, however I believe it's indeed possible for black to lose in it if they trust engine too much. For another example of engines messing up in much more typical structure see my game in nimzo: https://www.iccf.com/game?id=1247790 My opponent followed engine straight into lost position. – Matija Sirk Jun 12 '23 at 08:38
  • @MatijaSirk Fair enough, at that point it's difficult to judge. I can only speak for my experience and haven't played it with Black. (and some GMs calling it less certainly objectively good than e.g. a semi slav may be) But then, presumably it's objectively a draw so arguably all of those defenses are 0.00 objectively. :D In any case, for a decision like Bd6 vs Be7 I would usually trust the engine since I see no reason why it would mess up such a position. – koedem Jun 12 '23 at 19:29
  • for this position i agree of course, I am just saying that in general engines are not to be trusted too much. – Matija Sirk Jun 12 '23 at 20:06