What does this ...h6 move do?

In the post mortem analysis SF 12 NNUE came with this move ...h6. After 13 hours of thinking it chose for this move. But I so no reason to do so. What does this move do?

I mean, you know that h6 is often played to stop the queen bishop battery from being used to trade the bishop.  That probably isn't much of a benefit in this position, and gives white the move for h3 or something else, so I guess the bigger question, at least to me, is how are the other candidate moves flawed?  What were you looking at in this position?  The intent of ng4 was what, ne5, bxn or f5?

To prevent the exchange of the g7-Bishop?

Obviously that was my first thought, as it makes space for the king to assist. But I don't see any benefit in that at all. Black has enough resources to prevent the exchange anyway.

Not sure what this means. White would be the one who WANTS to exchange the g7-Bishop. Black would be the one trying to prevent it, with Kh7 to stop White's Qd2 followed by Bh6.

Looking at the stockfish lines (not after 13 hours lol) it appears that Stockfish does indeed want to play qd2 and exchange if allowed.  It doesn't surprise me that h6 offers some small benefit.  I mean, isn't black's bishop the best minor piece on the board right now?  If the other moves don't actually do anything useful, why not prevent the swap?  Rb8 seems rather empty, and I can't see how ne5 accomplishes much.  What else were you thinking or hoping to accomplish?

I'll take this for an answer. Meaning, there isn't much better. But I wasn't sure. But it's hard to accept  a system of this strength that it makes a move because "there isn't much better". So figured I ask here, I might have overlooked something.

@ blueemu , yes black, I meant black, not white of course.

In this position there are a dozen moves that are about equal. It really does not matter if some position is evaluated 0.05 points higher or lower than the others: no engine is that precise.

Moreover, it is really absurd to let an engine think for 13 hours in this flat, boring, uninteresting position. What is the point?

Its been answered.  Trying to exchange off the fianchetto'd bishop is a common manuever as it often renders the king's defense weak.  Even if that isnt currently a factor here (note that the epawn hasn't moved helping the defense of f6) black's bishop is currently probably the best minor pieces on the board controlling a meaty diagonal.  H6 prevents the exchange of bishops, and is a clearly useful, if not terribly exciting move.  Its hard to find another move for black that is clearly useful.  This isn't one of those chess engine moves that is too deap for humans to parse -- h6 makes sense using basic positional analysis.  

@Barefoot_Player

Yes, you're right, one shouldn't overuse the engine. Well said.

I think the idea is Kh7. Pretty common idea in a lot of lines where the bishop is fianchettoed. 

Black wants the R on f8 and the B still on the board but black can't play e5 and f5 without preparation. The logical place for the K is h7 because of the difficulty in getting the white sq B to train on that square.

Therefore h6 was a good move if black is trying for an advantage and it shouldn't take 23 hours to work that out. It took me two minutes.

I don't think black wants to play e5 here.. The dark squared bishop is probably the best minor piece and e5 makes it much worse. I'm not so sure if black wants to play f5 either as this weakens g6 significantly. After the king is tucked away I like ideas like Ba6, Qb6, Rb8, and Rd8. Playing h6, Kh7 gives added support to the h6 square itself which I think is a critical point.

Well, e5 and f5 demand about five or seven moves of preparation, after which the game plan might be different. White is going to play h3. I think white's best chances lie on a q-side pawn advance, fully supported by pieces. White would like to get the dark squared bishops off, to that end. Black can only play e5 after adequate preparation or the d pawn is going to be very weak.

h6 is a critical point?

And even without that... this is not the anslysis of a 2400 player. Reported, lol

Qd2, Bh6 is an idea, and h6, Kh7 prevents this. I have used this idea in the fianchetto variation of the King's Indian. I'm happy to talk more about the position if you would like to discuss.

When I said this is a critical point I didn't mean the square.

Re8 Qd2 Qa5 h3 Ne5 Bh6 Bh8 also avoids the the bishop exchange. h6 is a pawn move, and Re8 is not. If it were about preserving the bishop then Re8 would be preferred.

The answer is a little more sophisticated than looking at the single square the immediate move influences. White obviously mixed his systems playing Nxc6 when it was incorrect and now he's struggling to find play. Black's minors are superior, and e.g. a sequence involving Ne5 Bxe5 would be bad for white.

That's closer to a real analysis / what's actually happening in this position than "hurr-durr h6 Kh7 prevents the trade of bishops"

And that criticism isn't aimed at any single person, it's basically aimed at all of you.

h6 feels like a pretty useless move right now. I prefer Rb8 to bring the rook to the open file.