My latest game... please explain why it's telling me this is a blunder?

First, I know I'm not good and made a myriad of mistakes, so take that as a given.

I just want to know why in my latest game vs. DanielServenti, my (black) move 12, that it's telling me that was a blunder. I know I could have gotten a "free" bishop if I did its suggested move of taking on a3. But white could have responded with Knight c7 taking a pawn, forking my king and rook. My bishop move to e5 saved my bishop while preventing this fork. Should I really have taken and allowed the fork? Why?

I would have been +3 from the bishop, he'd have been +1 and then +5 from the rook. So even if I somehow trapped the knight and took it without losing another piece, it would have been roughly even but with a worse pawn structure.

But I get: Bishop (+3), MAYBE a free knight if I pull it off (+3 later on)
He definitely gets, right now: Pawn (+1), Rook (+5), fracturing my 3-pawn structure and preventing any future castle. 
Looks to me like the 3pt bishop sac is totally worth it because he's done 6 pts of damage plus the disruption I mentioned, and who cares if the knight is trapped because it will cost me at least 3 moves (pawn, bishop, bishop) to capture it. 

You also need to consider positional elements. Look at the position that results from stopping the Knight fork, particularly after 15. Bxd5 ... where Black is completely lost. All of Black’s pieces are sitting undeveloped on their original squares, Black's uncastled King is dangerously exposed in the center while White's King is safely castled, the isolated d4 and d7 pawns are weak, the f7 pawn is weak and about to fall, White has two active Bishops, White’s Rook controls the critical semi-open f-file, etc. White’s two Bishops are deadly, particularly the dark-squared one with no equal, which is one of the reasons why capturing it earlier would have been prudent.

I definitely understand why this is difficult to see. But as others have said, it's not about material. it's about position.  If white had taken with the Bishop with check  on move 16 (17. Bxf7+) the position is absolutely lost. The engine considers it a blunder because of what will follow. Even when your next moves are perfect, you are in a losing position.

But even when looking at the position where he (sub optimally) plays 16. Rxf7. After you played Nxd5 he blundered by retaking with the pawn, losing his Rook in the process. 

If he had played correctly 17. Rbf1 your position looks terrible. And his Bishop on a3 is definitely part of his overwhelming attack. White is at +5 or 6 depending on depth.

If you could go back in time to move 12, you would take the roughly neutral exchange of Rook versus two minor pieces.

I guess. In all the games I watched, allowing the rook fork was always an enormous mistake. Here, in that situation, apparently it's not. I've just never seen that happen. 

What if there were a similar situation, where there's a "free bishop" but it's actually worse to take it. Would the engine still claim it's a blunder to not take it?

The point I was trying to make:  It's not a "free Bishop". It's simply the best move in this position.

Alternative options would be to move your Rook (Rb8), move your King (Kd8), Ne7 (an inaccuracy) so you can capture the Knight after he takes your Bishop.

However, you played a move that put you in a completely losing position.

If you can take a Bishop without compromising your position, by all means please do
(edited)

In the middle game, two pieces is often better than a rook.  Queens are off the board, so king safety (or the inability to castle) is less of a worry.  Finally, after the fork, your move of Kd8 already traps the knight if they take the rook and they're going to have a hard time getting it out without a dark-square bishop able to cover c7.  Sure, it may take a few moves for you to take the knight, but why bother wasting time taking it if it is stuck in the corner doing nothing?  It might as well already be off the board.  Develop your knight, bishop, and your remaining rook and start making threats on dark squares (where their remaining bishop will be of no use) and they won't have time to free their knight.  They will effectively be playing down two pieces for the rook.  Pick up a few central pawns or one of his pieces and then you can make time to officially win the knight.

I think you're overlooking the fact that after your 12...Be5 13.f4 your king is trapped in the center and White has an overwhelming position. In fact , my engine evaluates the position as +7.00. If White had played 16.Bxf7+ or 17.Rbf1 your position was hopeless.  So 12...Bxa3 was your only decent move. Whether you got one piece or two for the rook was of little importance

Top 5 lines + Max Analysis:


After 12. ...Be5 13. f4 blacks' position is in shambles:

Thanks computer, that pile of letters and numbers really clarifies everything. 

Well, those numbers simply verify what I said in #10. 12...Bxa3 was the only move that gave you a ghost of a chance. After 12...Be5 you were completely crushed against any decent play by white. 

Also, that "pile of letters and numbers" were chess moves and their evaluations. It's pretty easy to read, if you would care to try

Keep in mind that "blunder" just means a move which gives away a lot of the advantage, it does not necessarily mean a "BAD" move, as in a move which makes you lose.  Ba3 is a fine move however it is not nearly as good as the top move in that position, which is why it's called a "blunder"

This seems a little harsh. To be fair to the OP, posting a bunch of engine evaluations is a terrible answer to their question. It demonstrates no understanding of the position whatsoever. The engine’s assessment is what generated the question in the first place. 

"This seems a little harsh."

Which part? I don't think that pointing out that his position was totally lost after 12...Be5 is harsh. That's just chess. I think it's simply instructive to point out why his choice of moves was a mistake.  While it's true that dumping a bunch of engine lines may not be the most user-friendly approach, those engine lines did contain a lot of information that wasn't too hard to read

The OP is asking why the engine thinks 12... Be5 played by Black is a blunder, not 12. Ba3 .... played by White. Coincidentally, 12. Ba3 ... is also a blunder.

"that "pile of letters and numbers" were chess moves and their evaluations. It's pretty easy to read, if you would care to try" is all I was referring to. It comes across as suggesting the OP has not put enough effort into trying to understand the engine's suggested lines. Perhaps I misread your tone. When I was around the 1050 range, I would have found this position and the engine's recommendations quite confusing. People, including yourself, have done a good job of explaining the position in human terms. 

ah ok im sorry about that. This makes the answer much simpler, it's a bad move due to f4 followup