Why is move 12 a blunder for both sides?

Question

I have trouble understanding why move 12 for both sides was such a blunder? Evaluation went roughly from -1 to -6 and back to -1.3. The engine suggests 12... Ne4. I understand that e4 is a good outpost for the black knight. But so is e5 for the white knight and they can go there with a tempo (hitting the queen). That is why I (as black) decided to retreat the knight to d7 to prevent the white knight coming to e5. Analyzing with a computer I see that somehow black is able to make progress after 12... Ne4. But I get lost in the amount of variations and I still don't get it. Is there an easy way to see it?

[FEN ""]
                                                                                                     1. d4 d5 2. Bf4 c5 3. e3 e6 4. Nf3 Nf6 5. Nbd2 c4 6. c3 Nc6 7. b3 b5 8. b4 a5 9. a3 Be7 10. Ne5 Qb6 11. Nxc6 Qxc6 12. Nf3 Nd7 13. Ne5 Nxe5 14. Bxe5 O-O 15. Be2 axb4 16. axb4 Bd7 17. O-O f6 18. Bg3 Be8 19. Rxa8 Qxa8 20. Qa1 Bg6 21. Qb2 Qb7 22. Ra1 Ra8 23. Ra3 Be8 24. Qa2 Bc6 25. Rxa8+ Qxa8 26. Qxa8+ Bxa8 27. f4 Bd6 28. Kf2 Bc6 29. Bh4 Be8 30. g4 h6 31. Bg3 Bg6 32. Bf3 Be4 33. Bxe4 dxe4 34. f5 e5 35. h4 Kf7 36. dxe5 Bxe5 37. Bxe5 fxe5 38. g5 h5 39. Kg3 Ke7 40. Kg2 Kf7 41. Kf2 Ke7 42. Kg3 Kf7 43. Kh3 Ke7 44. Kg2 Kf7 45. Kf2 Ke7 46. Ke2 Kf7 47. Kd2 Ke7 48. Ke2 Kf7 49. Kf1 Ke7 50. Kg2 1/2-1/2 

Answer 1

Leading up to move 12, black has gained space on the queen's side and can open a file with axb4.

E4 is an important square, and white's Nf3 move gives it up for the short term, granting opportunities for black. As a London player, I would be trying to contest that square and work up to a pawn push to e4 so I can begin to fight back in the center or on the King's side, while undermining black's queen's side pawn chain.

Once both players' knights assume their outputs, the difference between the two is that black's knight attacks c3, a key part of the pawn chain, while white's knight doesn't attack anything substantial. The black queen losing a tempo isn't too important, since it has plenty of squares to go to, and white will need to respond to the threat on c3 after the tempo.

If white defends c3 with the queen, both sides playing axb4 means that black controls the a-file, and white still has work to do to claim e4 and begin any sort of offensive.

If white defends c3 with the rook, they lose the a-file immediately.

Answer 2

The point seems to be that after 12 Nf3? Ne4 Black crashes through on the lightly defended Queenside. For example 13 Qc2 loses directly to axb4 14 cxb4 Bxb4+ because Ra1 is undefended. So 13 Qc1, but Black still has Qa6 followed by opening the a-file, e.g. 14 Rb1 axb4 15 axb4 Qa3: Black soon wins the c-pawn and then the b-pawn, and White will lose the house to the connected passed b/c pawns. (Black can also play it fancy with 13 Qc1 Nxc3 14 Qxc3 axb4, either getting three connected passers for the piece with a clear winning path bxa3, b4, b3 etc., or winning the ending after 15 axb4 Bxb4 16 Qxb4 Rxa1+ and 17 . . . Ra4 when White is obliged to offer a Q-trade on d6. Here too it ends with Black getting connected passed b/c pawns that White has no good way to block.)

Answer 3

Moving a piece twice in even a somewhat "slow" opening as the London can be counterproductive. Here the idea behind Nf3 is probably to waste more time with the aggressive-looking Ne5 which would achieve little (and, as Dr. Elkies points out, runs into 12...Ne4). That said, White's position seems difficult regardless but perhaps a more effective plan would be to either castle quickly or try to create kingside/central play with h3, g4, Bg2, etc.