KGD: Falkbeer Counter Gambit

Question on the Falkbeer:

1. e4 e5 2. f4 d5 3. exd5 e4 4. d3 Qxd5 5. Qe2 Nf6 6. Nc3 Bb4 7. Bd2 Bxc3+ 8. Bxc3 Bf5? 9. Bxf6

Removing a defender of e4 and damaging Black's kingside pawn structure.

9. ... fxg6 10. g4 Bd7 11. dxe4 Qa5+ 12. c3 Nc6

Ordinary move is 8. ... Bg4 9. dxe4, so Black must have played poorly here. Despite being up a pawn and having the superior structure, White cannot castle long without doing something about the a-pawn (perhaps sacrificing it) and his kingside remains undeveloped. Moreover, after O-O-O/Re8 Black is ahead in development and is not without threats. I drew this game, after 13. h3
Protecting the g-pawn before Nf3.

13. ... O-O-O 14. Nf3 Rhe8 15. Nd2 f5 16. gxf5 Bxf5 17. Nc4 Qa6 18. Bg2 Re7 19. Kf2 b5 [19. ... Rde8 20. Nd6+ cxd6 21. Qxa6 bxa6 22. exf5] 20. Nd2 Red7 21. Rhd1 Be6 22. a4 Qb6+ 23. Kg3 bxa4 24. f5 Bb3 25. Nxb3 Rxd1 26. Rxd1 Rxd1 27. Qxd1 Qxb3

What do you think of White's play in the opening? The main line of the Falkbeer, as I have said, is inferior to the position after 10. g4.

Where did White go wrong in the middlegame? First, White's inability to catch up in development resulted in a poor king position and cramped development. Second, White's large center became a liability quickly. In particular, the black rook opposed against the white K+Q on the e-file proved very annoying. Perhaps White should have played to trade off queens . . . would it have been worth giving back the pawn? Seemingly, White could have worked from here to obtain an advantageous endgame position.

Any endgame thoughts?

All help is appreciated.

castle and then try to develop, also exchange pieces