Alternatively, White has the opportunity for an interesting gambit: 4. Bc4 cxb2 5. Bxb2, after which White is two pawns down but has both bishops already out.

The line nuntar mentions is a fun line. My only example on Gameknot I had the Black pieces gameknot.com. Otherwise the only other danish gambit is this one: game

It's not so good here, because White is losing time after having gambitted a pawn. Better is 5. Bc4.

I think I would have played the straightforward 5.Nf3, and let Black exchange on c3. For the battered Q-side structure, white would have gained a little time, Black having to begin all over again to develop his game.

Yes I think I should have kept with the opening and tried to ensure no piece survived. I like your option Ion

There is a flaw in jkarp's analysis: 6. Nd5 Bxd2 is check, so Black has time to save both bishop and queen. However, White additionally threatens the c7 pawn, so Black must meekly retreat with 7... Qd8, leaving White well ahead on development.

Joe is quite right: 6.Nd5! was the right move, and a killer, pretty much. The plot goes like this: 6.Nd5! Bxd2 7.Qxd2 ... and Black has to move the Queen and protect c7. There are three possible defences: [A] 7...Qd8 8.Qc3! ... attacks c7 again, and g7 as well! 8...c6 and now White gets the advantage by 9.Qxg7 (see Line C), but better still is 9.Qe5ch! Kf8 10.Qd6ch Ne7 11.Nc7 ... and Black has a miserable game; [B] 7...Qc6?? 8.Qg5 with a winning game for White; [C] 7...Qd6 (best) 8.Qc3 c6 9.Qxg7 ... (White doesn't have the Qe5ch option) 9...cxd5 10.Qxh8 Qb4ch (say) 11.Kd1 Kf8 12.exd5 ... When the the exchange plus pawn extra gives White a near-on winning edge.

Not a good place for the queen, as she is blocking the development of Black's other pieces, and vulnerable to tempo attacks.

Best for black was to leave the Queen on f6 (hoping White will exchange it) and develop the Q-side, by ...Nc6 or ...d6. The Queen is starting already to cost Black time, and will continue to do so.

More forceful would be 7. Bc4, attacking the weak pawn on f7.

Long castling (7.0-0-0) would have forced black to think about his Queen again.

That Queen is costing Black huge.

A nice double-purpose move, also defending against Black's threat of 9... Bxc3+ 10. bxc3 Qxc3+ forking king and rook.

Very good move - so good hardly anything else came into consideration. White's pices are active and mobile, and castling, available on either side, will have Black guessing.

I agree Black was probably better off keeping his bishop, but I don't think it is yet fatal.

Nxd4 was not possible, as White's king was in check.

Either recapture left White with a big edge. The text is fine; Nxc3 was equally strong.

Harassing the queen is only worthwhile if White gains tempo (moves her pieces to more useful squares) during the attack. Here, for example, supposing Black played 11... Qd6, the attitude of "it's there, I must attack it!" could lead to 12. Bf4 Qb6 13. Be3 Qd6 and an early draw. I would suggest instead either 11. 0-0, to complete development while Black has barely started, or maybe 11. Rb1 -- harassing the queen but with a definite purpose, to reduce her available squares and (depending on where she goes) threaten either to trap her or win the pawn on c7 at least.

(!) Now Black's queen is in real danger of being trapped in a couple of moves.

Only move. Otherwise White plays 13. a3 followed by 14. Rfb1, winning the queen.

12...Nc6 13.a3 Qb3 and escapes.