I have some apologies to give as well- I should've provided a link for that correspondence game:
www.iccf-webchess.com/GetPGN.aspx?id=93060 The game in full:
White: Israel De la Calle (2302)
Black: Prof. Mario Valverde Lopez (2059)
60yearsBdF-WS25, ICCF, 15/11/2006
1.d4 e5 2.dxe5 Nc6 3.Nf3 Qe7 4.Bf4 Qb4+ 5.Bd2 Qxb2 6.Nc3 Bb4 7.Rb1 Qa3 8.Nd5 Ba5 9.Rb5 Bxd2+ 10.Qxd2 Kd8 11.e4 a6 12.Rb3 Qxa2 13.Be2 Qa1+ 14.Bd1 Qa5 15.Qxa5 Nxa5 16.Ng5 Nh6 17.Rh3 Rf8 18.Nxh7 Rg8 19.g4 d6 20.exd6 cxd6 21.Rh5 Nc4 22.Nhf6 Rf8 23.Be2 b5 24.Bxc4 1-0
After 8.Nd5 Ba5 9.e4 Nge7, I have looked at three alternatives against 10.Bc4, and think 10...Qa4 may be better:
A) 10...0-0 11.0-0 +/- (11...Ng6 12.Rb3! +-)
B) 10...Qc5 11.Bb3 Bxd2+ 12.Qxd2 Nxd5 13.exd5 Na5 14.Ng5 +/- is similar to the main line with 10...Bxd2+,
C) 10...Qa4 can lead to a long, forcing line with 11.Bd3 Bxd2+ 12.Qxd2 Nxd5 13.exd5 Qxa2 14.c4 Qxd2+ 15.Kxd2 Na5 16.c5 b6 17.d6 cxd6 18.exd6 0-0 +=, when White can regain the lost pawn but I don't think the pawn on d6 offers many winning chances. Or 11.Bxa5 Qxa5 12.Qd2 (12.Nxd2 Nxd5 13.exd5 Nxe5 14.Qe2 Kd8 with ...Re8 to follow, is pretty unclear) 12...Qxd2+ 13.Kxd2 Kd8 +=.
On the other hand I think 10.Rb5 should lead to +/-. I agree that after 10...Bxd2+ 11.Qxd2, 11...0-0!? 12.Nxc7 Rb8 may be a better practical try, in view of 11...Nxd5 12.exd5 Nd8 (12...Ne7 13.d6 cxd6 14.cxd6 Ng6 15.Bc4, 13...Nc6 14.dxc7 a6 15.Rb6) 13.Bc4. It's worth mentioning that 12...Qxa2 almost works for Black here, but after the accurate 13.dxc6 dxc6 14.Rb3! (otherwise Black gets a perpetual), Black's two pawns should not be enough for the piece. Indeed I am tempted to suggest that 9.e4 is superior to 9.Rb5 here.
Regarding 8...Bxd2+, objectively it may be no worse, but I'm not sure that White has more chances to go wrong there, judging by the article- at the very least there are numerous ways to get into an ending where Black is suffering and begging for a draw.
Interesting about the lack of 4.Bf4's at high levels. It would also be interesting to see the proportion of 4.Bf4 games that continue with the horrible blunder 6.Bc3, as those might be inflating the percentages at lower levels. In my various games with the gambit (mostly casual/blitz but also with one tournament win over a 1700-rated player) White played 4.Bf4 only about a dozen times, and on all but one of the occasions, followed it up with 6.Bc3. The one game that continued 6.Nc3 went 6...Bb4 7.Nd5?! Bxd2+ 8.Kxd2 Kd8, when I suspect Black is already at least equal.