Missed Chance at Copycat Immortality

Just for fun, I decided before the following game to play it in the style of Capablanca: straightforward, simple. This is almost the opposite of my normal style, more Kasparov-like, of taking the fight to the opponent, setting him problems, and seeing if he can solve them. On move 16, I threatened to capture on d6. How would you have defended for Black - what 16th move would you play?

Black played 16...Bxe5 and the game continued (click on Black's 16th move):

A pretty good game! But I was sorely disappointed that Black had not defended with 16...Rd8. Because then I was ready to spring...17.Bxd6 Rxd6 18.Rxd6 Qxd6 19.Rd1 Qb6 (if 19...Qc7 20.Qe5! Qb6 [Of course, if 20...Qxe5 21.Rd8# is the recurring theme of this blog: removal of the guard of the back rank, by deflection] 21.Qc5!+- as the back rank falls), which was the main line I saw, but 19...Qb6 requires similar means, as computer analysis confirms: 20.Qe3! Qc7 21.Qe5! Qb6 22.Qc5! and again the back rank falls, so White wins. [corrected]

However, I can hardly take credit for this idea, although I found this variation during the game. One of the most famous combinations, which occurred in Adams-Torre, is an even more complicated version of this, and I was quite aware of that pattern during the Pennig game:

The famous line went 18.Qg4! Exploiting the same back-rank theme seen in my game (but my game was 52 years later...) 18...Qb5 19.Qc4! Qd7 20.Qc7! Qb5 21.a4 Qxa4 22.Re4 Qb5 23.Qxb7 1-0

So I missed my chance at copycat immortality . However, I got to put my game in full, with notes, in the Appendix added to the 4th edition of my first book, Elements of Positional Evaluation.