Middlegame or Ending? aka Bad Wiessee (part seven)

SOME positions are obviously in the middlegame and others have clearly reached the ending, but inevitably there is a grey area where arguments can be made on both sides.
I have found a useful rule-of-thumb is what I call the 13pt Rule.
I cannot recall where I first saw it, but the basic idea is this: a game has reached the ending when the points value of each side's pieces (not pawns) is under 13, using the traditional scale (ie queen-9, rook-5, bishop/knight-3).
Not everyone agrees - chess is too complicated for a strict dividing line to hold complete loyalty - but over the years I have found the 13pt Rule to be a useful guide.
Take a look at this position, which arose today in round eight.

Boris Miskevicer (2074) - Spanton (1878), after White's 17th move

Would you call this a middlegame or an ending?
Under the 13pt Rule, it is clearly the former as each player has two rooks and a bishop - exactly 13pts.
But I can well understand many people reckoning the position has all the characteristics of an ending, with the possible exception that there are six pawns aside.
Anyway, see how the game proceeded and make your own mind up.
17...Bc5 18.Re1 f6
Playable is 18...Bxf2 19.Rxe5+ Kd7, but I felt it accelerated White's development while leaving my pieces less-well coordinated.
19.f4 Kd7!?
I rejected 19...Bd6 because of 20.Be3, which forced, so I thought, 20...b6 or 20...a6. But the analysis engine Stockfish9 gives 20...Kd7!, and if 21.Bxa7, then 21...Ra8 22.fxe5 Bxe5, with compensating pressure.
20.fxe5 fxe5 21.Bh6!?
This is Stockfish9's choice. The point is that after 21.Rxe5, Black has 21...Rhe8, with lots of activity based on Black's superior development. Even so, it is far from clear that Black has full compensation for the pawn sac.
21...Rhe8 22.Re4?!
I was concerned about 22.Rf1, when 22...Rc7, 22...e4 and 22...Bf8 are all serious candidates.
The problem with the text is that White's e4-rook is rather inflexibly placed, which in turn makes his passed d5-pawn more of a weakness than a strength.
22...Kd6 23.c4?
I expected 23.Rd1, which my main analysis engines reckon is equal. The text leaves White vulnerable to, believe it or not, a middlegame king-hunt.
23...Bd4 24.Rae1 b5! 25.cxb5
Engines prefer 25.Be3, but still reckon Black is winning after the obvious 25...bxc4+.
25...Kxd5
Even more convincing is the immediate 25...Rc3+.
26.R4e2
26.Bd2 covers the c3 square, but after 26...Re6 White's queenside quickly collapses, and White is unable to muster any meaningful counterplay.
The game finished:
26...Rc3+ 27.Kb4 Rec8 28.a4 28...R8c4+ 29.Ka5 Ra3 30.Ka6 Raxa4+ 31.Kb7 Rcb4 32.Bg7 Rxb5+ 33.Kc7 33...Rc4+ 34.Kd8 Rb8+ 0-1
Meanwhile, here is a puzzle from outside of the tournament hall that I have not been able to figure out:

An impressive carving by the Tegernsee in Bad Wiessee, but is it an effeminate god or a butch goddess?