When is the middlegame over and the ending starting?

Currently, I am working with a DVD series on "Winning Endgame Strategies" (Kuljasevic), Soltis famous "Grandmaster Series - Endings" as well as Shereshevsky`s "Endgame Strategy".   All of these experts hardly care about any theoretical lines. (Hardly; of course, there are a few standard rook endings to know. But that´s it. It is, by no means, about the 20, 50 or even 100 most important endgame techniques or something alike.) For them, it is all about plans and schematic thinking.

However, one question that strikes me: When is the endgame beginning?

According to Soltis - referencing to Belavenets, as soon as the queens have left the board. Whereas Romanovski´s definition depends upon the king: as soon as the king assumes an active role, he calls it an endgame. Glenn Flear (Practical Endgame Play - Beyond the Basics) defines it as the final phase of the game, where significant simplification marked the end of the middlegame. Many experts argue that it is about "reduced material being left on the board"...
Mednis tried it the other way around and approached it from the negative: Endgame starts, when it is not anymore about better development or space advantage (but about pawn structures).

Of course, each of the above definitions can be falsified easily. Nevertheless, none is completely wrong. No doubt, it is not clear cut and it doesn´t matter how many more (past and present) experts one consults, it remains somewhat fuzzy and foggy. It simply "depends"...👦

Its recognition, though, seems vital! And there is one thing they all agree - and even Wikipedia "knows" it: Middlegame and endgame show different characteristics and require different strategies! E.g.:

• The king becomes (much) more active.
• It is no longer about better development.
• It is no longer about "more space".
• Wing pawns gain in value.
• Piece value in general changes (as opposed to the middlegame); especially pawns.
• The better side shall trade pieces but not pawns (whereas in the in middlegame it is rather vice versa).
• Initiative becomes vital.
• It is hardly about precise calculation, but about decent plans --> schematic thinking.

Honestly, I never really thought about this (and definitely not the way indicated by the above bullet points). For me, endgames were always about deep and accurate calculations - something I am bad at; about pawn races and rooks to be placed on the right square.
Hence, better finish things off before the endgame or accept a draw, if in doubt.  Right?!
No, not anymore - thanks to the authors mentioned above!

"Studying the opening is just memorizing moves and hoping for traps, but studying the endgame is chess." – Joshua Waitzkin