@23
You posted:
"I ran SF 15.1 (no NNUE) against itself from the initial position in question in Arena with 5 mins for all moves each (repeating)."
That is 1 game by one SF 15.1 entity piloted by a 860 rated human with 5 minutes for all moves.
All automatic. Not piloted by any human, 860 rated (whomever you might have had in mind) or otherwise.
I don't draw any conclusion from the game, merely use it as an illustration that the assumption of messrs. @VerifiedChessYarshe and @saderfewpro_SHB that the position is drawn is questionable.
I posted results from 68 games by 17 entities piloted by ICCF (grand)masters with 5 days/move.
Why?
They do play perfect games.
I don't believe you.
There are innumerable instances of similarly strong players (human and silicon) playing imperfectly in much simpler positions that can be checked by tablebases. It defies common sense to assume they play better as the positions get more complicated.
The 3-fold repetition rule is the major drawing mechanism.
I said your conjecture was interesting. I didn't say wrong. Feel free to elaborate.
@23
You posted:
"I ran SF 15.1 (no NNUE) against itself from the initial position in question in Arena with 5 mins for all moves each (repeating)."
That is 1 game by one SF 15.1 entity piloted by a 860 rated human with 5 minutes for all moves.
I posted results from 68 games by 17 entities piloted by ICCF (grand)masters with 5 days/move.
They do play perfect games.
The 3-fold repetition rule is the major drawing mechanism.