This movie was honestly really boring to watch, especially once I realized it's just the same minigames being played repeatedly, with slightly different themes/difficulty. It's clear to me, by the detailed submission notes and gameplay, that the author put a lot of work into this movie. Ignoring the possibility of memory manipulation for a moment, it does appear to be fairly optimal (though I'll admit I'm not a good authority on that aspect).
I also checked the speedrun.com leaderboards, after reading what CasualPokePlayer said about memory manip. For this 1 player 200% category, the time difference between the NoMM and MM world records, is 3 hours, 37 minutes. So it seems to me that this TAS could have been at least that much shorter, if MM was used.
Entertainment-wise, using MM would have also made this far more entertaining to watch. I read through
a forum thread that gave some examples of different MMs that are possible. And who knows what more could be done in the TASing environment.. There were also some periods where the player was waiting, motionless, for the level to end (e.g. waiting for the timer to count down). During these moments, it would've been more enjoyable if the player was doing
something, such as moving to the beat of the music, or doing something with frame-perfect accuracy just to show off.
While the submission was made under the "NoMM" RTA category, I'd have to ask: What benefit does not using MM have in regards to a TAS? Are there different strats/tricks that are unique and interesting to NoMM?