Hello. I’ve lurked the site for a long time (years actually), and I decided to create an account and share some feedback.
I suggest that TASVideos shift its rules away from which types of games are allowed or not allowed. Instead, the rules should subject runs of games with certain central elements to increased standards of publication. This allows sidestepping pointless discussions about whether a certain game technically is or is not of a certain genre, whether a certain genre should or should not be allowed, and whether a past publication should be interpreted as a genre precedent while also addressing the valid concerns raised in this thread. Some games perhaps can never be TASed in such a way that this site should never accept, but it is unwise and almost certainly premature to blacklist a game—and certainly an entire category or genre—in this way. Under these principles, a rejection would mean only that the run is rejected, not the game itself, leaving the door open for innovation in the future.
More concretely, instead of saying that “educational games” are not allowed except when they are, the rules should identify more precisely the elements which cause “educational games” to be unsuitable for TASing. Consensus about which elements typically make “educational games” unsuitable for TASing will be a lot easier to achieve than about whether “educational games” should or shouldn’t be banned. Identifying whether a game focuses enough on these elements to warrant higher standards of acceptance will be trickier and will still be a judgement call subject to disagreement, but this should still be much easier than deciding whether a game is or is not an “educational game”.
I suspect that I’m only expressing what the community has understood intuitively but maybe has struggled to codify logically.
If I were judging this run, I would decide that the game does focus on elements that typically make “educational games” poorly suited for TASing enough that a run of this game must meet a standard of publication higher than simply finishing the game quickly. (I submit that PRNG manipulation alone is insuffcient to exceed this threshold.) Since this run shows that deliberately answering incorrectly is actually faster than answering correctly as the player should, I would decide that this run does indeed meet such a higher standard, although barely.
I hope this helps.