How does NL differentiate between:
- Your replay vs someone else's - maybe the tick was erased (or rather, turned orange) because the level has been modified, but your replay still solves it; in this case the green tick (and records) should definitely apply.
- A replay that solves the level, vs an attempt replay (maybe your own) you used as a starting point then finished off the rest manually - to some extent this could be determined by "has the user made any action since the replay ended / was cancelled", but it would get very complicated to determine whether the action is meaningful (maybe the action was assigning a leftover floater to a lemming that doesn't ever actually fall again); and even more of a problem, how does it differentiate between "finishing off an attempt replay" vs "cancelling a solution replay before the last couple of assignments, then doing those manually". One suggestion may be "the replay file should include data on whether the replay solves the level or not" - except, this data wouldn't be present on replays that are saved before they complete the level (but require no further action beyond waiting to do so).
- Someone else's replay that solves the level, vs your own replay that you've edited by hand. Given that there's no result that can be produced this way that can't be produced by playing the game directly, and that NL's focus is on solving rather than execution, I would argue that it's legitimate, even if it may be very unconventional.
Not to mention, replays affecting progress / records / talismans is extremely useful when creating content (talismans in particular, as the mass replay checker's output itself doesn't indicate anything to do with talismans - but one can verify talisman replays in a mass check, by first clearing their save data, then running the mass check and seeing which talismans do / don't get unlocked).
Don't get me wrong, I get your point for sure - but I don't see any practical way of actually fixing it without causing even bigger problems as a result, other than maybe manually asking the user "do you want this attempt to count in records / completion" which would get annoying.
