I guess I'll do the start then, uh? The mapset is (obviously, congratulations once again!) very nice, yet I have some suggestions I'd like to mention in order to polish it. This won't take very long to get bubbled and ranked.
[Normal]
some missing new-combos, everything else seems fine to me.
00:33:283 (4) - NC ?
02:28:737 (3) - NC
02:30:392 (3) - NC
02:31:220 (1) - remove NC
02:32:047 (4) - NC
02:38:668 (3) - NC
02:45:289 (4) - NC
02:41:979 (4) - NC
02:45:289 (4) - NC
02:48:600 (5) - NC
02:51:910 (4) - NC
[Hard]
AImod says:
01:37:837 (3) - unsnapped sliderend
01:39:492 (3) - unsnapped sliderend
(probably happened during submission or during my editing, lol)
That's all. I found actually nothing I would change for real.
[Insane]
00:02:040 (2,3) / 00:03:695 (2,3) - inconsistent spacing. While the first patterns receives 2,02 DS, the second has only 1,81 DS, which has been slightly offthrowing. FOr the sake of consistency, you should decide for one and make it consistent here.
00:08:868 (3,4) - I actually dont understand why you change the pattern here. Variety? The song doesn't change, maybe you should stay with letting the first slider end on the red tick and let the other start on the blue tick,like you did before. I assume that your stacking indicated a change, but actually it made me think that it is stop-go movement, where the gap between (3) and (4) would be larger (timing-wise). I'd recommend to change the hitobjects-timing and unstack them like you did before.
00:12:179 (4,5) - ^
00:17:248 (5,6) - oh, this looks so nice patternwise, but I humbly believe that the repeatslider should start on the downbeat :/ Unfortunately the blanket would die then, because you have to shorten (5) for that. (this is pretty much optional)
00:20:662 (3) - (nitpicking) should be a bit more left to perfectly line up with (2)
00:22:938 (1) - ^
00:23:558 (5,6,7) - inconsistent spacing? (wasnt sure if I should mention it, I do it anyways)
00:37:835 (1,2,3) - oooh, that is hard to catch, because the way you placed it patternwise makes it difficult to read on the first attempts what you tried to follow. (and especially in which timing you placed the objects) I guess an easy fix would be to reduce the spacing between (2) and (3). The sliderfollow circle also just barely pokes the (3), thus making it a bit more closer to the sliderend would decrease the chance for sliderbreaking randomly. I am sure you can improve the readability here.