I'm no mid-dan, 1k myself IRL, but one commment on the top left corner/top side anyways (I agree mostly with your analysis on bottom before that).

I don't like 33+35 for B at all. Seems like really asking for trouble. First idea that comes up instead is standard sente 33-point invasion in upper left (rarer variations for W don't look that good at a glance), and extension on right side (probably Q10) after that. As a refinement, I would look into whether I could play M3 before the 33-point invasion - I don't have enough strenght to say anything definite about that though.

Finally, if I for some reason played 33 at H17, I would follow up with H15, not M17.

Thanks for the comments. H17 is maybe an awkward spot for an invasion in this position, but in this kind of double star point double wing formation, what are the alternative invasion points other than 3-3? I think I can agree with you that M17 might have been a bit much, mainly because I couldn't afford to treat H17 lightly in the game.

Whats the reason to look for alternative invasion points? 3-3 looks ok instead of H17 (golem7 had similar idea in his variations...)?

The clamp at is a nice move. But after W connects, which B stone is most important, F5 or E6? Clearly F5, right? B must protect this cutting stone to attack the three W stones which have been cut off. So for move 27, F6 seems like the obvious move. The moves played in the game force W to take a large territory on the left and do not do much to protect the important cutting stone.

is the wrong direction. B must strengthen the still weak center group; it is premature to take away the W base, because when W jumps into the center, his stones are stronger than the B cutting stones. In fact, W found such a good move that B gave up on those stones. This was clearly a failure for B. If instead B played something like G5-H5-H6-J5-G7, it would be a reasonable fight.

The invasion at is fine. I do not much like though. It seems like the wrong direction, trying to make territory near strength, while giving the opponent an extra move to attack the weak invasion stone. Jumping out to H15 directly seems better. Still, this move does make some territory and keeps W from making a base, so I cannot criticize it too strongly. It just feels like an overplay to me. Not as clearly bad as , but much the same idea.

seems like a reasonable alternative to invading the upper left corner. Actually, I like this move better. But is not a good continuation. Do you really think you have a chance to kill W here? If not, this move will end up as a very small play. Much better would be to expand the right side with a large move like N4, making full use of previous moves. This move would also be harder for W to ignore, as it threatens to capture the W stone on a fairly large scale (compared to M2 and N2 in the game).

For an invasion of the upper left corner, I like C16. It takes a bigger corner than C17 if W blocks on the outside, and it breaks up the left side nicely if W takes the corner. The B stone at F13 should make it fairly easy for B to live or escape. The game move makes less use of the existing B stones in this area, and in the end left W with too much territory here.

Tewari analysis on black 33-35:
Imagine if black played m17 immediately, and white responded with k17 (both moves would be kind of OK in this position).
Then black h17 would obviously be a very bad move (and white f17 a good answer to it).

I am a very logical person but I've often had difficult understanding some tewari analyses. It's clear that if black played M17 directly and white responded with K17, H17 would be a bad invasion. But if black played M17 directly and white responded with F17, H17 would be less-obviously a bad move.

To make a deduction from a tewari analysis, do we only need to create one possible resequencing of the moves in question that gives a bad result?