H1. 100bbs deep I just check my option pre most of the time from BB. Iso'ing is ok in theory (bc it's so likely that you have the best hand, etc) but I prefer hands that play better post. Imo it doesn't make sense to play big pots (especially OOP) with marginal hands...99 is actually a "good" hand but the vast majority (~88%) of the time, you will not flop a set and (~80% of the time) there will be at least 1 overcard on the board and you will be in a marginal spot (even the <1%
of the time when you are HU.) I'd rather iso something like A5s, which doesn't flop much better but still plays very well in raised pots (especially when it's multiway.) You can semibluff or check and have enuf equity to continue sometimes with as little as a bdfd (b/c it's to the nuts)+bdsd (wheel draws are always very well hidden, which increases their implied odds.)
As played sizing pre is ok...I just use a formula of 3x+1bb per limper and I always add 1bb when I'm OOP. When you bet 4k into 12k, I think you have to continue vs a smallish raise (and hope that the other 2 callers fold) and call again on a lot of turns. You're really only losing to 1 combo of 22 and A2s for SB (since 44 always slowplays.) If the villain is indeed tight, he will probably check it down and you will win a good % of the time. I would CBet bigger tho...~60% pot.
H2. Once you check the flop you really shouldn't be doing much leading on future streets. Check raising is fine otf at a low freq and ott but as played, I would probably rather go for the c/r otr on a safe board like QJ575xx. I will say that you ended up getting max value by leading vs AJ but you also will just fold-out all the missed draws that could elect to bluff. If QJ5 flop had a FD then I would Cbet like 2 combos of top set when it's multiway.
H3. I limp behind almost always but bet bigger otf and ott. FWIW I think this is a better spot to iso than H1, but still not that great with this holding. I would DEF 3bet the turn and iso btn (results aside - this is by far the biggest mistake imo*.) The pot is already big and your hand needs some protection with the bdfd and 3 to a str8 on board. There's way more value in denying equity to the BB than there is in slowplaying since the board can have a lot of bad runouts where you will either be beat or it will be much tougher to get value. As played the board otr is: 56J73xxx...just check back, although it might be closer if BB is the type that is likely to lead when they complete their draws. If you're gonna bet for thin value on 4-str8 boards then you gotta really size down to get value from 2pr (usually ~1/4 - 1/3 pot is good.)
* I was just calculating the potsize and it should be ~15k ott, so if you bet ~7k and btn jammed for <14k then it wouldn't reopen the betting afaik. If you bet 5k (which would be 1/3 pot,) the 12k jam would reopen the betting.
Re: the implied odds theory question, there are some good older threads where this question is debated at length. I believe that the conclusion (that the good players came to) was we need a minimum of 15 to 1 implied IP and 20 to 1 implied OOP to profitably setmine. Check it out tho.