The real bidding problem on the first deal is:
What I often do in situations like this is to try to figure out the distribution around table is and then apply LoTT (with adjustments).
More specifically, before I apply LoTT I assume that the actual distribution around the table is close to the mode distribution(s) around the table or, rather, what I get when I try to divide each suit as evenly as possible consistent with the bidding and the requirement that each player get 13 cards. (If that's a silly method, please let me know.

) Here, for example, I get
N: 2434
E: 3235
S: 2533,
so I'd assume there are close to 18 total trumps. My poor trumps and (in my experience) also South's assumed 5T(332) shape suggest that total tricks may be lower than this, although rarely as low (16) as on the actual deal. So (by LoTT w/ adjustments) I'd assume 17 total tricks and bid 3
♠ (competitive, NF) accordingly.
On the second deal I assume the 2
♥ response was forcing to at least 2N. Then one possible agreement is that it sets up forcing pass to 2N, and North's pass over the 2
♠ advance would actually be forcing. It doesn't seem like you have that agreement, so in this position
you only had to decide whether to raise or not. You passed, which looks kind of defensible in view of the poor trump holding and the K of their suit (collectively suggesting total tricks < total trumps), so now your partner got the following problem
instead. LoTT would have told him not to pass, of course.
LoTT (or a corollary to it, "bid to the level of your fit") would also have told you what to do in this position
on the third deal, and your partner what to do in this position
taken from a very recent thread.