Three proof games

Here are three proof games. For each position, construct the only game that leads to the given position after exactly the specified number of moves.

(Note: The only correct things about each position's FEN is the setup and the player to move. In particular, castling rights and whether en passant is possible are not necessarily correct.)

R. Muller
Rochade, 1985

G. Donati
Problem Paradise 10, 1998

Second puzzle is beautiful!

Third puzzle looks too difficult.

Third puzzle's hint (highlight): [The shortest path for White to reach the desired position is 30 moves.]

First puzzle's answer (highlight): [1. a4 d6 2. a5 Bg4 3. a6 Bxe2 4. axb7 Bxd1 5. bxa8=Q Bg4 6. Qf3 Bc8 7. Qd1]

I'm actually still not sure how to effectively solve short PGs besides largely trial and error. Unlike longer PGs where you can analyze them with retro deductions as in other retro puzzles.

I think it all depends on how you define "effectively". Even with the limited experience I have with solving these, I can say pretty confidently that shorter PGs will be solved significantly faster by nearly everyone, regardless of proficiency. I did need ~10 minutes of trial and error for the first puzzle, but when I found the solution, it wasn't even connected with anything I had thought about or was thinking about at the time. The "missing rook, missing knight's pawn" pattern just leaped out of nowhere and hit me in the face, and that was it. Afterwards, I was a little disappointed that I hadn't solved it in seconds - I mean, looking for all possible promotions is how I normally begin solving! 

Yeah, that's what I meant; there doesn't seem to be any logical way to deduce, or even to guide, you to look at the possible promotion.