_ _ _ _ _ A _ _ _ _ _
Then, perhaps the Devil goes here: _ D _ _ _ A _ _ _ _ _
Angel moves: _ D _ A _ _ _ _ _ _ _
Devil places: _ D D A _ _ _ _ _ _ _
The angel must now change direction. For any given direction and any given angel power, as long as the devil starts placing pieces far enough away he can force the angel to change direction.The difficult part of the problem is proving your argument to others. There-in lies the difficulty of mathematical proofs.
Conceptually, the devil will start constructing both traps at the same time at half speed, and decide which one to put more work in as the angel approaches.
Nope. Even if you restrict the angel to 3 directions (up, left, right), the devil has a winning strategy. This is mentioned in the linked article:
> If the angel never decreases its y coordinate, then the devil has a winning strategy (Conway, 1982).
There's a simple, informal proof in the references:
Conway, H. "The Angel Problem" http://library.msri.org/books/Book29/files/conway.pdf
If the Angel backtracks, this is good for the devil because it's amost like getting a free move. You can imagine that if the angel moves right and then moves left, he's back in the same place but the devil has place 2 pieces. So the Angel is a lot more restricted than it first appears.
But I see it considers movement as multiple movements of a chess king, so you're right. It doesn't have that restriction!
Starting off, if the angel can move 1 every 1/9th of a turn, a devil could block off a 3x3. Every subsequent full turn would add 2 that the devil would have to block off.
So for move rate M on an NxN grid (which can represent any wandering linear path) necessary to trap an angel is 8+2NM (interestingly, close to 13). If an angel moves at all, the devil cannot trap it without player error.