In my interpretation of the rubric, I would think #4 earns the point since it's pretty clearly one long algorithm.
The rest might depend on how finicky College Board gets with what's considered an algorithm. Strictly speaking, students should be identifying an algorithm, not showing a portion of their program and getting a point if there's an algorithm "someone in there".
Sample #1 contains an algorithm "somewhere in there" in the first onEvent, but that's not identified as the algorithm. By selecting that entire piece of code as the algorithm, the student is claiming that is an algorithm, which I would argue it's not - a series of onEvents is not an algorithm.
Sample #2 also contains at least one algorithm "somewhere in there", but I would argue that specific selection of code is not an algorithm. There's an algorithm in the onEvent and an algorithm outside the onEvent.
Similar with Sample #3... I see two separate algorithms.
I wouldn't be too concerned yet, since it really comes down to how they parse this out at the reading during the summer. Before you posted this, my initial interpretation was "meh, everything's an algorithm, as long as it's got at least two lines of code! weee!..."
I'm sure this point can be interpreted different ways. Sorry I'm unable to give you a non-definitive answer. :o)