(I've been skiing, so I'm still very behind - that's ok.)
Well, this was a bit annoying. Basically, Part 2 we had to make up a definition of what a Christmas tree might appear like, and then test for that made-up definition to see if we made up the right thing. I ended up testing for "step count with largest number of robots in a single row" - other definitions go by "step count with lowest safety value", "step count with lowest entropy", or sone kind of clustering / nearest-neighbors approach. All make sense - all are a bit irritating. I think the main thing I'm taking away from this day is that part 2 should probably use part 1 somehow, even if they seem totally disconnected.