You need only the fact that the Poisson Distribution is wonderfully scalable.

We have 1 dent every 2 plates.

This translates smoothly to

1/2 dent every 1 plate or

15/2 dents every 15 plates or

50 dents every 100 plates.

Of course, at some point, often around 30 or so, it may become burdensome to calculate what you want using the Poisson directly. That's where the Normal approximation comes in.

Let's see what you get.