You are right. My point still stands. On collision debris might have significantly different velocities than the original objects thus have different orbits.

> One satellite smashing into another would not add sufficient energy to the debris to alter its speed significantly.

I think that's completely dependant on the mass of the specific particle of debris. The smaller the particle, the larger possible speed change attainable in a messy collision.

I'd really prefer that there was some scientific study simulating possible results of impact between two SpaceX satellites. Because there's non-zero chance of that happening. For example if there are some errors during the deployment of new batch.