Spikes of Vacancies in cumulative PKA cascades

Hello everyone, Greetings.
I am trying to do a cumulative PKA damage cascades and exploring few options ahead.
I have run a 300 consecutive damage cascade on 72x41x45 hcp Zr substrate by taking a PKA randomly from a spherical shell, providing velocity directing towards center of the box.
After this, extracted nos of vacancies for each cascade and obtained following graph.


I am not sure why the spikes occurred. I also ran another cumulative cascade which also turned out quite similar to this one.
Can anybody guide in this regard?
Thank you for your time.

I think that he spikes are likely not an error by themselves; they are probably the natural result of rare high-damage cascades in a highly stochastic irradiation sequence

Thank you for your reply. Would you elaborate a little bit more? The spikes are bit too high to corelate actual physical phenomena.
Additionally, should these be regarded as random error and filtered out?

You probably shouldn’t filter them out. From what I know, big spikes can be real rare events from channeling or local damage overlap or even boundary effects. You should only remove them if they came from artifical or bad run