I believe pseudorandom is the best, but others may have a different opinion.
Honestly though fredted40x, a single pass will probably be good enough for you.
Have you found any evidence that a single pseudorandom pass is not sufficient ? Is there any forensic documentation you have found that proves it is possible to recover overwritten data from a “modern” hard drive after a single pseudorandom pass ?
I am not having a go at you here I just want to know if you know something we don’t.