Pondering how the three-way multivibrator worked in the way home, I decided that the base waveforms must be overlapping triangle waves – which would give a good reason for the nice 120° output waveform, and a reason for it to be predictable. Sadly, my speculation was not confirmed – instead the bases look like this, …
This story continues at N-way multivibrators, and an increasingly confused ex-engineer
Or just read more coverage at Electronics Weekly