The only thing I can think of, other than the character being ironic, is that truly random events will tend to bunch up more than people think they will. For instance, if you flip a fair coin a whole bunch of times you will get enough long stretches of heads in a row to surprise most people, who think of random as being something more like “uniformly distributed.” Way back in the day they used to generate random numbers for scientific and statistical applications by having row after row of people sitting at desks with ten-sided dice and rolling them and recording the results, which were then published in tables and sold. It turns out (or so the story goes) that a lot of these tables were biased, because the people doing the rolling were throwing out long sequences of the same number, or obvious patterns like 1, 2, 3, 4, 5…instead of faithfully writing them down, because they thought they weren’t “random” enough!

I caught that last night too. I figured some writer screwed up their technobabble vocabulary and forgot what “random” means. It might have made a bit more sense if he’d said “random events CAN cluster together”, which is what we call “coincidence” (and what I think the writers were driving at in that exchange).

I enjoyed the episode, but I feel like the show is kind of treading water with the character relationships this season – they had a great mix on the last few episodes last season and this season it feels to me like the writers are going out of their way to try to keep everything more or less static.

I’m not an expert in randomness (though looking around, I seem to generate a lot of it… well, that’s really more chaos), but Joshua’s got it right. If you deal with randomness, you will expect to find clusters not because randomness generates clusters but because true randomness fails to avoid clusters, and thus clusters occur.
If you Google “Clustering Illusion”, you’ll find explanations about the tendency of humans to note clusters and then assume non-randomness in random situations.