This is the thing: it;s a random stochastic process. There is no pattern. It could happen again tomorrow - but given the frequency, that is extremely unlikely. Or it could never happen again. You certainly won't be due another after 17,000 km. But this I do know: your minds are very good at seeing patterns where there are none. Such clusters are actually quite common, more so than most people realise. I'd be very suspicious if there were no such clustering - though that's unlikely to be of much comfort to you!
That's some old school mansplaining right there.
Specifically, it's a Bernoulli process.
Imagine if I had a giant opaque lotto machine full of red and white balls. I want to know how many red balls there are. Obviously the best way to do this would be to open the machine and count them. But I can't do that, all I can do is draw out a single ball, observe it's colour, then return it and repeat. If the first ball is white, then all I know is that there is a white ball in the box. If the first 10 balls are white then it's probable that a majority of the balls are white. It's possible that there is only one white ball, and all the rest are red but that's unlikely. If the 11th ball is red, the only thing I know for sure is there are some red balls as well. Even if there are a million white balls and only one red ball it is equally possible for me to draw a red ball on the first try or the millionth, but of course it's much less likely to a red ball in the first 100 tries than to get one in the first million tries. It would not be foolish for me to suspect there is a high number of red balls, as the more there are, the higher the chances of getting one in the first dozen tries. That's the so-called demoninator effect, which worries someone after their first crash or reading about other people's collisions. But that's not the case here, I had - I'm guessing - 20,000 trials before I got my first "red ball", and a lot between that and my 3rd. It is possible it's a "cluster" - albeit one spread evenly over a long time - but it's more likely to be showing regression to mean, and the actual P value is quite close to 3/55,000. I was joking about having another collision within 17,000 km, but if someone opens a book on the chances of me being hit by a car in the next 55,000km*, the smart money would be on the 1-to-5-collisions bracket.
To add to that anecdata, I'll just say that in the 40,000+ miles I've done in the last 10 years, I've never been knocked off, for what that's worth.
Not a lot. Why should I be interested in the contents of your lotto machine when I have so much data from mine? Sure, maybe I have been unlucky but maybe you have just been lucky. Or maybe my P is bigger than your P <pause in memory of Fnaar>, because we are different people with different riding styles on different roads.
* Do not do that!
