Not THAT unbelievable.
We should all remember that statistics are just that: numbers accounting for a kind of large numbers reality.
Riding IS statistically more risky to human beings than driving, both in probaility of occurrence and in consequences. This never changed along the decades of studies.
However, stats are largely irrelevant when it comes to indivual subjects. Between myself (obsessively 'active' (mostly) legal riding, bordering on paranoia) and the ordinary sc*mbag not giving a sh!t about his life nor other's (I know it's much more complex than that rethorical shortcut, but this post will be long enough already), the average risk per kilometer is several orders of magnitude away.
That's not to say I'll maintain my winning streak forever (rear-ended while stopped at a crosing once since 2001), for at least the two following facts:
- I AM doing mistakes. Despite all my good will on the road, there are portions where my attention is not 100%. This DOES increase the risk.
- Even when at 100%, I just can NOT mitigate every ill behaviour of every other road user.
We live in quantum reality really. It's all about probabilities and not one of them reaches 1 ;-)