Of course it may be disputed in which way to measure exposure when comparing risks for different modes of transportation. Distance is often chosen, as it may be argued that this is a fixed value.

But this is not always the case. Actually, in many situations you may choose your destination according to your preferred mode of transport. Shall I take the bike to my local grocer, or shall I take the car to the far-away supermarket? Shall I walk to the near-by park, or shall I go by car to the distant recreation area? Shall I choose a living close to work, so I can go by bike, or shall I settle in Suburbia and commute by car?

So accidents per hour or per trip is in many cases a more relevant measure than accidents per kilometre.

Moreover, risk per kilometre is not as obvious as it seems. The national averages for cycles versus cars hide big differences in local environment. Cycle data consist mainly of urban trips in complicated traffic situations, while car data to a great extent consist of rural or highway kilometers with low accident rate. If you compare data for the same trip, and include accidents that cars incur to other road users, the difference between car and bike decreases substantially.

And then, of course, the most dangerous thing you can do is sitting still by your television set or computer. Per kilometer, the risk is infinite!