It's probably not a linear relationship between total experience and chances of dying in the next hour of exposure. Also, there are too many other variables: you can read BFs to learn about riding strategies, think about your riding and decide to ride smart, you can choose your routes... and so on and so forth. However, if there is a plateau of experience, I know I personally am not anywhere near it. I'm continuing to learn things. And many cyclists more experienced than I told me on more than one occasion that it's the same for them.

That's true. Even in a city of several million there aren't enough cycling fatalities to draw significant statistical conclusions (which, of course, is a very good thing). One year we in Toronto might have five cycling fatalities, another year - only one. It doesn't mean though that the city became five times safer to cycle in.

I read a study somewhere that examined cycling crashes per mile. Crashes are much more common and, if one can properly track them, can paint a much more complete picture. Novices/casual riders were on one extreme, long-time cyclists who have taken a safety course were on the other. The difference between them was quite dramatic. Unfortunately, I don't remember what the study was or where the data came from.

City of Toronto also once did a study of reporting cycling crashes from one year. It's somewhat useful in some respects, but there is some important data missing. For example, it showed that 30% of crashes happened to sidewalk cyclists. However, the authors of the study do not know what percentage of cycling in the city is done on the sidewalk, so the figure is completely meaningless...