@MrDK, yes, jet transport accidents are all small sample statistics, fortunately. I agree with you it says something about air safety, namely that it got a lot safer since the 1960's. Also, that part of the statistics where you look at how many accidents occur per million departures, is obviously NOT small sample statistics. But anything to do with the accidents themselves, is.
Since about 2000, the statistical sampling is simply insufficient to provide trend information, and the occurrence of accidents takes on the stochastic character of randomness.
Look at page 18 in this report:
http://www.boeing.com/resources/boei...df/statsum.pdf
NOTE: these numbers take the increasing traffic into account, i.e. these are accident rates per million departures.
You can see that from 2000 onwards, just about the only information that can be deduced from those numbers is that the accident rate over several years in the US is non-zero. For the rest of the world, it is still higher than in the US initially, and that's where a trend can be seen, but from 2009 onwards, it appears to have reached similar levels as in the US.
Extending from this to taking lives lost in an accident as a safety indicator is misleading at best. You cannot take lives lost as a safety indicator when such small numbers of accidents are occuring, as the outcome will be random.