Originally Posted by
ATC Watcher
Let's wait a bit more to see if it is a pattern or just a anomaly.
Oh please. Let's not wait.
The statistics of the occurrence of rare events (search for Poisson distribution) is a complex subject. And not easily understood by those not versed in the mathematics. It can take a significant sample of these to determine whether or not there is an underlying change in the driving datasets. And since these events sometimes manifest in the form of significant losses of capital equipment and lives, can we afford to sit around and count them?
The aviation industry (at least back in my time) preferred to do a root cause analysis on each one with the goal of eliminating the cause. Even if that meant pushing the economics of the fixes to the back seat. Just to see if things would regress back to the mean or not.
There are mechanisms in place to capture near misses and use this data to extrapolate to the expected frequency of more significant events. With the goal of identifying underlying adverse system changes and fixing them before the next big tragedy. Again, this is a complex statistical problem in that rare events can cluster and give the appearance of trouble when none really exists. I hope that the FAA and NTSB have some smart math geeks concentrating on this data.