I've noticed in my limited experience that some caches seem to be more prone to trackables going missing that other caches. Personally I try to help a trackable along its goal and I would be willing to take into account how likely the next cache I place it in is to be a good home.
I believe there is sufficient data to provide some insight into what percentage of trackables go missing once they arrive in any given cache. The statistics could be updated to include the recent trend vs the entire history as well as updating when a trackable that was thought to be missing is in fact later recovered somewhere else. Sure, the statistics wouldn't be perfect, but they may provide some insight in behavior around a certain cache.
Another idea would be use to data analysis (who viewed what caches) to try and determine what geocaching users might be keeping trackables that they shouldn't be. This would be no exact science, and since it would not be prove-able, I'm not sure what you would do with the results. SO maybe we'll table that one for now...
If you are looking for volunteer programmers to work on this kind of thing please email me, it would certainly be an interesting project. What about making all the data open, or at least what might be considered public data?
-lex3001