What really matters to us is knowing the probabilty of a failure happening to any particular car. I would like to know how the chance of belt failure increases if I change it at 60k vs 48k? Taking only those vehicles that have already failed just has to biase the data. Without knowing how many engines don't fail (ie the true population size) then we cannot know this. I would expect the mean to plateau with increased sample size as most people seem to get their belts changed at 60k so there will be very few engines failing above that mileage whilst nearly all cars are in the lower mileage camp. Basically your sample size is decreasing markedly as age/mileage goes up. I'm not disputing there is an issue, rather pointing out we are missing information to make a true risk-assessment.
But that's the problem with statistics Keith, you're never going to get the full picture. We can only go with the data to hand. Those members reporting failures have agreed to have their details added, but this is only a very small number of failures occuring just in the UK alone. None of the dealers will provide information - presumably because of confidentiality issues but more likely there will be an inference of liability if they did. When have we seen dealers in general, and Alfa Romeo dealers in particular, admit liability for anything?
As you say, there is a problem, and the sticky thread only indicates the potential of the problem. But the real problem also lies with Alfa Romeo, in that:
1) They have not publicly admitted the problem.
2) The waterpump is still not on the schedule list for change at 72,000 miles (the cambelt needs to come off to change the waterpump).
3) The idle bearing isn't on the schedule for change at the 72,000 miles point either, but these have been known to fail, as noted by members on here.
The only way round it is to change the lot at an earlier interval, to reduce the possibilty of anything breaking or seizing. Quantifying it more precisely in the way you ask would be great, and from a statistical exercise, extremely interesting, but it ain't gonna happen.
Alfa Romeo are aware of the problem. The dealer technicians go on Alfa sponsored 'workshops' and the issue of this waterpump has been discussed. This has been forced on them because of the significant number of failures with more current cars such as the GT and 159 fitted with this waterpump. Unofficially, the dealers are encouraging people to change the waterpump when the cambelt is changed. There is also a cambelt kit for the JTD multi-valve engines, which includes the idle bearing. Dealers now tend to use this kit. The 'problem' is therefore 'unofficially acknowledged'. All you need to do as an owner of these engines is whether to change the running gear including the waterpump earlier or not.