WRT frames, we should probably make a distinction between the (ultimate) probability of failure (which will always be 1) and the hazard rate (the instantaneous probability of failure).

Any frame will break with certainty if ridden for eternity. (somewhat like death, or taxes...) But no one cares about this. What we care about is the hazard rate, i.e. the probability of failure at a point in time. Note that in the toss of a coin example, events are presumed to be independent, meaning that every time you toss the coin, the probability of head = the probability of tail = 0.5. Not so with frames because of material fatigue, such that we expect that the hazard rate increases over time, and this rate depends on very many other circumstances, both associated with the frame itself (material, geometry and such) and the environment (load, type of road, weather, etc.). The problem is that because failures are rare, and because of a self-selection bias, it can be very difficult to come up with answers. [self-selection bias : most tourers planning for a difficult/remote trip will ride on a steel frame rather than, say, carbon or bamboo. If the type of riding has an impact on the hazard rate (which is very likely), we shouldn't be surprised to find more anecdotes of frame failure involving a steel frame.]

Hmmmm... many are probably familiar with the UCI rule stipulating a minimum weight for bikes (6.8kg) used in competition. Interestingly, as technology has improved, several production bikes ridden by Sunday amateurs weigh less than the regulation weight... Touring being are marginal sport, we do not have such guidelines. Yet it would be an interesting development.