Sorry guys, I am in a pedantic mood. If the failure of "one chamber" in a given period of time is a rare and random event, then the probability of a chamber-failure is most likely Poisson distributed.
This would mean that the probability of two failures in the same period is not quite two times (or, not quite 1/2 the probability). And, the probability of two failures in double the given period of time is also less likely than having the first failure.
So, I think that the dual-chamber master cylinder is more than twice as safe (in terms of total failure). And, you would, of course, replace the master cylinder very soon after a one-chamber failure.
I think this type of reply most likely belongs in the Greasy Spoon?
