In comment to the original post regarding the likelihood of system failure, it must be noted that any gadget ..whether mechanical or electrical/electronic ... can fail or function contrary to the design intent and do silly (and often tragic) things.
As a result the manufacturer has
(a) a set of design tasks to determine failure modes, their probability of occurence, likely outcomes, and to put in place defences against undesirable outcomes.
(b) a routine review group to assess in service problems which often leads to the various SBs which beset operators.
It must be kept in mind that all of this takes place against a backdrop of potential litigation - the manufacturer is not always in an easy situation.
Sometimes things slip through the net for whatever reasons. For instance,
(a) the Concorde accident may be in that category.
(b) As another, I was involved in the investigation of an unpleasant fatal several years ago where it appeared that a rather unfortunate STC design decision led to a quite unexpected system problem which, in turn, led to a very confusing set of operational symptoms - which the pilots, in turn, misinterpreted with dire consequences when their handling of the immediate problem was inappropriate.
(c) The Sioux City DC10 is another pertinent example.
(d) How many times, in the early FMS glass ship days, did we hear the catchcry "What is it doing now ?"
Aviation, like any other area of human endeavour, is imperfect. A well trained and intelligent pilot (CRM plays its role) is the current best last line of defence. The human can be innovative when non standard situations require new and imaginative strategies for successful resolution.
Doesn't always work, and we just have to accept that there are times when it would have been better not to have got out of bed that day.
From such cases the lawyers develop their investment portfolios.