Moved from D&G. This topic should get a better airing in Tech Log, perhaps ?
I've been involved in various aspects of MEL development and the problem is that the folks around the table arguing this way and that can only dream up so many situations of risk. We might opine that the MEL is reasonably sound and should get us along the way in MOST situations and circumstances. No one of rational thought would argue that every MEL is ironclad through and through. Engineering risk assessment doesn't go that far.
The PIC's role is to accept the initial work for what it is worth but then consider the "how about TODAY question". If he/she can't come up with a feasible problem which might cause heartache on today's flying, perhaps it is a reasonable call to say, "fine, we'll take it". On the other hand, it might just be that today's combination of night, IMC, weather, other snags on the aircraft, etc., etc., could cause him or her to suck in the breath and wonder "what about if we run into a problem along the lines of XYZ" ? Maybe we gracefully decline on this occasion ?
One problem which we have all seen is where the line pilot declines the MEL activation and then, Johnny Management comes to the rescue five minutes later with a hearty "I'll take it" .... no system is perfect, we can only try to do our best to achieve the gameplan goal.