I think when you approach this question you have to first look at the strategy for building a team. You have success in this league mostly because of strategy, not because of luck. The best built teams are the most successful, most of the time. Sometimes luck plays a part, but the first measure of a GM has to be their strategy.
When you look at strategy, two GM's who grade out really poorly are Lou and Bergevin. I know their teams have had success this season (and also last season for the Islanders), so it'll seem impossible to some that they could be bad GM's, but those teams success wasn't because they had good strategy. It was for other reasons. Lou and Bergevin have a common problem of they don't understand what wins in the current NHL. They don't know how to build a current cup winner.
Almost all the recent teams that won the Cup had the same formula. It requires a bunch of star forwards, usually including a 1C, and at least a star defensemen or two, usually including a 1D. It might include a goalie, often times its a top heavy roster. These two guys think that you build a Cup winner by accumulating middle of their career and middle of the league NHL veterans on big contracts. That is the absolute wrong strategy. Star power is what wins in this league, not depth, unless its Tampa's type of depth in great players. These two don't understand, which is why I think they are fundamentally bad GM's. It might yield the occasional good season or playoff success, but the team that wins the Stanley Cup is usually the best team in the league. You probably aren't going to luck your way into that achievement with a poorly built team.
People can criticize Dubas for a lack of results, and I don't think a GM is good just because they build a team properly, so you can criticize the guy, but that Toronto team is absolutely structured correctly with the core of the team. The answer in Toronto is not to try to be like Montreal because they lost to Montreal. You can criticize Benning for doing a lot of crazy stuff, but he has the right idea with building around players such as Pettersson, Horvat, Hughes. Those are the types of core players that lead to a Cup winner, if you have enough of them. There are more bad GM's in the league than good GM's. I think people are often too kind in assessing GM's, but you have to start with the poorly structured teams as the worst GM's. I don't know how you overlook the first step in winning a Stanley Cup.