The term "shut-down" line is, of course, a generalization. We use it to refer to players who generally score less than their assigned check or matched line opponent, but the truth is, no two checking lines are the same, and all of them will play on teams with different strengths and weaknesses, so it can be difficult to make broad statements about whether or not it's a good idea to employ such line combinations and when they should be used.
We can all recall a few examples that worked well, such as Jordan Staal matching up against some of the Red Wings' finest in successive playoff runs, or Sammy Pahlsson matching up against pretty much anyone in the Ducks' Cup run. The inevitable counter-argument of the anti-shut-down line crowd to these examples is that (at least during those Cup runs) both Staal and Pahlsson didn't just shut down - they scored. Yes, this is the ideal, but Jordan Staals are few and far between, not to mention rather expensive to acquire and keep. In other words, what if Jay McClement is your best hope at limiting chances against?
A look at the current debate
There seem to be two schools of thought emerging on this issue, and like so many other hockey-related (and more specifically Leafs-related) topics, there is a tendency for the debate to polarize. I'd like to suggest a different approach, but for now let's have a look at the two arguments:
A better ratio is better: Cam Charron recently wrote about Team Canada's World Junior Hockey Championship team, and concluded that Canada spent too much time worrying about shutting down the opposition, and not enough time looking at the overall ratio of chances for and against. In other words, limiting chances is great, but it's better just to have more than your opponent. Focusing exclusively on the defensive side of the game is simply playing not to lose, and if you are willing to play large sections of the game content to sit back and not score, the odds are pretty good that you won't.
But wait: Our very dear commenter not norm ullman wrote an interesting piece a few days ago about this very point, and, if I may be so bold as to distill his idea (well, at least one of them), basically he said something like this: NHL coaches are, in the third period, correct in ignoring the overall ratio of shots (or shot attemps a.k.a. "events") for and against a given player, and choosing the low-event player instead. Your team doesn't get any extra points for winning 3-1 or 2-1 - you just have to prevent the goal that could tie the game. Because of the NHL's point system, it is beneficial to simply avoid a tie/loss, and the lowest-event player is the best one for that job.
The middle way: Overall, players with good Corsi/Fenwick ratios are going to be the players you want more of on your team. Ideally, even your shutdown players generate shots... Unfortunately, this isn't the case for the Leafs. The problem, of course, is one of timing: at what point is it OK to 'sit back' or simply focus on not being scored on? Jumping out to an early lead is great, but it doesn't do much good to then sit back and wait for the other team to start making mistakes, because if the play is usually in your end, odds are better that the really costly mistakes are going to wind up in the back of your own net. So. Shutdown players? Sure, maybe, but you have to be in good shape in terms of both the score and the clock in order for them to be of any use.
One last thing: it should be noted that Team Canada shouldn't have been playing for "2 points". If the score is tied or you're trailing (as Canada was, for most of their semifinal loss against Finland), why bother trying to 'shut the game down'?
At what point is playing McClement a good idea?
Well, it's certainly not when the team is losing, because all the shot suppression in the world doesn't help a team score goals. Get him right the heck off the ice. Behind the Net tells me that 60 minutes of McClement would result in just over 16 shots for. Nope. Even in a tied-game scenario, he's probably not the guy you want. Who plays for a tie? (Answer: randycarlyle.jpeg)
Now, looking at Behind the Net again, we can see that McClement is one of the Leafs' best players at suppressing shots at 5-on-5. Moreover, he does it with a 31% offensive zone start rate, and faces some tough competition to boot. Heck, allowing only 28.4 shots per 60 minutes at 5v5 is fantastic, when you consider (thanks to Extra Skater) that the Leafs' 5v5 Shots Against/60 number is the worst in the league at 35.0.
If we assume for the sake of simplicity that the Leafs are facing an average opponent, and we know that the league's current 5v5 Sh% is at an average of 7.77%, a bit of quick math tells us that playing Jay McClement for a full game leads to 2.21 goals against, and thus, per period is 0.74. In other words, if the Leafs play McClement for exactly the amount of time they played him last night (OK, that's not all at 5v5, but still), odds are, they're going to get scored on.
A more reasonable amount of ice time for McClement might be in the 10-minute region. I see no reason to play Peter Holland or Nazem Kadri less than McClement unless the team is in full-on shell mode, and the Leafs really shouldn't be protecting a lead that way for 20 minutes. With McClement on the ice for 10 minutes (OK, again, this is just the 5v5 numbers), the rate of goals looks reasonable, because 0.37 goals per game is well... usually not a goal.
In summary, to answer the titular question, the Leafs should be using McClement late in the game to protect leads, but not as much as they have been. If McClement is on the ice for twenty minutes, odds are that the opposing team is going to tie it, and the Leafs would have been better off giving extra time to scorers who stand a chance of getting another goal, because the Leafs will need all the offence they can get if they're going to get scored on anyway. Again, the numbers seem to line up with what we see (that is, the Leafs 'shutting down' or 'sitting back' way too early in a game), and I guess it makes Jay McClement's usefulness seem pretty limited.
Addendum: You might think that McClement would be among the Leafs' best shot suppressors on the penalty kill as well, but that's not really the case. This opens an argument about the style of play that a player has been coached to perform, and whether the likes of Phil Kessel would make a great penalty killer if he set his mind to it, but then, James van Riemsdyk also doesn't look so hot on the PK... This reminds me of a couple other posts I did (here and here) on the idea of PK specialists and whether or not they really exist that might prove useful to discussion.