Examining the circumstances of Toronto’s third period collapses

Photo Credit: Dan Hamilton/USA TODAY SPORTS

You know the story. The Leafs score a bunch of goals, pull ahead in the third period, and then everybody holds their breath. They’re going to cough it up again, of course. In 37 games this year, Toronto has led in the third period 23 times and won 15 of them, losing one in regulation and seven in overtime or the shootout. That’s still 80.4% of possible points, but if they had the nine that they gave up, they would be just a single point back of Montreal for the Atlantic Division lead and right near the top of the standings, likely tied or a point ahead of Washington for sixth overall.

So that really stings, especially when sitting two points back of a playoff spot. It stings even more of late, too, seeing as Toronto has pulled this act in 3 of their last 4 games; winning against Florida after conceding two early in the third period and narrowly winning the Centennial Classic after falling from 4-1, but losing to Washington last night made the issue glaringly obvious once again.

Why does this keep happening? Here are a few thoughts.

I’ve been beating this drum on Twitter for the past couple of weeks, but that’s because its a drum worth beating. Simply put… the Leafs give up a lot of leads because they spend a lot of time in the lead. This year, Toronto has been ahead of the opposition for 940 minutes; nearly half (42%) of their time spent actually playing. Only the Columbus Blue Jackets, a team that is on a historical 16-game winning streak, are ahead of them in this regard, and at their current pace, Toronto will have eclipsed their time with the lead total from last season by the end of this weekend.

A huge reason as to why nobody remembers the Leafs collapsing this frequently previous years was they were usually tied or trailing. Toronto has never spent more than 1498 minutes ahead in the fancy stat era (since 2007/08); they’re on pace for 2084 minutes, sixth highest of the decade, this season.

You’ve probably heard about score effects so many times that you’re sick of it. For those who haven’t, it’s the idea that teams that are ahead will pull back to protect their lead while teams that are trailing will get a bit more aggressive and take more risks, valiantly attempting to come back.

Ten years of data pretty clearly backs this up. Teams have historically attempted 44.2% of the game’s shots while leading, 45% of the unblocked shots, 45.1% of the shots on goal, and Corsica’s expected goals model assigns a typical weight of about 45.6% of goal differential to those efforts. Interestingly, leading teams exceed that, landing right around the 50% that they would in any score state, which over 426,000 minutes of sample, might just mean that the better team is a bit more likely to keep chipping away.

Trailing teams also see an uptick in the share of attempts faced that they block, perhaps because they tend to sit back and play passively instead of chasing for another goal. Even still, shots against tend to go up, so playing for blocked shots isn’t incredibly opportune.

While the goals have balanced out in the long run for these teams, teams that do a lot of winning should theoretically be capable of having a better goal gap than even. Not to mention, playing at the mercy of getting outshot makes smaller samples more likely to swing in a negative way, which might be part of the issue here. Toronto has walked away with points in all but one of these collapses; there could simply be luck of the draw.

As much as most would love to make this another “lets completely dump on the depth players” segment, it’s a little bit more complicated than that.

For example, while the pairing of Matt Hunwick and Roman Polak, unsurprisingly, gives up the most shot attempts, unblocked shot attempts, shots on goal, expected goals against, and scoring chances against, they really haven’t given up too many goals; their pairing has actually been the most effective. Now, as you all know, I’m a huge skeptic of using goals to predict the future (most people are), but when we’re talking about “who has given up the lead”, that’s very much results based and the results say that part isn’t on them.

Where they might be influencing goals against, though, is the fact that the team takes a more negative penalty differential when they’re on the ice, likely due to a lot of puck chasing. Jake Gardiner is in a similar boat. While the goals might not come on their even strength shift, they put the Leafs in an even tougher spot to protect their lead when making them go down a man. Ben Smith is in a similar boat to that pair; bad at just about everything, the worst at having his line take penalties, contributing almost nothing offensively and not even helping much at the faceoff dot, but not giving up goals.

Hunwick himself has also done a better job than most at directing attempts away; while he gives up a lot of them, possibly too many for this to matter, only 45% of attempted shots turn into a shot on goal when he’s on the ice while the Leafs are up. Meanwhile, most of other defencemen over at 47-50%. I don’t know how much that really means, but it’s a tangible example of him keeping shots to the outside, if nothing else.

Moreover, though, these guys can’t really be blamed because they aren’t the ones who are playing the most minutes when leading. Granted, the Leafs also take leads quite early in games and I don’t have the capability to break this data down by period; perhaps looking at just the third period would paint a different picture, but it appears that Zaitsev/Rielly get the most significant leading minutes and that the Kadri and Matthews lines are the most relied upon while ahead. In terms of limiting shots and getting some back, those forward choices appear to make sense of the long run. The Gardiner/Carrick pair does a little better than Rielly/Zaitsev in this regard, but being less gassed as the game elapses probably helps.

The Leafs seem to have its stretches where these goals flood in most consistently; in games where Toronto has had a third-period lead and either gone to overtime (nine games) or lost in regulation (one game), they’ve given up half their goals in the first two or last 90 seconds of the third period.

Now, both of these timeslots do make some sense to be the ones that stick out. Early goals could come as a result of a shift in strategy, and last-second goals could simply be a matter of the other team throwing it all on the line. But that also goes the other way; Toronto seems to get caught standing shortly after carry-ins on most of the early goals, and in almost all of the late ones, they’re trapped and chasing.

There are a couple of schools of thought here. Maybe the coaching staff isn’t adjusting to other teams quick enough to start the third period, and maybe their late-game strategy isn’t up to snuff.

Or honestly, it might just be in the players’ heads right now. They’re coming out of the gate thinking the same thing we are: “oh crap, here we go again”. Most of these goals follow similar patterns; rushes created by bad passes or absent minded skating, or in tight goals created by failed clear-outs and multiple skaters heading towards the same opposing player, leaving open men in dangerous areas.

Mental mistakes, created by worrying too much about the situation.

So what we know is this; the Leafs are a very good team that are leading games longer than almost everybody in the National Hockey League. Once they’re ahead, they’ve mostly put out their top players, and they’ve done a middle-of-the-pack job statistically at controlling the flow of play. But, for whatever reason, the team seems to stumble out of the gate and in the closing moments without fail, making mistakes that scream pressure, be it a giveaway, bad dump, or a penalty that makes the situation worse.

Somebody who is looking for a rapid-fire solution to cut goals against off would likely deeply their depth players, but while the group has done a surprisingly good job at avoiding goal concession, history implies that relying too much on them would be playing with fire. 

I don’t know if there’s a direct solution right now. Maybe you keep forcing the team to press and attempt to widen the cap in the third period, instead of sitting back, even in the final minutes. Maybe the staff just has to do their best to keep the guys comfortable with the idea of hey, things happen, but just because disaster before doesn’t mean it has to happen again. Showing confidence that things will shake out rather than messing with players’ heads further by experimenting solely based in scattered mental mistakes seems like the sensible thing to do; that’s what Mike Babcock did with Frederik Andersen early in the season and it paid off extremely well.

The good news in all of this, though, is that issues like this are something that through a combination of confidence, strategy, and reversal of some bad luck, can be mended. It’s better to have led and lost than to have never led at all, and while its frustrating to see the odd fumbles right now, it’s a much better situation than the team has been in for a long time.