Early on in this season, there were two things most often discussed about the Maple Leafs. Their power play was to hot not to cool down, and their five-on-five play was barely tepid. The penalty kill was largely ignored as something that was fine, and fun to watch because the Leafs won games no matter what.

The penalty kill gets more attention now that the recent success percentage hasn’t been that good, and the power play is getting suspiciously ineffective. Some bad goalie performances have proven the fulitily of looking only at the win/loss record, so a deeperlook is in order. How are the three parts of the Leafs holding up?

To have something to compare to, I’m just going to chop the 32 games played so far directly in half and look at the first 16 vs the second.

Special Teams

Success Percentage

Success percentage is the standard measure of power play and penalty kill value, and the most recent run of x goals in y tries is trotted out in every game as some kind of description of the special teams effectiveness. Much like plus/minus, the way it’s calculated is so bad, it’s not even useful as a descriptive of anything but perception of the special teams by people who watch TV. Each power play is held as a single unit regardless of length, and the quality of the play is reduced to a binary based on goal or no goal.

The Maple Leafs have 141.77 minutes on the power play so far, and 148.5 shorthanded, which is 15% of their total time on the ice this season. The smaller the stretch of time, the more variance decides the outcome, and in this case, that’s luck — good and bad — as well as the uneven representation of opposing goalies or skaters, injuries, and an unbalanced schedule along with a host of other things.

The worst part of success percentage as a measure of special teams, though, is that it folds the goalie performance into the mix in a way we’re all grown up enough to understand doesn’t work for five-on-five play. What good is something that can’t tell lucky from good or unlucky from bad?  To look at the Leafs special teams, I’m ignoring the goalie most of the time, so other measures are necessariy.

This is not only the simplest way to look at the special teams at the team level, it’s just funny. And it used to be worse, with the Leafs way off to the right so far, the rest of the chart was almost too cramped to read.

Power Play

The power play was absolutely too hot not to cool down, but it hasn’t cooled off that much.

Switching to Evolving Hockey’s Expected Goals (xG) model (which is broadly similar, but not identical to the above) for some ill-considered chopping up of numbers, I get this:

Maple Leafs first 16 games - 10.3 xGF/60 or first in the NHL by 1.84 xGF/60 over the Bruins — an absurd amount.

Maple Leafs second 16 games - 8.33 xGF/60 or third in the NHL by .92 behind the Oilers — a slightly more reasonable gap, but still large.

Overall, the Leafs have cooled back to very good, but not absolutely absurd territory, and they’ve sustained a level of play for their most recent 16 games that is excellent still. And this comes as the tinkering around players, positions and process of the power play itself goes on endlessly. Some of that tinkering was motivated by taking the strain off of Auston Matthews, so was necessary.

Shooting percentage, still mixing in the goalie, luck, random factors and player skill, but not as horrible as success percentage, shows the Leafs have produced goals at a high rate — they’re fifth with 17.59% over all 32 games, but not one that’s outlandish and must come down.

Penalty Kill

The vertical axis in the chart above is the penalty kill, and the Leafs are very good by that measure of performance. But, I hear you saying, the PK is terrible, everyone knows that.

The Maple Leafs outperform their Expected Goals on the power play, and we should expect that, particularly over a longer stretch of games with less influence of variance. They have very good shooters, not average ones, after all.

But do they have very good goalies while shorthanded to keep the penalty kill performing at or near Expected? So far the season, the answer to that is a resounding no. On 6.15 xGA/60, the Leafs have allowed 9.04 G/60. In raw numbers that’s 22 goals against on 15 expected (score and venue adjusted).

In the first 16 games, the Leafs were fifth in xGA/60 with 5.79 and allowed 12 goals to 9 expected. The save % was 80.6 or third worst.

In the second 16 games, the Leafs were 15th in xGA/60 with 6.69 and allowed 11 goals to 7 expected. The save % was 81.8 or sixth worst.

So, this is interesting, and neatly illustrates how the goaltending is so mixed in with everything else shorthanded, that it’s impossible to judge just from goals against, what’s going on. The other side of this coin, of course, is that the goaltending is not independent of the penalty kill behaviour of the skaters. And if you can decide who is “at fault” for the goals against being higher than it seems like it should be, well, good luck.

Frederik Andersen, in his recent four seasons on the Leafs, is approximately average shorthanded. It’s never been one of his strengths or a particular weakness even deeper into his younger years. The Leafs have had some 40 minutes of shorthanded play from the backups this season, where Campbell has been good and/or lucky and Hutchinson has been the opposite. They both have pretty wild swings over their careers, as is usual for low-minute backups, so it’s very hard to set a baseline with them.

The bottom line, is the penalty kill has been consistently a team weakness, and the only certain remedy for that is taking as few penalties as possible. It’s not at all clear, the skaters are the main problem, though.

Five-on-Five

Overall, the Maple Leafs are not as exciting offensively as they were last year when you look at their on-ice results on HockeyViz, but they more than make up for that in improvements in defensive results. This is good, and it shakes out to a team that is good, but not quite top of the league at five-on-five. There is also a disconnect between just shot rate — either Corsi or unblocked shots — and Expected Goals, which are of course just unblocked shots weighted for quality.

In the first 16 games, the Leafs had 52% Fenwick (unblocked shots) and 53% xG. The ranking goes from 10th to 7th. The disconnect comes into focus when you sort that out by for and against. The Leafs were 4th in xGF and 14th in xGA. Meanwhile the FF and FA have them 13th and 10th, so the xG weighting flipped their offence to good and their defence to middle of the pack.

This was very confusing, and was more marked in the first 10 games, and led to partisan support for the idea that offence is all about xG and defence about shot rate itself so it’s all perfect in Leafs Land. I think the logical explanation is that the Leafs have declined in shot rate, while taking fewer shots from the points, raising their overall quality for. They don’t actually execute defensively very well, lowering the overall quality against. This seems like a simple explanation that fits the team and doesn’t require switching horses mid-stream to rank the Leafs as good.

In the more recent 16 games, The Leafs have dropped to 51% Fenwick and in rank to 14th, and in xG% they are way up to 55% and are fourth. The spread got bigger, which only shows that 16 games is a small enough set for oddities.

Breaking it up for and against, they remained nearly unchanged in FF but the FA got worse and has them down at 16th in the league. xG brings them up to sixth offensively and seventh defensively. Suddenly you need to like xG and ignore the shotrate to make the Leafs look good, so this is a conundrum for those who went hard on the early theory of only shot rate matters defensively. The execution defensively has picked up, and that fits with the impression of the team. It might be driven by the unbalanced schedule as much as anything though, so all this chopping up is more about how we got here, than a prediction of where the road will take us next.

Where do we end up with it all stitched together into 32 games? Tenth in FF%, 6th in xGF%, driven by excellence offensively and a mediocre defence that never rises out of the middle of the pack.

Or the Maple Leafs, in other words. Better than last year in xG% by a considerable amount, with a lower pace of shots both for and against that helps considerably in raising the defensive half of the xG to the top  of the middle of the pack, not the bottom of it.

Now add the special teams, and the Leafs don’t quite need the power play to be on fire to win a game, but it sure helps. The tinkering needs to stop by the playoffs, however.

I can’t help feeling like the lower pace of shots for, which many people do believe is necessary to have lower shots against, is settling for a worse offence than is possible just out of conservatism. I think the intent is actually to limit rush chances against, not overall shots, and given the goaltending, that makes some sense, but it still nags at me as an area where the Leafs skillset should eliminate some margin for error in net, and yet they are choosing not to.

When the power play works, which you can’t expect it to every game, nothing else matters, and maybe that makes it harder to see or to generate the kind of five-on-five play the team needs.

There is no reason to expect the final 26 games to be dramatically different to the first part of the season, and the overall effectiveness of the team is more than enough to dominate the North Division, as we’ve seen.

The Leafs are going. The only question is how far.

Analysis