Canucks: Elias Pettersson’s performance in light of a new advanced stats metric

facebooktwitterreddit

Vancouver Canucks Elias Pettersson ranks fourth among the NHL’s top young forwards when factor of “defensive strength faced” is taken into consideration

Among the NHL’s young star forwards, Vancouver Canucks centre Elias Pettersson ranks right near the top in a comparison of these players’ first two seasons in the league.

More specifically, he is second in terms of “points per game,” and also second in terms of “points per 60 minutes” (“pts/60”). However, both of these metrics fail to consider an important factor in the assessment of offensive statistics, and that is the collective defensive strength each player faces over the course of a season. And this article will work to develop a new metric that takes this factor into consideration.

We start by noting the large spread in team “goals against averages” (GAA) every year. For example, in the 2016-17 season, the Washington Capitals lead the NHL with a 2.22 GAA, while the Colorado Avalanche had a league-worst 3.39.

Consider, for a moment, what this means for a comparative analysis of our 17 young star forwards. Because of the NHL’s unbalanced schedule, some players got to feast on Colorado’s porous defence five times, while having to face the stone-walling Caps only twice, and other players only got to play Colorado two times and had to play Washington four times.

Clearly, the unbalanced schedule has made for an unlevel playing field in terms of the defensive strength faced by the 17 players under examination.

A look at the collective GAAs of the Western and Eastern Conferences over the past five years reveals another factor contributing to an unlevel playing field:

15-16          16-17          17-18          18-19          19-20
2.73              2.79              2.85              2.96             2.91       WEST
2.70              2.75              3.09              3.06             2.99       EAST

Note how the GAAs in the two columns on the left side of this chart are distinctly lower than the GAAs in the two columns on the right side. This indicates that those among the 17 young stars who entered the league during the past two or three years—and especially those in the Eastern Conference—have had the advantage of facing weaker defences than did those who entered the league five years ago.

More from The Canuck Way

The rest of this article will be an experiment in “leveling the playing field,” that is, introducing a new metric designed to neutralize these advantages that some players have over others, in order to put all 17 of them on an even footing for the purposes of comparing their offensive outputs.

First, we need to establish some way of measuring in numeric form the defensive strength the players faced in each of their first two seasons.

Fortunately, we have a measure of defensive strength in numeric form readily available to us, and that is the seasonal GAA of each of the NHL teams.

To illustrate how this data can be used to calculate the defensive strength a given player faced in a given year, we will use Auston Matthews’ rookie year (2016-17). He played all 82 games that season, so we need to collect the seasonal GAA of his opponents in each of those games. Imagine you have a chart with 82 slots. Matthews’ first game—featuring his unforgettable four-goal performance—was against the Ottawa Senators, whose GAA for that season was 2.61, so you put that into slot 1.

The next game was against the Boston Bruins; their seasonal GAA was 2.59, so a “2.59” goes into slot two. And you continue this process all the way to game 82, which was against the Columbus Blue Jackets, whose GAA for the year was 2.38, and so that goes into slot 82.

Next, you take all the figures you’ve collected in the 82 slots, add them together, and then divide the resulting sum by 82 in order to get the average GAA of all those games, that is, the average defensive strength Matthews faced, which turns out to be 2.757.

Calculating the defensive strength faced by Pettersson in his rookie year is a little more complex, for it involves an extra step. As we’ve seen, Matthews played all 82 of the Leafs’ games during his rookie year, and so, calculating the defensive strength he faced involves simply consulting the Leafs’ schedule for a list of all of his opponents, and how many times he played each.

Pettersson, on the other hand, missed 11 games his rookie season, and so, before we can begin calculating the defensive strength he faced, we need to determine the exact games he missed, and eliminate the opponents in each from the list.

This turns out not to be an arduous task, for just a glance through his game log for the season quickly reveals no games in a gap between Oct. 15 and 27, and none in another gap between Jan. 3 and 20.

Therefore, one need only take note of the six opponents faced during the first gap— the Pittsburgh Penguins (2.94), Winnipeg Jets(2.98), Bruins (2.62), Capitals (3.04), Vegas Golden Knights (2.80), Arizona Coyotes (2.72)—and the five faced during the second—Toronto (3.06) / Arizona (2.72),  the Florida Panthers (3.41), Edmonton Oilers (3.34) and Buffalo Sabres (3.30)—and eliminate their values from the tabulation. Therefore, what would have been an 82-game total of 258.10 ends up being a 71-game total of 212.64, with an average per game of 2.995.

Going through these calculations for each of our 17 players, we get the following chart, with the players listed in alphabetical order (note: the figures in parentheses are for first seasons consisting of 25 games or less).

| 2015-16     2016-17     2017-18     2018-19     2019-20
|                          2.755           2.920                                                    Sebastian Aho
|                         (2.330)          2.972          3.054                              Matthew Barzal
|                         (2.650)          2.926          3.015                              Brock Boeser
|                         (2.828)          2.858          2.991                              Kyle Connor
|                                                 2.925          2.960                             Alex DeBrincat
|    2.714           2.762                                                                         Jack Eichel
|    2.685           2.773                                                                         Nikolaj Ehlers
|                          2.815            2.950                                                  Patrik Laine
|    2.734           2.769                                                                         Connor McDavid
|                          2.758            3.009                                                  Mitch Marner
|                          2.757            3.009                                                  Auston Matthews
|   (2.640)         2.764            3.009                                                  William Nylander
|                                                                      2.995          3.025        Elias Pettersson
|                          2.763           3.006                                                   Brayden Point
|   (2.640)         2.831           2.937                                                   Mikko Rantanen
|                                                                      3.037          2.995       Andrei Svechnikov
|                         2.771           2.957                                                   Matthew Tkachuk

One thing that stands out in the chart is the figure “3.009” three lines in a row in the middle column. Since those three lines are all occupied by Leafs, it should come as no surprise that they might all have the same “defensive strength faced” values; after all, each of their schedules would have them facing the exact same opponents the exact same times.

However, looking at these three players’ values for the preceding season, we find three different numbers. This also is easily explained. While these players all had the exact same schedule, they all did not play the exact same games. Matthews managed to go the whole season without missing a game, but Nylander missed one, and Marner missed five. This means the pool of opponents faced was different for each player, and so, it’s not surprising that their “defensive strength faced” values would differ from each other.

It might appear that all the values on the chart cover such a narrow range that there would be no significant difference between them. However, if you take the highest value—the “3.054” for Barzal—and compare it to the lowest value—the “2.685” for Ehlers, the difference indicates Barzal had a 12.1 percent advantage over Ehlers, just by virtue of facing weaker defences. So, if we wanted to compare the actual offensive abilities of these two players, it would seem only logical that we would need to adjust Barzal’s numbers downward to compensate for the 12.1 percent “defensive strength faced” advantage he had over Ehlers.

Related Story. Canucks could have plenty of representatives at 2022 Olympics. light

In fact, for the task of comparing the offensive abilities of all our 17 players, we should be doing the same with all the numbers on the chart. Ehlers’ “2.685” represents the toughest defensive strength faced, with all the higher values representing weaker defences faced, and thus an advantage over Ehlers in terms of ability to produce points.

We need to calculate the percentage advantage of each of those higher values, and then adjust each of the players’ “pts/60” ratings downward in the amount of the corresponding “defensive strength faced” percentage advantage, thus showing the “pts/60” each player would have produced if facing the exact same level of defensive strength.

This is all pretty complex, so I will provide an example of working through these calculations, again using Pettersson’s rookie season for the purposes of illustration. Earlier, we determined that the “defensive strength faced” value for Petey’s first season was 2.995. This indicates that the defences he faced were significantly weaker than those faced by Ehlers. And to calculate exactly how much weaker, we first determine the percentage difference between Petey’s “2.995” and Ehler’s “2.685”—which turns out to be 10.4%—and then we adjust downward Petey’s “pts/60” (3.059) by that amount, yielding an adjusted “pts/60” of 2.741.

Working through this set of calculations for all the players, and then averaging out each player’s “pts/60” over his two or three seasons, we get the following chart (with each player’s ranking in the former “pts/60” chart in parentheses):

1)   3.351     Connor McDavid (1)
2)   2.897     Auston Matthews (3)
3)   2.829     Patrik Laine (4)
4)   2.793     Elias Pettersson (2)
5)   2.769     Mitch Marner (6)
6)   2.707     Matthew Barzal (5)
7)   2.648     Alex DeBrincat (7)
8)   2.530     Alexander Nylander (9)
9)   2.500     Brock Boeser (8)
10) 2.406     Jack Eichel (15)
11) 2.398     Matthew Tkachuk (10)
12) 2.336     Sebastian Aho (13)
13) 2.292     Mikko Rantanen (14)
14) 2.286     Kyle Connor (12)
15) 2.267     Andrei Svechnikov (11)
16) 2.262     Nikolaj Ehlers (16)
17) 2.131     Brayden Point (17)

We see that Pettersson has dropped from second to fourth, which should come as no surprise. We saw in our first chart (above) that league-wide defensive strength was the strongest at the beginning of the five-year period under examination, but became significantly weaker by the end.

And that was right when Pettersson was breaking into the league, meaning his “defensive strength faced” values would have been among the highest on the chart, translating into significant downward adjustments in his “pts/60” ratings. On the other hand, the two players who moved past him—Matthews and Laine—started their careers in the 2016-17 season, before the “defensive strength faced” values ballooned up, meaning their downward adjustments in “pts/60” ratings were much smaller in comparison.

There, then, is a demonstration of a new advanced statistics metric, one designed to refine the findings of “pts/60” analyses.

dark. Next. Canucks roundtable: Who should Jim Benning extend?

While the “pts/60” metric does represent a vast improvement over simple “points-per-game” findings, it does not account for the fact that some players hold an advantage over others simply because the opponents they face over the course of a season are defensively weaker overall as compared to the opponents other players face. And it is my hope that the new metric introduced here may help to rectify this imbalance.