The numbers behind the NHL’s Top 20 Centres

A noted trademark of August Hockey Social Media is the absolute grief that comes out of the NHL Network’s annual positional rankings. Designed almost entirely to generate a lot of discussions and quote tweets off of a series of short videos, the TV station puts out Top 20’s for centres, wingers, defencemen, and goaltenders that usually leave people frothing at the mouth.

Usually, it’s because their team’s star player in that position is too low for their liking. Or, it’s a player from a rival or a team they feel gets too much attention being ranked too high. Whatever the case, having 20 different spots to dissect and a pool of snubs leads to some great bloodbaths. Below is this year’s ranking in the centre position:

Curious to whether there’d be anything definitive hiding in these players’ numbers that would make certain player’s rankings look incorrect beyond subjectivity, I ripped open some visual tools. Specifically, I used Bill Comeau‘s SKATR charts (which use Corsica’s data) with a two-year outlook and CJ Turtoro‘s All Three Zone comparison tool, which uses data from Corey Sznajder‘s All Three Zone’s project.

I used these visuals’ data because they output easy-to-consume percentile ratings for each number. After all, we’re not trying to build our own Top 20 here, simply trying to see if anything is truly a rage-inducing degree of out-of-place.

Individual Numbers

(GS = Game Score, PTS = Points, G = Goals, A1 = Primary Assists, A2 = Secondary Assists, iCF = Individual Shot Attempts, iXG = Individual Expected Goals, SH% = Shooting Percentage, Pdif = Penalty Differential. All metrics besides SH% are rated per-hour at 5-on-5)

I don’t think it comes to anyone’s surprise that most of these players are in an elite offensive class. Some rack up numbers through playmaking, some through finishing, and some through both, but almost all of these centres are in the Top 10% of the league in even-strength point production rates over the past two years. Three players are outside the Top 20%, but even they come with caveats (Seguin being an elite powerplay finisher, Bergeron being used largely for his ability to push his linemates forward and shut down opponents, Kopitar being similar to go with a down year in 2016/17 that lowered the average).

One thing I will point out: I’m not an “Auston Matthews at fourth is correct” truther, but I don’t think it’s as outlandish as it was made out to be. Outside of secondary assists, he is by far the closest to keeping up with Connor McDavid in terms of individual contributions. The fact that two young talents playing their first two full seasons are running the show here is pretty insane.

Shot-Based Metrics

(CF% = Corsi (Shot Attempts)-For Percentage. Crel = CF% Relative To Teammates. CF60 = Rate of Shot Attempts players’ team has taken with him on the ice. CA60 = Rate of Shot Attempts player’s team has conceded with him on the ice. xG = Expected Goals, final four columns follow the same pattern as Corsi columns)

Also known as the “Patrice Bergeron is not human” section, this is where the defensive wizards and cycle-driving robots make their bread and butter. A few things worth noticing: being a superstar isn’t a guarantee of good underlying numbers (in large part due to the fact that you’re tasked with facing other top players), and for the most part, this list is straight up below the league average in defensive rates (likely because high-skill players play with more pace, both because they like their odds of finishing from a risk-reward standpoint, and because they’re more confident with the puck to not waste much shift time on dumping and chasing).

Team-relative percentages are usually a better indicator of control than purely looking at concession, and even then, who they face matters (I’m really interested to see how Matt Barzal fares next year, for example, and you’ll see why below).


(TOI% = Percentage of team ice time played. QoC = Quality of Competition. QoT = Quality of Teammates. DZS% = Percentage of shifts started with a Defensive Zone Faceoff)

I don’t think anyone is too shocked by anything they see here. Your best players play a ton unless you have the depth to support them, and usually, teams will send out their best players to take them on. Quality of Teammates with star players is something I wonder about – time on ice is the preferred proxy for judging both Competition and Teammates, and in both cases it typically smooths out, but if you get a high-minute player on your wing for long enough under the guise of chemistry/balance, does the number get skewed a little much?

I guess what I’m saying here is that I’m not convinced that Connor McDavid has Top 15% support on his wings when he has to drag around Milan Lucic and Drake Cagguila for extended periods of time. As referenced before, I also wonder about Barzal’s, as he only has this year to work with and the Islanders spent a good chunk of time figuring out which wingers would work with him, which would work with Tavares (who sheltered some of his competition time), and who would get sent to purgatory on their sparingly-used bottom six.

William Karlsson’s numbers, are of course, what you’d expect for a two-year outlook on a player who went from a bottom sixer to a superstar (allegedly) in the blink of an expansion draft pick.

As for the D-Zone starts, it looks a lot like the against stats in the previous block. Most of everyone is bad. Barkov and Couturier are good. most teams want their best players trying to get them goals rather than blocking shots off the draw, so this isn’t horrifically shocking.

Transition and Distribution

(SC60 = Shot Contributions [Shots+Shot Assists] Per Hour. SA60 = Shot Assists [Passes that lead to shot attempts] Per Hour. Ent60 = Controlled Offensive Zone Entries Per Hour. Ent% = Percentage of Zone Entry Attempts Succeeded. Exi60 = Controlled Defensive Zone Exits Per Hour. Exi% = Percentage of Zone Exit Attempts Succeeded)

Once again, it should surprise no one that just about everyone here is extremely good at most of this stuff. Nearly the entire list distributes the puck at an elite level, and most carry the puck into the offensive zone often and with success. A few players stick out as outliers, but play on lines where another elite linemate (or two) is responsible for the breakout, making their attempts less frequent and often more of a “Plan B”.

Zone exits come with a lot of dominance across the board as well, though more names are infrequent. I’d imagine this is a similar situation: their winger or defencemen are usually the ones skating it out, unless its a moment of desperation. I wouldn’t be surprised if in the cases of guys like Bergeron and Seguin, that their lack of success is systematic more than it is inability.

Final Impressions

As I’m sure you’ve gathered, nearly everyone on this list is elite at a lot of things (Karlsson being the exception, which can both be explained by his Columbus year ‘anchoring’ him downward, while those expecting regression might have him somewhere near this middle ground anyway). That’s because most of these guys are, in fact, in the top 15 or so percent of centres in the National Hockey League, and they all tend to get there in different ways.

We all have our personal slides up and down that we’d make with the list, but assuming we have one and two more or less set in stone (McDavid at 1, Crosby at 2), the rest of the pack is more fluid and “what have you done for me lately” driven than most will admit. So many of these players have been argued to be Top 20, Top 10, even Top 5 at least some point in the last 3 or 4 years, that I have a hard time giving anyone too much heat for going with personal preference in the 3-20 pocket. There are so many arguments that can be made for so many different combinations that it someone would have to be purposefully difficult to make an arrangement that is truly, undeniably off-base.

You might be interested in …


  1. Matthews when up against some of the top names in this list was schooled more then a few times this past season. I expect a huge jump this season with the addition of reflection and JT.

Leave a Reply

Your email address will not be published. Required fields are marked *