One Timers

Using Advanced Stats to Classify NHL Center Types

When it comes to advanced stats and player evaluation, Domenic Galamini is a big deal.

The McMaster University student is the creator of the HERO chart, which pulls together several key analytics on NHL players and gives a succinct description of a player’s talent as a first-, second-, third-, fourth- or replacement-level player. These charts have taken over the blogosphere and likely don’t need much more explanation than this.

But, if HERO charts aren’t something you tend to look up, here’s a quick example.

Los Angeles Kings centre Jeff Carter‘s HERO rapidly displays some key indicators. Carter tends to play when his team is leading and he receives a second-liner’s amount of ice time (relative to other NHL forwards). His goal-scoring and puck possession (Corsi For per 60 minutes) are elite. His ability to suppress shots (Corsi Against per 60) is almost replacement level.

Thanks to Galamini and the charts, an analyst or fan can gather some keys fancy stats knowledge about any player. It takes seconds and the visual tends to resonate in a way that a spreadsheet with a billion numbers never could.

Building from this idea, I’ve been working to add a wrinkle that sheds a little extra light on player analysis.

Based on the HERO chart, Carter would likely be described as a rare example of a sniper who plays centre. His goal scoring is elite, his primary assists total is second-liner quality, so he’s a scorer. However, by graphing Carter’s production over the past three NHL seasons — and comparing his scoring rates with other qualifying centremen — a slightly different picture emerges.


Scorers, Playmakers, Bottom-Six, and All-Around – The Player Types

Below, I’ve created a graph drawing statistics from War-On-Ice.com here. Like the default HERO chart option, this data set includes player stats between 2012-2015. All 60 centres who have played 2500 mins at even strength during the past three years are included.

On the x-axis, a player’s assist totals per 60 minutes are plotted. By using a per/60 rate stat, players that benefit from greater playing time don’t necessarily dominate the visualization. For example, Jaden Schwartz appears in the “all-around” category despite lagging behind stars like Ryan Getzlaf and Matt Duchene who benefit from heavier use.

On the y-axis, a player’s goals per 60 are plotted. The same playing time equalization factors in here, allowing Cogliano to appear in the “goal-scorers” quadrant in the upper left portion of the grid.

The differences in TOI are indicated by colour. The grey circles represent average playing time among this group, red represents below average, blue represents above average.

But the most interesting feature is the creation of four quadrants by adding the intersecting average lines.



Bottom Six

Players that score and assist on goals at below-average rates are found in the bottom-left group. This group includes “bottom-six” skaters like Steve Ott, Sam Gagner and Nick Spaling.


Goal Scorers

Found at the top-left of the graph, players that score at above league-average rates but gather assists at a below-league-average rate, are considered goal-scorers. Predictable names like Steven Stamkos, Joe Pavelski and Ryan Johansen highlight this group.



In the bottom-right quadrant of the graph are players that assist on goals at above-average rates but score at a below-average rate. They are considered playmakers. This group includes celebrated setup artists like Henrik Sedin and Joe Thornton.


All Around

The league’s brightest all-around stars populate the top-right portion of the grid. All-world talents like Sidney Crosby and Ryan Getzlaf lead this pack of team-leading all-stars.


On Jeff Carter and Some Other Thoughts

Think back to Carter. He’s widely reputed as a natural goal scorer. And there’s no question that Carter is a talented sniper. He’s managed 283 goals in 718 NHL games, scoring on an impressive 11.3 percent of his shots during his career. He’s also scored more goals than assists, which lends to the notion that he’s a pure marksman.

But, when compared with other centremen who have skated more than 2500 minutes since 2012, Carter actually lands just inside the “all-around” quadrant. He’s managed enough assists per minutes played that he’s more appropriate counted as a contributor in goals and assists, not just a marksman. He is close to the line here. But the point is that Carter’s body of work, relative to his peers, actually shows a more complete game than he may be credited with otherwise.

There are some positive surprises for some players.

*old photo. Some storylines never change…

Despite a certain lack of fanfare, Matt Duchene plots near Jonathan Toews and Tyler Seguin, just back of Crosby and Getzlaf in the all-around quad. Though he isn’t often counted among the league’s best players, this chart reveals the all-around contributions Duchene makes on the ice. Toronto Maple Leafs centre Nazem Kadri finds himself near Jaden Schwartz, Anze Kopitar, and Eric Staal. This suggests that, if given a greater opportunity with the Leafs, the young centre may develop into a full-fledged NHL star.

In both cases, these young scorers are shown to be high-level all-around contributors.

On the opposite end of the spectrum is bad news for Paul Stastny and Travis Zajac. Though Zajac has been forced to carry the load as New Jersey’s first line centre, his scoring rates place him in the “bottom-six” quad; below-average rates in goals and assists. Stastny suffers a similar fate. Despite his role as the St. Louis Blues’ second-line centre last season, his production likely only warrants a role on a team’s third or fourth line.



What This is and What it isn’t

This graph can’t pretend to tell you everything you need to know about a player. Maybe Kadri’s off-ice issues will derail him from translating strong G60 and A60 rates into top-tier counting stats in the NHL.

And maybe there’s an unquantifiable reason why Zajac should skate on New Jersey’s top unit.

Even beyond player types, HERO charts, and analytics there are much broader issues to consider. Can a player be reduced to a single “type” at all? For more on that idea, please take a look at Stefan Wolejszo’s “Person, Player, Pineapple” for a treatment of the numerical commodification of hockey players.

But this chart does offer some sprinkles of insight. Comparing a player with averages among his peers allows for a fan to see what skills truly stand out. Relatively, Daniel Winnik is a decent playmaker and Andrew Cogliano does well as a goal-scorer in limited minutes. These insights offer some value as analysts move more and more towards well-rounded player evaluations.


What do you think? Which player types surprises you most? Does the graph pass the eye test for you?

To Top