On Monday morning the NHL debuted a new public resource for advanced analytics. Called Edge, the new site uses player- and puck-tracking data to deliver new metrics for player performance. In total candor, there isn’t much information on the new site that wasn’t already (and better) delivered through sites like HockeyViz. But there is some new stuff on Edge, and it does offer some new insights, and one of those insights may shock you:
The Washington Capitals are slow.
What’s New and What’s Not
Most of the metrics on Edge have been available before. Shot-location data, for example, has been around since the 2007-08 season, and they still provide the foundation of our best resources. Edge offers some new options using this data, but it doesn’t quite have the sophistication of a site like HockeyViz. Below, at left is Edge’s visualization of Alex Ovechkin’s shot locations last season; at right is HockeyViz’s.
I prefer HockeyViz’s visualization, which has more detail — both in location and shot type, plus I find it easier to scan.
But to its credit, Edge has resurrected an important statistic that has not been available for the last twenty years: zone time. During even-strength play, the Capitals spend 37.5 percent of their time in the offensive zone, 17.8 percent in neutral, and 44.7 percent in the defensive zone — ranking them somewhere in the bottom half of the league.
(I do not know what it means for a team to be worse than another team in neutral-zone time.)
When the NHL implemented its Real Time Scoring System (RTSS) regime about twenty years ago, zone-time data stopped being provided. The early shot-attempt statistics were used as a proxy to replace that lost zone-time data, and they’re still in use today. According to Natural Stat Trick, the Capitals control 40.5 percent of the shot attempts when they’re on the ice, meaning — through some abstraction — that the Caps are in the attacking zone substantially less often than their opponents.
Skating Speed
To me, the novel contribution of Edge so far is its skating data. In addition to raw distance, Edge tracks top speed, which they define as the “maximum sustained skating speed a player has achieved,” and bursts, which is a count of “times a skater achieved a sustained speed above a given threshold.” I have a lot of questions about these measurements and their usefulness, but for now it’s at least interesting — if incomplete. We don’t know how long a player has to sustain their top speed or how top speed relates to a player’s overall speed. It’s also unclear what the implication is when a player has a lot of those 22 miles-per-hour bursts. Could a lot of bursts suggest a player is chasing opponent breakaways too often?
There’s a lot of still-to-come wisdom we’re missing in these numbers, but it’s a cool start, and it prompts some fun conversations. At great personal cost (the site is slow and unstable), I collected the top skating speeds for every Caps player. This number is the peak miles-per-hour speed each player achieved over the first four games of the season. Also included is that speed’s difference from the league average for their position, plus the player’s percentile. I’ve ordered the players by their difference from league average for a kind of position-adjusted stack ranking.
| Player | MPH | vs Avg | Percentile |
|---|---|---|---|
| Fehervary | 21.9 | +1.2 | 83th |
| Jensen | 21.4 | +0.7 | 61st |
| Mantha | 21.7 | +0.4 | 76th |
| Malenstyn | 21.7 | +0.4 | 76th |
| Riemsdyk | 21.1 | +0.4 | <50th |
| Sandin | 21.0 | +0.3 | <50th |
| Kuznetsov | 21.6 | +0.2 | 68th |
| Milano | 21.4 | +0.1 | 64th |
| Protas | 21.4 | +0.1 | 62th |
| Oshie | 21.4 | +0.1 | 61st |
| Wilson | 21.3 | – | 56th |
| Carlson | 20.7 | – | <50th |
| Phillips | 21.1 | -0.2 | <50th |
| Strome | 21.1 | -0.3 | <50th |
| McMichael | 21.0 | -0.3 | <50th |
| Johansen | 20.3 | -0.4 | <50th |
| Backstrom | 20.3 | -1.0 | <50th |
| Dowd | 20.2 | -1.1 | <50th |
| Ovechkin | 19.9 | -1.4 | <50th |
I want to emphasize that Edge does not specify a player’s exact percentile if they’re not in the top fifty. This is apparently done to not embarrass the weakest players in each metric. That’s fine; we can reverse-engineer it.
Alex Ovechkin, 38, has the slowest top speed on the Caps, 19.9 miles per hour, which he achieved in the second period of the season opener. While it’s not possible to quickly compare him to the rest of the league, he appears to place around the bottom few percentiles. Jack Johnson, Cody Cedi, Erik Gudbransson, and Ryan Reaves were all faster. The only player I could find with a lower top speed was Corey Perry, who peaked at 19.7 miles per hour. Ovechkin appeared to finish around the 25th percentile last season, and has been an apparent drop over the course of the data collection.
Note: The league-average line here is according to the 2022-23 season.
For reference, the top speed in the league appears to belong to Winnipeg’s Rasmus Kupari, who hit 24.0 on October 17. Washington’s fastest skater is 24-year-old Martin Fehervary, whose top speed was 21.9 miles per hour, placing him in the 83rd percentile. If we go back a few years, Washington’s fastest skater in the sample was Tom Wilson in 2021-22. He topped out at 23.2 miles per hour, the 94th percentile for that season. Wilson has since had ACL surgery, and his top speed has dropped by nearly two miles to 21.3, placing him now in the 56th percentile.
Much has been made of the age of Washington’s roster, second highest in the league. One predictable consequence of having older players is they’re slower, and in aggregate we see Washington is perhaps the slowest team in the league, at least in terms of top speed, at least according to the few games we’ve seen so far.
One neat feature of the Edge site is their comparison tool. Using radar charts, which are confusing to me, we can visualize how the Caps compare to their next opponent, the Toronto Maple Leafs. In the visualization below, Washington is the black shape, and Toronto is the gray shape. The bigger the shape is in each direction, the better the team is in that metric. Read it and weep:
Washington barely registers on the radar chart in any direction. We can hope that Washington will improve in at least some of these stats as the sample grows, but at the same time I feel obligated to note that their opponents so far averaged just 84.4 standings points last season. These were supposed to be easy games, but they’ve made the Caps look small.
Headline photo: Alan Dobbins/RMNB
