The Washington Capitals’ season has ended — at least for now — in a way none of us could have predicted. The coronavirus has “paused” the NHL season, leaving the Caps alone to consider their fortunes. They’ve been a mediocre .500 team since around Christmas, but they still own first place in the division. They’ve got star talent and a deep roster, but they have played some devastatingly bad games recently.
In this week’s snapshot, I want us to carve out a tiny corner of our minds where we can talk about and think about hockey nerdery, free of blackpilled geohell worry. In this space, we’ll wonder what it might look like when the world and the sport and the Caps finally emerge from the pandemic. In this space, we will keep calm and carry on without putting anyone at risk. And we will do what Chuck Noland in Cast Away says about living in troubled times: “Keep breathing, because tomorrow the sun will rise. Who knows what the tide could bring?”
Forwards
| Player | GP | TOI | SA% | SA% Rel | GF% | PDO |
|---|---|---|---|---|---|---|
| Eller | 69 | 833 | 55.9 | +5.7 | 51 | 0.99 |
| Dowd | 56 | 456 | 54.7 | +3.4 | 62 | 1.02 |
| Panik | 59 | 605 | 54.7 | +3.3 | 65 | 1.04 |
| Hagelin | 58 | 644 | 54.4 | +3.2 | 54 | 1.01 |
| Hathaway | 66 | 612 | 53.8 | +2.3 | 54 | 1.00 |
| Leipsic | 61 | 521 | 53.4 | +1.9 | 54 | 1.00 |
| Backstrom | 61 | 827 | 52.4 | +0.8 | 50 | 0.99 |
| Vrana | 69 | 851 | 51.3 | -0.7 | 54 | 1.01 |
| Ovechkin | 68 | 980 | 50.9 | -1.2 | 48 | 0.99 |
| Wilson | 68 | 900 | 50.7 | -1.7 | 45 | 0.98 |
| Oshie | 69 | 913 | 49.8 | -2.8 | 52 | 1.01 |
| Kuznetsov | 63 | 840 | 45.4 | -8.3 | 50 | 1.03 |
Defenders
| Player | GP | TOI | SA% | SA% Rel | GF% | PDO |
|---|---|---|---|---|---|---|
| Orlov | 69 | 1275 | 53.3 | +2.5 | 49 | 0.99 |
| Dillon | 10 | 165 | 51.9 | +3.2 | 44 | 0.98 |
| Gudas | 63 | 855 | 51.6 | -0.4 | 55 | 1.02 |
| Carlson | 69 | 1192 | 51.4 | -0.7 | 52 | 1.01 |
| Jensen | 68 | 1011 | 51.3 | -0.9 | 49 | 1.00 |
| Siegenthaler | 64 | 780 | 50.7 | -2.2 | 55 | 1.01 |
| Kempny | 58 | 941 | 50.7 | -1.4 | 58 | 1.03 |
Glossary
- GP – Games played.
- TOI – Time on ice in minutes
- SA% – Shot-attempt percentage. The share of total shots attempted by Washington while the skater is on the ice. 50% means even.
- SA% Rel – Relative shot-attempt percentage. The difference in SA% when the player is on the bench versus on the ice. 0% means even.
- GF% – Goals-for percentage. The share of total goals scored by Washington when the skater is on the ice. 50% means even.
- PDO – The sum of Washington’s shooting percentage and saving percentage when the skater is on the ice. 1 means even. The acronym doesn’t stand for anything, and yes, that it is annoying.
Notes
- The Caps were in a precarious moment before the pause. Since the middle of January, they had strung together back-to-back wins exactly once (Pittsburgh and Winnipeg) as they fell from first overall in the league to fending off the Flyers for the division lead. With all the apparent dysfunction in systems (which I won’t get into today), I suspect head coach Todd Reirden was pretty darn close to losing his job on a couple of occasions. But with the frustration and boredom of those recent losses fading from my memory, and now that I’m looking at pattern-level data, I wonder if the Caps had recently turned the corner. Below are two line graphs for the Caps performance during five-on-five play. First is their cumulative difference in expected goals (using Natural Stat Trick’s data) and second is their rolling 10-game average in the same stat.
- To my lying eyes, it sure looks like the Caps hit their nadir just prior to the NHL trade deadline on February 24 — around games 62 and 63, which were incidentally those two messy wins over Pittsburgh and Winnipeg. And while play could not really be called “good” since then, it has probably been better. (Though I want to stipulate that those recent games have still been stupefyingly hard to watch and amounted to just two wins in six games.) There often is a lag between systemic improvement in a team and that team seeing those improvements manifested in wins. I wonder if the coronavirus’s outbreak came just between the two for Washington. There’s some alternate history we’ll never see where maybe the Caps continued losing and made a drastic late-season change, or maybe they started a late-season rally that pushed them into the playoffs with momentum. We’ll never know which it would have been.
- Garnet Hathaway is a rare duck. He has scored 8 goals during five-on-five play this season, powered by a decent shooting percentage: 10.5 percent. But, during his shifts, the Caps as a whole are shooting a team-low 6.7 percent. Everyone else on the ice with him is shooting a miserable 5.4 percent, so while Garnet is generating just a quarter of the shots (24 percent), he’s scoring about two-fifths (38 percent) of the goals. Those are some wild fluctuations in shooting and saving, but if we stopped our analysis at top-line PDO, he’d be a perfectly boring 100.
- I still think PDO is a useful statistic, but its utility come at the beginning of analysis — and it should never be the end of it.
- Hathaway’s production may not be a fluctuation; he’s generating more “danger” per shot attempt than any other Caps player.
- But this is still a volume-based game. Alex Ovechkin has a lower danger per shot attempt than any Caps forward except Richard Panik, but he makes up for it by taking an absurd number of shots. Also, every author of expected-goals models that I know freely acknowledges that available data limits the ability to quantify some tactics and players in particular — Ovechkin being a chief example.
- When I was glancing at the snapshot table, Tom Wilson‘s row jumped out at me. For a player I think of a strong possession driver, I was surprised to see him so low in the stack rank (50.7 percent of shot attempts belong to WSH) and with a negative impact (minus-1.7 percentage points when on the bench versus on the ice). I think I’ve figured out why. You will not be surprised.
| Forwards | TOI | SA% | xGF% | WSH goals | Opp goals | Goal diff |
|---|---|---|---|---|---|---|
| Wilson and Kuznetsov | 398 | 44.5 | 44.7 | 18 | 22 | -4 |
| Wilson only | 502 | 55.3 | 54.5 | 27 | 33 | -6 |
| Kuznetsov only | 442 | 45.6 | 45.4 | 27 | 22 | +5 |
- The Caps control about 45 percent of play when Wilson and Evgeny Kuznetsov are on the ice — a pretty poor showing, with a goal percentage and differential to match. Once Wilson gets away from Kuznetsov (i.e. when he plays with Backstrom instead), he improves a full 10 percentage points — but his on-ice goal percentage is not any better. Meanwhile, when Kuznetsov has played with a different right winger (often Oshie), he has been crushed in process stats like shot-attempt percentage and expected-goals percentage (SA% and xGF%), but commands a positive goal differential. To me, it seems clear that Wilson doesn’t help Kuznetsov much, but Kuznetsov drags down Wilson a lot, which should be instructive for coaching after the pause. This visualization from hockeyviz, which I have butchered, helps clarify the impact — top-right is good, bottom-left is bad:
- I don’t want to beat this dead horse any further, but I think — for posterity — it’d be good to record where Kuznetsov stands among all NHL forwards in a few stats. Of 334 forwards with at least 500 minutes played, Kuznetsov ranks 316th in opponent high-danger chance rate (13.9 per hour), 311th in the difference between opponents high-danger rates when he’s on the ice or on the bench (+2.0 per hour), and 289th in the on-ice percentage of high-danger chances belonging to his team (44.5 percent). I hope this break does him well.
- Back in October, our pal Ben asked me when Richard Panik would get bumped to the fourth line for his poor play. I told Ben he’d like Panik more soon, and by soon I obviously meant February 25. That was when Ilya Kovalchuk (not enough ice time to be on the snapshot yet) took Panik’s spot on the third line. It’s also when the fourth line got back to being dominant. Panik has scored a point per game since then and has been on the ice for just one opponent goal.
- The Kovalchuk impact/Panik adjustment is one of the big reasons I’m optimistic — or maybe was optimistic — about a late-season return-to-form. The other is Todd Reirden’s experimentation with defensive pairs. We’ve seen a ton of different combinations over the last few weeks, with some inconsistent returns. Dmitry Orlov has played well (as expected, with the exception of a stinker in Minnesota), as has Nick Jensen (maybe not so expected and probably driven by PDO). The one emerging pattern I trust is Radko Gudas as the seventh defender and presumptive healthy scratch. I think JP made this case pretty well in a tweet last week:
If you think Nick Jensen has been one of the Caps' best defensemen over the last dozen games or so, there's a pretty good reason for that… (data via @NatStatTrick) pic.twitter.com/DNkXKZc2sh
— japersrink.bsky.social (@JapersRink) March 9, 2020
- I am both happy and surprised to see Michal Kempy doing well. He had been pulling up the rear in the stack rankings, but has been trending upwards since right around new years. If his recovery from last year’s injury is finally over, then I wonder exactly how high his new ceiling may be…
- Games played: Nice.
I’ve put my data for the snapshot and a behind-the-scenes video of me assembling it up on the RMNB patreon.
That’s all for now. Please lay low and stay safe for the next couple weeks. Both with the disease and with the Caps, I have a feeling brighter days are ahead.
This story would not be possible without Natural Stat Trick. Please consider joining RMNB in supporting NST on Patreon.
