Since the end of the season, the position that has been weakened the most on the Capitals has undoubtedly been the defense. Losing Schmidt, Alzner and Shattenkirk looks really bad on paper and raises quite a few questions: “Who is going to take Alzner’s heavy minutes?”, “Is Carlson good enough to anchor a second pair with a rookie?”, “Will the rookies be good enough?”, “Is Orpik going to be an anchor (the bad kind though) on the third pair?”. Well, let’s try to answer these questions.
After reading this really good article on The Athletic on Ron Hainsey and how quality of competition impacts a player’s performance, I asked Dom Luszczyszyn to make a graph of the Caps defense and he was very kind to answer:
(QoT: quality of teammates, QoC: quality of competition)
(The percentages in the boxes represent which “percentile” of that category the player fell, i.e.: a player in the 90th percentile for QoC means that the competition he had to face was higher than 89% of the players in the league. Colors also help – red means “tough”, so hard competition/bad teammates and blue means “easy” so weak competition/good teammates.)
(Game Score is a stat which focuses on 5v5 play and ignores special teams. If someone wants a thorough explanation of the stats, here’s where Dom breaks it down)
(GS effect represents how QoC and QoT influenced the player’s performance as represented by the Game Score. A GS effect below 0 means that the player’s deployments impacted negatively on the player’s performance, so we’ll have to adjust the player’s Game Score by adding that value. Mirror it for GS effects above.)
Takeaways
- Alzner’s minutes may have been tough from a competition standpoint (especially at forward), but his teammates were amazing. Very few players in the league benefited from better teammates than he did. His negative Game Score could mean that losing him is literally addition by subtraction.
- Niskanen is the best Dman on the team and Orlov is not far behind. Good that we’ll have them for a long time. Carlson is good, but not that good.
- Schmidt to Vegas isn’t that big of a loss. He had a good showing in the playoffs, but in the regular season he benefited from some really easy deployments. To put it in perspective, Schmidt was closer to Orpik than Carlson as far as impact at 5v5 goes.
Let’s answer some of the questions
“Who’s going to take on Alzner’s heavy minutes?” – Orlov will be just fine. If we gave Alzner’s minutes to Orlov this past season, his impact would drop from 0.27 to 0.24, nothing huge and still very good. To give some perspective on Game Score values, Erik Karlsson is at 0.37 (adjusted). If needed, Orlov and Niskanen could also take on an even bigger role.
“Is Carlson good enough to anchor a second pair with a rookie?” – If we put Carlson in Orlov’s 2016-17 minutes, his impact would rise from 0.11 to 0.15. That’s not something to laugh at. If he doesn’t get paired with the likes of Alzner, our American Hero will do well on the second pair.
“Is Orpik going to be anchor (the bad kind though) on the third pair?” – Orpik in a sheltered role does just fine, not 5.5M a year fine, but not on the levels of screaming for a buyout. One could say that being paired with Schmidt helped him more than numbers can measure, though my counterpoint would be that he’s competent enough to not drag his partner down when the partner is good.
This leaves us with one last question, “Will the rookies be good enough?”. For this I go back to Luke Adomanis’s piece on analysis of the Caps’ prospects and here are the bits on the guys most likely to make the jump on defense:
Djoos
Even last summer, some had Djoos pegged a Top 6 or depth defenseman at best and even then some were skeptical. He stepped up in a big way once Bowey and Aaron Ness went down with injuries and he had a monster season. Most believe he has the ability to be even better. Is a second-pairing role out of reach? It’s hard to tell, but that’s not saying he can’t do it. All he does is prove people wrong. A few think he can at least solidify himself as a strong bottom-pairing defenseman with potential to be more.
Bowey
Known as a vocal leader with a booming shot and great puck-moving abilities that can put players on their bottoms with devastating hits, Bowey has developed into a great all-around player.
Lewington
Of all the defensive prospects, Lewington probably fits the “defensive defenseman” title the best. Many never really hear about him, but the organization absolutely loves this guy.
Takeaways
- Given handedness, Bowey will likely slide on the third pair, he’s of the “puck-moving” mold, so we should expect at least a baseline level of play from him and Orpik.
- The Caps should be fine as long as Djoos will be the one on Carlson’s side (assuming that the Swede will gain some pounds). Things will get dicey if Lewington will be there, “defensive defensemen” are in extinction for a reason. Choosing Lewington over Djoos would mean that the team is going against the evolution of the game, which would not be good.
- There could always be some big surprises in camp – Siegenthaler, Johansen, and Hobbs have all taken big strides this past season.
Conclusion
The Caps have such a good first pair in Orlov and Niskanen that they could give them all the tough minutes, to the extent of sheltering both the second and the third pair. This could be needed to begin the season, as the rookies will need to adapt to the NHL given that neither of them has ever been played at the highest level. Ideally, throughout the season, the second and third pair’s roles will increase. Best case scenario, the Caps defense next season will be even better than last year’s.
Obviously, with my arguments based mostly on stats, they don’t tell the whole story and have to be taken with a grain of salt.
By Liviu Damaschin
I think there’s a typo in the explanation of colors and quality of teammates. I was reading the graph and thought Alzner had played with not-so-good forwards, until the text said otherwise. It seemed to make sense, since he’s often deployed with checking players, and the Caps’ checking players, while tough and effective, don’t score a ton? Anyway, I like your final grain of salt. I wonder if there is a statistical measure of how much a player’s performance per minute declines as his minutes increase from one year to the next. Obviously there’s room for huge variation in summer conditioning, but it would seem like a curve worth studying, if GMs want to acquire players likely to improve. And even with a great summer, there must be a limit to how much muscle a player can build. The variation is why there is training camp. The predictable element is why there are statistics. Anyone want to figure it out?
Hi!
There’s no typo on the table. The colors and the percentiles reflect how much tougher the category considered makes it for the player.
A high percentile in one of the Competition column means that the the player is facing good players, while a high percentile in one of the Teammates columns means that the player is on the ice with bad teammates. Those two conditions mean that the player is being put in a tough environment, so the color used to represent them is red.
On the other side of the spectrum a low percentile in Competition means that the player is facing bad players and a low percentile in Teammates means that the player is on the ice with good players. These two situation represent an easier deployment for the player.
Specifically to Alzner’s situation, while he played some minutes with the Beagle line, he also played much more with Kuzy and Backstrom. All of this while not contributing much and actually being a drag to them, this results in him being in one of the sweetest spots in the league as far as quality of teammates goes.
Erik Karlsson is in the exact opposite scenario (100th percentile in QoD and 99th percentile in QoT), and this is how Dom explained his situation:
“Erik Karlsson, thanks mostly to very harsh teammate numbers. With his numbers, you’d expect a 5-on-5 Game Score QoT of 0.49/0.38, when in reality they were at 0.33/0.16. It’s possible those are inflated because they’re hard numbers for his teammates to attain, but since it’s compared to what’s expected of a player posting his numbers, it means he carries a huge burden with little help given how far the rest of his teammates are from him. The rest of the list isn’t too surprising.”
In fact the measurements used in the graph not only define in how good teammates and competition are, but are also a way to give credit of what is happening on the ice.
Alzner’s teammates have always played better when they were apart from him, that means he’s the problem and his teammates have been bringing him up and they did so to the extent of him being in the 2nd percentile (so only one percent of the players in the league are in a better position than him).
This is how Dom calculated QoC and QoT:
“Quality of competition was easy to measure, it was just the weighted average of every opponents 5-on-5 Game Score above average as provided by Alam. Using the same method for teammate quality created some hiccups, though, as it was highly correlated with a player’s Game Score, which in turn would grossly overrate fourth liners and underrate top players. To combat that, I measured teammate quality relative to what’s expected of the player given their Game Score, which should give appropriate credit to the players providing the most value compared to the strength of their teammates.”
There is a typo in the table description. It reads:
Colors also help – red means “tough”, so hard competition/bad teammates and blue means “easy” so weak competition/bad teammates.
I suspect blue is supposed to read “weak competition/good teammates?”
Oh, you’re right, thanks! I read the comment as “there’s a typo on the table”, not in the description.
My bad. Thanks for pointing it out.
i think there’s still a typo:
” red means “tough”, so hard competition/bad teammates and blue means “easy” so weak competition/bad teammates.”
presumably, it should read “blue means easy, so weak competition/good teammates”?
oops, got in after that other comment was posted. JR way
Pingback: How Much Can Brain MacLellan Be Blamed For This Summer? | NoVa Caps
Pingback: How Much Can Brian MacLellan Be Blamed For This Summer | NoVa Caps
Pingback: Optimizing the Capitals’ Forward lineup | NoVa Caps
Pingback: The Two-Year Plan for the Washington Capitals’ Defense | NoVa Caps