I understand I’ve given you over 3,000 words, a bunch of charts, and a whole lot of numbers and stats about Ben Lovejoy. All of this can be summed up in the following pithy statement:
Ben Lovejoy has been very bad for the Devils in 2016-17.
So before I get into the article, let me set the scene here. I am aware that Ben Lovejoy has been one of the— if not THE — most maligned Devil this year. I am fighting an uphill battle to convince the masses that it is better to have Ben Lovejoy on the team than to not. So I’m gonna try to bring my A-game, and you should try to bring an open mind. Though I can’t stop you if you choose not to.
My case is going to be defined as follows. I am going to make a logical argument, then give a quick statement about the evidence for each assertion, before finally providing statistics to back up each claim.
Logical Thread of Assertions
- Ben Lovejoy is contributing positively to this team.
- His production correlates to wins.
- Therefore, we should hope/like that he will be back next year.
This is the goal. My contention is that I can provide evidence for each assertion, and if I do so, there will no space to accept the alternative, which is that we should not like/hope that he will be back next year. Seems like a Herculean task right? Well I claim that I can support each of these statements using the following evidence. The numbers for the evidence-based claims correspond to the numbers for the logical arguments.
Evidence For Assertions
- The best defensive pairings on this team have included Ben Lovejoy.
- We win more when Ben Lovejoy plays more.
- Ben Lovejoy has been improving over the course of the year.
The Case for Ben Lovejoy
Okay, enough messing around — let’s do this thing.
Evidence for Assertion #1: Our Best Pairings have included Ben Lovejoy
If you go to Corsica’s “Combos” section, they have some really interesting tools to investigate WOWY-type information. Among them is a pairings tool. The chart below was created with Score, Zone, and Venue adjustment and a 200 TOI minimum. All other settings were default.
The xGF% is “expected goal ratio” and it accounts for several factors that contribute to the likelihood of shots becoming goals. The SCF% is “scoring chance ratio.” You will notice that Greene-Lovejoy and Merrill-Lovejoy were the best two pairings that played significant time. Better, even, than every Severson pairing.
How can this be? John’s research was all accurate so how is it possible to reconcile this Tale of Two Lovejoys? Well, there’s a few reasons.
Most obviously is the fact that Lovejoys next most common pairings were among the worst on the team. I will address this fact, and why I think it’s an unconvincing counterargument in my Evidence for Assertion #3.
Alternatively though, is the fact that John looked mostly at CF% which is a proxy for possession. And the narrative of Lovejoy being a poor possesion player — even if you adjust for his zone-starts — definitely has strong support. However, contrary to what some may tell you, possession is not the whole story. According to these statistics the Lovejoy-Greene pairing is definitely our best one with regards to scoring chances and so, while the possession numbers are not favorable, their net impact on the game is positive. Furthermore, this is not just buoyed by Greene, If you investigate Corsica’s WOWYs you’ll see that — while most skaters still decrease even in xGF% with Lovejoy, the most frequent partner and the one with most comparable situational starts actually improves with him. That being said, Jon Merrill is likely more to credit for the second pairing succeeding — Lovejoy significantly improves with him, and Merrill gets slightly better without him.
Lastly, I should mention that Corsica’s correction for zone-starts vaults Lovejoy-Greene up way to the top. Why is this? Of those 6 pairings listed above, the two Lovejoy Pairings started in the offensive zone 14% and 18% of the time. Those are the two LOWEST FIGURES IN THE NHL.I repeat, Greene-Lovejoy, and Merrill-Lovejoy start in the defensive zone more than any other pairings in the NHL. Lovejoy is on both of the most defensive pairings. The other 4 on the Devils ranged from 30%-45%. The next lowest in the NHL is the Ericsson/Ouellet pairing on Detroit with 21%. So, as you can see, the Greene-Lovejoy pairing is off the map entirely when it comes to how much is asked of them — reminiscent of Larsson-Greene in that respect. The fact that the do so well at it is a true testament to the success of that pairing.
Evidence for Assertion #2: Ben Lovejoy Helps Us Win
The next thing to prove is does this correlate to wins.
For defenders, 20 minutes seems to be the marker for being relied on heavily. Sixty minutes in a game divided by 3 for the number of pairings make it the average. Lovejoy’s happens to be closer to 20.8 minutes, but in the interest of simplicity I’ll use 20. Below is the Devils record in games that Lovejoy has recorded at least 20 minutes as opposed to less than.
Now, a portion of this can be attributed to the fact that Lovejoy plays more with the lead. He drops from first in ice time with the lead, to second in ice time when behind. Lovejoy provides 20% of all defensive ice time when behind, but that shows only a modest rise to 21% when ahead. So I can’t attribute THAT much value to this argument.
That beind said, there IS something to be said about that utilization. Should he lead the team in ice time when ahead? Well if you go to Corsica’s skaters page and navigate to the “Custom Query” you can filter to look only at stats when the team is “leading.” If you do so, and adjust for zone effects, the following is the result.
Lovejoy is a positive force when leading in both expected goals and scoring chances. And when it comes to scoring chances, the team improves their scoring chance ratio by almost 10% when he is on the ice — the highest such figure among all defenders. He clearly has a role, and it helps us win.
Evidence for Assertion #3: Ben Lovejoy Is Improving
I mentioned in my evidence for assertion #1 that I would explain why I excluded Lovejoys worst pairings from the output in this section. The reason is that it doesn’t demonstrate the player he is right now or the role he currently inhabits. The fact that his numbers are improving shouldn’t be surprising at all. Lovejoy had previously played with Auvitu and Moore which produced 2 of the 3 worst pairings in terms of expected goals. He was moved (at about the new year it seems based on my perusal of shift charts) from that to the two BEST pairings on the team. It took a while for that shift to manifest positively, but by mid-February, Lovejoy seemed to rise reliably to a positive player in his new position (if you change it to a shorter rolling period — say 10 games — the shift is visible earlier..
In fact, using the Custom Query, it can be shown that Ben Lovejoy’s Rel xGF% (expected goal ratio relative to the team) was -2.46 (7th worst among NJ skaters) before Greene’s return on February 3rd, to +1.53 (6th best) afterwards.
Ladies and gentlemen of AAtJ, the defense of Ben Lovejoy rests on my ability to prove the logically unassailable thread of statements beginning with Ben Lovejoy’s positive contributions to the team, passing through the comparative success it creates, and arriving at the conclusion that we are better with him than without and we should be excited for him to return. I have provided evidence that Ben Lovejoy’s most recent, and most frequent pairings have been the best on the team in terms of zone-adjusted expected goals. The role of his game in contributing to wins is demonstrated by our win percentage improving by 30% in games in which he plays 20 minutes or more. This is likely due to his status as the best lead-protector on the team, logging the second highest expected goal ratio among defenders and highest scoring chance ratio on the team when leading. This is made even more encouraging by the fact that he has steadily improved over the course of the year, before ultimately becoming a top 10 skater on the team in expected goal ratio in the past 2 months.
Ladies and gentlemen, of AAtJ, the defense rests.