About a month ago, now, I wrote an article about the Devils defenders and which ones I thought should be getting more/less playing time. At the time, I argued that there was no reason to not try out Dalton Prout a little more, Mirco Mueller was getting buried, John Moore should see decreased playing time, and Steve Santini was playing strong in difficult situations.
I make too many claims to not get used to saying the following: It’s possible I was wrong.
I still don’t know why we don’t try Prout. He was fine(ish) when he played, and there just no reason to keep a guy on the roster if you’re going to sit him 12 games in a row. I still think Moore needs less playing time. Mueller, however, played much better before the injury. And Santini’s previously justifiable possession stats have since completely fallen out of any realm of respectability. I’ll look at what’s gone wrong, see what is his fault, what is not his fault, and what the team can do about it.
What’s Been Happening To Him?
At the time of that article, Santini was just 3 high-danger chances short of 50%. He is now the lowest on the roster in percentage, and is allowing them at a rate higher than any defender not named John Moore (per Natural Stat Trick). Also, Andy Greene’s CF% plummets by over 10% points when he’s with Santini. Granted, this is likely in part due to the fact that Greene starts in the offensive zone 57% of the time when he’s away from Santini, and 41% when he’s with Santini. Santini’s ZS% actually goes DOWN to 31% when he plays away from Greene. This means that , in a departure from prior years’ philosophy, Hynes really uses Santini the most heavily in the defensive zone by far.
And this is the first example of part of the reason why Santini is struggling. Not only is he being given the most difficult zone starts to deal with, but he’s being given them in insane situations. Based on this chart from hockeyviz, Santini is being put up against the other teams top 4 skaters a ridiculous about of time. And he’s actually been facing them disproportionately frequently with our bottom 6 skaters.
Let’s use the Minnesota game as an example, a game in which the Devils were outshot 20-10 when Santini was on the ice. How much can be reasonably expected of you when you are starting in your own zone, playing against a line like Granlund, Koivu, Zucker which is among the best lines in the league with a line like Coleman-Noesen-Gibbons or, in the case of this game, Wood-Boyle-Stafford.
What IS His Fault?
I’m making some excuses here for Santini because he is undoubtedly put in the worst circumstances of any defender on the team — possibly the league. However, there’s simply no accounting for how terrible some of those stats have been. As a brief example, if you use Corsica’s adjustment, he gains 2% ... which increases his Corsi from 36% to 38% -- still lowest, not only on the team, but in the NHL.
This is a degree of futility simply unattributable to any context. Remember that zone start effects fade around 15-20 seconds after the faceoff. Also, no matter how often he plays against the top line with the bottom line, shifts are so mobile that the only really consequential on-ice companion/opponent is going to be Andy Greene — who’s actually good. It’s rare to see a top-pairing defender spend THAT much more time with any particular line. As an example, in the dumpster fire against the Panthers Monday, where Santini once again brought up the possession rear by losing the corsi battle 9 to 25, he didn’t play against any single player longer than 5:22 minutes.
So now I get to part of what I think the problem with his game is. I spent the better part of Thanksgiving weekend scraping, cleaning, compiling and assembling data from Corey Sznajder’s All-3-Zone project into a 3-page Tableau that paints a picture of a player’s contributions to the ultimate goal of getting out of the defensive zone (exits) into the offensive zone (entries) and generating shots for your self or others (shot contributions). You can find it here and I’ll embed it at the end of the article if you want to play with it. So far, he’s tracked 8 Devils games off which these conclusions will be based. If you want him to keep tracking extremely helpful data like these, please contribute to his Patreon, like me.
So how does Mr. Santini show in these statistics? Seen below, Santini is 5th out of our 7 defenders in shot contributions as well as in shot assists per game so he generates nothing offensively, but we probably knew that one. It also may not have anything to do with his offensive prowess, and might just be that he’s always in the defensive zone.
But that’s actually where we run into the big problem with his game. Santini is not particularly adept at escaping the zone. Seen below, he is 5th in zone exits per game, and 5th in possession exits per game. Given the fact that he starts so often in the defensive zone, you would have to expect him to have among the highest exit rates on the team simply because he has more opportunities to exit the zone.
What Do We Do About It?
There are a few options we have for how to fix this problem. The 3 options I’d imagine we have with how to address it are:
- Distribute Ice Time More Evenly
- Give Defensive Starts to Other Players
- Change the Linup
#1 Distribute time more evenly between pairings
This one is actually already being done. If you look at the shift charts from hockeyviz, you’ll see that their usage has gone down, and Moore has been the one making up most of the difference. The red+black in the image below is the Greene Santini ice time. You’ll see that those bars are decreasing in length, but by the time we get to Moore’s bar it’s leveled off, indicating he and Severson have filled in those vacated minutes. In fact, they have lead the team in ice time the last two games, overtaking Greene and Santini. In fact, Lovejoy and Butcher have been getting more time as well.
#2 Give other pairings more demanding shifts
If you look at the Florida game, you’ll see the Lovejoy was placed back in the defensive zone a lot — 0 OZS and 6 DZS. He actually handled it okay (10 CF, 13 CA). As bad as everyone thought Lovejoy was last year, the Lovejoy-Greene pairing actually did quite well considering the circumstances.
Notice that the Greenejoy pairing had essentially the same ZSR (zone start ratio) as Greenetini, but the CF% and xGF% were both 7% higher than this year’s pair. Remarkably, while Greenetini has given up 10 goals so far this year, Greenejoy gave up 10 goals ALL of last year.
By virtually any metric, Lovejoy can handle those starts better than Santini. Which brings me to the third option.
#3 Change the lineup
Changing the lineup is harder than it sounds. When you move one piece, you have to move others to compensate. That means that some players will play in tougher positions or more minutes than they are accustomed to. We can Corsica’s pairings tool to see what options we have. I assumed that Santini stays in the lineup, and that Mueller stays injured. I only used pairings that have already played 15 or more minutes together and I used historical time on ice information, and a little guessing and checking to fill in the projected 5v5 time on ice for each pairing. That aspect of this was stunningly unscientific. But it produced the following possible lineups.
As you can see, fixing the problem is easier said than done. We simply don’t have the defensive personnel to make a significant improvement by just moving bodies around. Every other feasible option I concocted had an even lower CF% than our current arrangement. Only the #QuickFix where we swap Severson and Santini does the xGF% improve.
Conclusion / Your Thoughts?
I would use a combination of all 3 options I mentioned to address this issue. I would probably go back to Greene-Severson. I’d up the ice time for Butcher/Lovejoy, use that pairing more in the defensive zone, and possibly occasionally double-shift Greene with Lovejoy or Santini in those situations as well.
What do you guys think we should do? Is Santini a problem? If he is, how do we fix it? Would you go so far as to sit him? Change the lineup? Ice time? Deployment? Thanks for reading and leave your thoughts below!
P.S. Here’s the tableau I used to generate the above visualizations. If you’re on a mobile device or something it may not show well here and you’ll have to go to the actual website and request the desktop version. Hopefully it works more or less the way I expect. Enjoy!