Jason Garrison, the Brian Campbell Effect, and Regression

Newly minted Canucks defenseman Jason Garrison was, like many of Mike Gillis’ recent draft picks, a "late bloomer." Garrison went undrafted and didn’t even break into the NHL until he was 24. At 27 (he’ll turn 28 before the beginning of the 2012/13 season) he’s only played two and a half seasons of NHL hockey, and needless to say that isn’t the deepest track record for a guy who was just signed to a six year deal…

In his first half season in the NHL (2009/10), Garrison was a sheltered third pairing defenseman who got his teeth kicked in by NHL competition. In his first full season (2010/11) he played extremely tough minutes in a shutdown role with Mike Weaver, one of the most under-rated players in the league, and fared extraordinarily well against top-level competition. Then last season (2011/12), Garrison received the gift of playing full-time with Brian Campbell, an extraordinary offensive talent, in a contract year. His offensive production exploded.

So how will Garrison fare in Vancouver without Brian Campbell on his left-side? Was his 16 goal season a mirage? Also, how responsible were Weaver and Campbell for Garrison’s success over the past two seasons?

These questions are pertinent – the Canucks just made a significant commitment to Garrison (six years and a full no-trade clause). In the minds of many Canucks fans, and observers around the league: Jason Garrison, the White Rock native with the staggeringly powerful point shot still has a lot to prove. So let’s wade into the data and see what the numbers can tell us about Garrison’s play the past two seasons, and his offensive breakout last year.

Let’s start back in 2010/11, the season in which Jason Garrison really "broke out." Sure casual fans still didn’t know anything about him, but he became a darling among those who pay close attention to advanced stats. Jonathan Willis praised Jason Garrison’s defensive play, and James Mirtle named him the first runner up for the Rod Langway award given to the NHL’s best defensive defenseman (Dan Hamhuis was #1 that season). 

During the course of that 2010/11 campaign, Jason Garrison faced the toughest minutes among all Florida Panthers defenseman, while starting most of his shifts in the defensive-zone. Despite the difficult circumstances, Garrison came out even in possession which, indicates that when Garrison was on the ice the Panthers won the on-ice battles, and managed to move play in the right direction. While Garrison deserves a fair bit of credit for this, his most frequent defensive partner that season was former Canucks cast-off Mike Weaver – who, by the numbers, is one of the best defensive defenseman in the game.

What follows is a quick "With or Without You" or WOWY table, and the numbers are taken originally from David Johnston’s Hockey Analysis site. Included below are Corsi numbers (CF = Corsi Event For, CA = Corsi Event Against) and goal differential (GD) numbers. They’ll allow us to see how the Panthers controlled possession with Garrison and Weaver on the ice, with just Garrison on the ice, and with just Weaver on the ice. I’ve included goal differential as well because frankly, results matter too – even if goals are mostly  random events! To the numbers:

(A quick primer for those new to advanced stats: a Corsi number is like an expanded version of +/- that accounts for all shots, blocked shots, missed shots and goals both for and against. It’s used as a proxy for possession, though it also correlates closely with scoring chance data and also: with winning hockey games).

A quick glance at this table and the Corsi differential numbers in particular, might lead you to conclude that Weaver and Garrison were worse off playing together than they were playing apart. That would be misguided because, remember: Weaver and Garrison were Florida’s shut down pair in 2010/11.  When Garrison and Weaver played together we can say with a reasonable degree of confidence that their circumstances were more difficult than when either player was on a different pairing (playing with Dennis Wideman, or Dmitry Kulikov). 

When Garrison played without Mike Weaver he was a positive possession player, but also a ridiculously high-event defenseman. Nonetheless, he finished with an even goal differential though a lot of pucks went in at both ends during those 200 odd even-strength minutes.

Without Garrison, Weaver was an even stronger possession player though again: lots of pucks went in at both ends during those 350 odd even-strength minutes, and the Panthers were burned more often than they scored on their opposition. Playing together the two limited events considerably, which, is quite an accomplishment considering their role and deployment. As a result: few goals went in either way when they were paired together.

This was Garrison’s first season in the NHL, and while he certainly wasn’t quite the impact possession player that Mike Weaver was and is – looking at the differential of each defenseman when paired together, and when split up – I’d still describe their partnership as exhibiting synergy. Certainly you wouldn’t describe Garrison as an anchor here…

This past season Jason Garrison played alongside blue-chip offensive defenseman Brian Campbell, a dynamic puck-mover, in a more offensively oriented role. He also spent a significant amount of time playing the right-side. The Campbell and Garrison pairing started less frequently in the defensive end than Weaver and Garrison’s pairing had in the season previous; and while they played against the toughest competition the opposition had to offer, they also played with the most talented offensive players on the Panthers. 

We’ll get directly to the numbers, which, are laid out exactly like they were in the table above (Corsi for and against, and goal differential are included):

Here we see a somewhat similar pattern in the symbiotic relationship between Campbell and Garrison, as we saw above between Weaver and Campbell.

Garrison and Campbell were significantly better when paired together than they were apart. When Campbell wasn’t paired with Garrison, he most often played with lesser possession players like Erik Gudbranson, and Dmitry Kulikov, meanwhile, when Garrison was separated from Campbell he played most often with Mike Weaver (presumably in a more traditional shutdown role).

When they were on the ice together, Campbell and Garrison’s +7.69/60 Corsi differential is impressive, and the fact that the Panthers didn’t outscore their opponents by a wider margin looks to me like pure luck (both Campbell and Garrison had an on-ice shooting percentage below 6% this past season). Regardless that’s some Bieksa-Hamhuis quality ice-tilting.

When Campbell played apart from Garrison, he played an appreciably more high-event game – there were more shots and chances both ways, and the puck ended up more often in Florida’s net, than it did in the opposition’s. When Garrison was separated from Campbell, the Panthers were outshot, but managed to score as often as their opponents did. 

The trend from the past two seasons indicates a couple of intuitive things. First of all, the numbers indicate that Garrison is more effective when paired with a quality defensive partner (who isn’t). Secondly, while Garrison’s partners tend to perform better than he does when they’re separated, he’s generally remained an effective possession player and an even player in terms of goal-differential regardless of who he is paired with.

The sample in this case is only about 500 minutes large, which, doesn’t tell us that much – but the available data does suggest that Garrison is effective on his own, and in fact helped both Weaver and Campbell to be successful. At the very least, there’s no reason to suspect that Garrison was "dead-weight."

Jason Garrison got paid this summer (27.6 million through 6 seasons, and a no-trade clause) largely as a result of his gaudy goal totals from last season (he tallied 16 times, good for third most among all NHL defenseman). On the Canucks, he’ll be expected to be what he is: a steady defensive blue-liner who can find the back of the net more than your average defenseman. Cam Charron today, tempered expectations when he wrote that he expected Garrison to score "upwards of five goals" this coming season. Looking over the data, that looks to me like a base-line minimum.

After all, based only on even-strength production (Garrison scored 7 even-strength goals this past season, and added 9 on the power-play), Garrison’s goal totals this past season weren’t insanely out of line with his totals the season previous. The table that follows includes all of Garrison’s even-strength goals (4-on-4 and 5-on-5), it includes his even-strength shots, his shot rate per sixty minutes, his personal shooting% and his on-ice shooting%.

In terms of Jason Garrison’s shooting, he actually shot less this past season, so the jump in his goal totals was driven by the percentages and not by volume. While Garrison’s personal shooting percentage took a jump, his on-ice shooting percentage was extremely low for the second year in a row. On-Ice shooting percentage tends to be very stable league-wide, and it is a rare outlier who carries an on-ice shooting percentage below 7% over a large enough sample. If you were wondering why Garrison recorded so few assists this past season, that’s probably the answer.

Many expect Jason Garrison to regress this season, and they’re correct, but maybe not about what direction that regression will be in. Either way, based on his even-strength goal scoring over the past two and a half seasons (Garrison also had 2 even strength goals in 38 games in his rookie campaign in 09/10) five goals is the bare minimum to expect from Garrison – assuming that he can stay healthy, and that the season isn’t shortened by a lockout.

If Jason Garrison had added 3 power-play tallies to his 7 even-strength goals this past season, he would’ve been worth about a million dollars less per season on the market. But he tallied 9 times on the power-play, second most among all defenseman behind only Shea Weber.

The Canucks paid up, but power-play production tends to be more inconsistent than even-strength production is. So if you’re going to argue that Jason Garrison rode Brian Campbell’s coattails to a 25+ million dollar contract – the power-play might be the place to start. After all, Brian Campbell assisted on 5 of Garrison’s 9 power-play goals, and was the clear quarterback of Florida’s man-advantage unit.

A quick side-note: before he was traded to Florida, the Blackhawks played Brent Seabrook more on the power-play in 2010/11 than they played Brian Campbell – who had thirty power-play assists this season (not a typo). I’m a big Joel Quenneville fan but talk a misallocation of resources…

Three of Jason Garrison’s power-play goals came while Florida had a two man-advantage, so, at the very least Garrison’s shot should serve to replace Salo’s artillery fire in those rare situations. The other six came with the Panthers up five-on-four and that’s what I’ve chosen to focus on, for the purposes of simplicity.

Heading into this season, Jason Garrison had never scored on the power-play in well over 125 minutes of ice-time with the man-advantage. So to what extent was Garrison’s power-play production an aberration driven by Brian Campbell, and luck more generally? The below table breaks things down based on goals and shots (both are also rated), as well as power-play shooting percentage, and on-ice shooting percentage:

These samples are miniscule, so we can’t proceed with too much certainty. It looks to me, however, like Garrison’s lack of power-play production in 2010/11 was largely bad luck. A 5.56% on-ice shooting percentage is ugly at even-strength, but on the power-play? To put it in perspective, no Canucks defenseman who regularly received power-play time has carried an on-ice shooting% below 10% since the 08/09 season.

When you look at the shot-rate in particular, it doesn’t look like the addition of Brian Campbell helped to provide Jason Garrison with appreciably more opportunities to unleash that cannon of his. Also, his 13.6% personal shooting percentage with the man-advantage is high, but it’s not outrageous. Six Canucks skaters (including one defenseman) carried a 5-on-4 shooting percentage above 13.6 this past season, and two defenseman in Alex Edler and Dan Hamhuis had a higher power-play shooting percentage in 2010-11.

I don’t mean to pull a Coach Q and undervalue the impact a puck-mover of Brian Campbell’s pedigree can have on a team’s power-play unit, but it’s hard to look at Garrison’s offensive production and see anything altogether that unsustainable about it. After all, Henrik and Daniel Sedin are quality playmakers in their own right, and while Alex Edler is often engaged in a mortal battle with consistency, his passing ability is beyond doubt. Surely they can replace Campbell’s cross ice feeds to tee up Garrison’s one-timers…

Looking at the data, it seems to me that if Garrison is able to earn regular ice-time in the top-four, and plays at least second unit power-play minutes: the idea of him hitting double-figures in goals again next season isn’t much of a stretch.