clock menu more-arrow no yes

Filed under:

Adam Larsson, Offensive Zone Starts & Goals For Percentage

New, comments

TSN's Travis Yost found a damning trend when Adam Larsson got favourable zone starts with the New Jersey Devils...

Marc DesRosiers-USA TODAY Sports

I wrote an article yesterday looking at New Jersey's offensive strategy through share of iCF by defensemen. At the end, I shared a final thought on a piece by Travis Yost over at TSN.

First, to recap Yost found the following:

So, the good news is that Larsson’s negative impact on shot differentials seems to dissipate some with more favourable offensive zone starts. Larsson’s career norms are right around the 25-45 per cent rate, which is why you see his numbers historically in the red. But when he’s really getting a lot of offensive zone looks relative to teammates, his team-relative shot differential improves to a small degree.

The troubling piece of it is that the tough zone start issue doesn’t seem to explain his poor impact on team goal-scoring rates. You can see that as his offensive zone start% climbs, the team’s goal differentials – both actual and expected – either stagnate or decline. That’s the discouraging piece of this.

That sure sounds funny to me, as I'm sure it does to you. Larsson's shot differential improves slightly in these instances when receiving more offensive zone starts, but his expected and actual goals for decline or stay the same.

Well, in my piece yesterday I offered the following hypothesis:

I imagine the games in which Larsson received more than 55% OZ starts were few and far between, but Yost's article says they exist. I wonder what the circumstances were that saw Larsson see so much offensive zone time when, on average, he saw so little over the course of the season. Could the drop in goal differentials have something to to with QoT? Was Larsson baby sitting a lesser partner or forward line in these games, resulting in this strange drop in actual and expected goals?

My trouble is that I can't find any website or tool that will allow me to look at zone starts for individual players on an individual game basis. If I had a table sorted by ZS% and each row was an individual game listing Larsson's fancy stats, then I would investigate on a game-by-game basis to see who he played with in those games in which he saw favourable offensive zone starts. Unfortunately, I can't.

Is this data available only to those who attend Vic Ferrari's weekly poker night at his stats palace beneath the Air Canada Centre? Am I missing something here?

He didn't reply :(

What I Could Find

is Adam Larsson's WOWY tables going back—I chose—three years. This table sorted by OZS% when players are together with Larsson quickly shows who he was playing with when he got the most cumulative offensive zone starts.

The problem here is that, if Yost were to have used a WOWY table to create his damning graph of Larsson in the offensive zone, then the cumulative OZS% would essentially negate his most common linemates, because as we know, over the course of a season he starts far more shifts in the defensive zone.

Basically, all his most common linemates appear in this table as < 50% OZS% and therefore the games they actually spent starting the majority of their shifts in the offensive zone would not appear on the right side of Yost's graph. All you would get for that > 50 OZS% data would be the small sample dregs—players Larsson has hardly played with, but when they did they were receiving favourable offensive zone starts that result in an average above 55% or so.

I'm not saying Yost used a WOWY table. He's a smart person and I admire his work and I don't think that he did this. But that's all I've got to work with, so let's take a look at who those players were over the last 3 seasons, the players that Adam Larsson saw the light of the offensive zone with to start his shifts.

And here they are, the rag-tag bunch. Note that blank spaces in the table are "0" values.

TOGETHER LARSSON apart TOGETHER LARSSON apart
Player Pos OZone% TOI OZone% GF% CF% GF% CF%
JANSSEN. CAM R 80 13:09 37.3 66.7 51.9 51 47.2
TEDENBY. MATTIAS L 68.7 29:59:00 37.3 33.3 48.6 51.6 47.2
LOKTIONOV. ANDREI C 65.2 75:33:00 36.8 55 51.6 47
CLOWE. RYANE L 65.2 41:28:00 37.2 50 50 51.3 47.1
SALVADOR. BRYCE D 65 32:32:00 37.2 33.3 48 52 47.2
GELINAS. ERIC D 62.7 288:51:00 35.2 46.2 59.7 51.7 45.8
CARTER. RYAN C 61.5 76:16:00 37 37.5 50.8 52 47.1
OLESZ. ROSTISLAV L 58.8 26:35:00 37.3 56.2 51.3 47.1
VOLCHENKOV. ANTON D 56 43:15:00 37.3 75 66.1 50.6 46.9
BRUNNER. DAMIEN C 53.5 147:37:00 36.7 57.1 52.7 51 46.9
JAGR. JAROMIR R 50 239:11:00 36.7 55.6 51.8 50.7 46.8
ZACHA. PAVEL C 50 3:24 37.5 100 77.8 50.6 47.1
BERNIER. STEVE R 49.7 308:12:00 36.3 47.4 51.8 51.8 46.6
RYDER. MICHAEL R 46.8 244:17:00 36.7 42.9 53.7 52.1 46.5
CAMMALLERI. MIKE L 46.2 476:22:00 35.9 62.9 46.1 48 47.4
GOMEZ. SCOTT C 46.2 179:43:00 37.1 62.5 52.3 50 46.8
SESTITO. TIM C 44.4 40:43:00 37.5 50 47.6 51.3 47.2
MATTEAU. STEFAN C 41.5 88:43:00 37.4 25 45.6 51.9 47.2
HENRIQUE. ADAM C 41.4 800:04:00 36.1 61.2 47.4 46.8 47.1
SCHLEMKO. DAVID D 40.9 85:50:00 37.5 50 59.6 51.3 46.8
ELIAS. PATRIK L 39.1 477:13:00 37.3 46.4 46.1 52.3 47.4
ZUBRUS. DAINIUS C 38.7 421:59:00 37.4 43.5 50.6 52.6 46.6
STEMPNIAK. LEE R 37.9 340:37:00 37.5 52 43.6 51.1 47.6
LARSSON. ADAM D 37.6 2931:44:00 51.3 47.2
BOUCHER. REID C 35.9 237:44:00 37.7 50 43.9 51.4 47.5
GREENE. ANDY D 35.4 2061:11:00 44.6 52.9 45.8 48.1 50.5
ZAJAC. TRAVIS C 34.9 910:39:00 38.9 54.5 43.9 50 48.6
RUUTU. TUOMO R 33.8 319:54:00 37.9 60 52.1 50.7 46.6
JOSEFSON. JACOB C 33.6 364:35:00 38.1 26.7 50.3 53.8 46.7
THOMPSON. PAUL R 33.3 14:24 37.6 100 37.5 51 47.2
PALMIERI. KYLE C 32.2 547:35:00 39.2 64 43.4 48.9 48
TLUSTY. JIRI C 32.1 134:33:00 37.9 100 50.7 50.6 47
TOOTOO. JORDIN R 31.7 381:16:00 38.3 33.3 45 52.7 47.5
GIONTA. STEPHEN R 31.5 532:31:00 38.6 34.5 48.9 55 46.8
BLANDISI. JOSEPH C 30.9 182:12:00 37.9 66.7 46.4 50.3 47.2
SISLO. MIKE R 30.8 90:10:00 37.8 100 43.4 50.6 47.3
KALININ. SERGEY C 30 333:46:00 38.7 53.8 41.9 51 47.8
MERRILL. JON D 30 139:51:00 37.8 44.4 44.1 51.7 47.3
ZIDLICKY. MAREK D 30 23:35 37.6 42.5 51.3 47.2
PIETILA. BLAKE L 30 20:47 37.6 44.4 51.6 47.2
HAVLAT. MARTIN R 29.9 148:33:00 38 27.3 42.2 53.1 47.5
HELGESON. SETH D 28.8 95:47:00 37.8 50 39.1 51.3 47.5
FARNHAM. BOBBY R 28.6 133:58:00 37.9 50 48 51.3 47.2
WHITNEY. JOE R 28.6 9:50 37.6 23.5 51.6 47.3
O_BRIEN. JIM C 27.3 15:44 37.6 37.5 51.3 47.2
SMITH-PELLY. DEVANTE R 25.7 77:43:00 38 80 46.2 50.3 47.2
KENNEDY. TYLER C 25.6 183:45:00 38.4 40 49.3 52 47.1
MOORE. JOHN D 25 51:25:00 37.8 100 38.3 50.3 47.4
HARROLD. PETER D 21.4 84:14:00 38 100 45.6 51 47.2
O_NEILL. BRIAN C 19.1 64:31:00 38.1 33.3 43.3 51.6 47.3
SEVERSON. DAMON D 42:20:00 37.7 29.9 51.6 47.5
FRASER. MARK D 26:16:00 37.7 50 17.6 51.3 47.5
WOOD. MILES L 6:04 37.6 80 51.3 47.2
FAYNE. MARK D 4:49 37.6 25 52.3 47.2
WARSOFSKY. DAVID D 3:25 37.6 100 51.3 47.2
MOZIK. VOJTECH D 1:40 37.6 51.3 47.2
GRAGNANI. MARC-ANDRE D 0:36 37.6 51.3 47.2

There are only a dozen or so players with whom Larsson has had favourable OZS over the past three seasons, and in most case he spent little time on ice with those players as well. Here we see what Yost concluded in his article: with seven of those ten players, Adam Larsson had subpar to miserable Goals For %, but saw his Corsi For % remain reasonable.

He managed to maintain a > 50% GF% when playing with Damien Brunner, Jaromir Jagr, Anton Volchenkov, Cam Janssen (though they only played 13 minutes total together so sample size is obviously an issue), and Pavel Zacha (sample size alert).

The rest of the players, with whom he saw the NJD share of goals drop, were almost all fringe NHLers or journeymen at the time, and almost all of them are no longer seeing a regular shift in the NHL.

I wish I could see the data Yost was working with, because the WOWY table above suggests that Larsson was playing with subpar linemates in almost all cases of favourable OZS%. Over a small sample with those linemates, the shot attempt share remained reasonable but the share of goals dropped considerably for the Devils.

Why might this be? Well, the bottom line is that these few players were not very good. When Larsson saw the ice with them, I imagine issues of positioning and speed—problems that generally see a player exit the NHL like many of the players in question did—led to odd man rushes and other high danger scoring chances for the opposition, resulting in the lopsided GA. As for the lack of GF to balance things out, this could be attributed to the same problem: bad NHL players don't score a lot.

We might never know what Yost was seeing when he created his chart (unless he was using a WOWY table), but the table above suggests to me that the bizarre dip in GF% can probably be attributed to exceptionally poor Quality of Teammates for our man Adam Larsson when he saw an increase in Offensive Zone Starts.