The Numbers Game: Carl Soderberg - Stats, Lies and Videotape

Alright team, take a knee. I may have been a little hard on you lately. We've been cruising with a pretty good record for the first half of the season, but some underlying numbers had been pointing to a little complacency creeping in. We've had a Corsi bag skate with Chris Bourque, done suicides with Shawn Thornton because of his meager performance against woeful QoC, and I've even called out the Captain. If I stay too much of a hardass, you're just going to tune out my fancystat systems, so today we're going to take a practice off of grueling statistical hammering of the current team performance and just have a little bit of a scrimmage.

Lets turn our attention to scouting. No, I'm not asking you to hop a plane, train and automobile to some god forsaken rink miles outside of Chicoutimi. It's awfully hard for fans to catch enough of a guy in other leagues to get a sense for how he might contribute to the club, so most of us are going off second hand info from people with better eyes than us. But what the heck are we doing just trusting eyes?! Why I bet there's information to be divined from data, by golly!

Player Traffic

Now why are we scouting outside of the NHL mid-season? Shouldn't we be looking for talent to trade for and playing Moneypuck to find cheap, undervalued guys in our own league? Not if you've got talent right under your nose and happen to possess its rights. As Cornelius mentioned in this morning's skate, Carl Soderberg of the Swedish Elite League looks to be on his way to Boston after his season wraps up. Originally drafted by the St. Louis Blues in the second round of 2004, Soderberg was exchanged for former goalie-of-the-future Hannu Toivonen in 2007 shortly after another netminding Finn arrived in the system (there can be only one). The big, left shooting forward has spent several years in the Swedish leagues putting up solid numbers, this year exceeding a point per game, but has up to this point been unwilling to come Stateside until a roster spot in the NHL opened up. In all the years we've held his rights, pretty much the entire Chiarelli era, Soderberg has never so much as attended a Bruins training camp. Well, with Bourque taking a one way bus to Providence that spot he's been mandating just materialized. But what can we make of that foreign production? Can it tell us anything about his expected output for the Bruins?

NHLe: The Informant

Gabe Dejardins, guru of the repeatedly referred to, invariably linked and totally indispensable Behindthenet.ca, created a metric called NHLe - NHL equivalency, simply enough - that attempts to measure how output in the many feeder leagues into the NHL would translate in the big show. This is done by adjusting the point-per-game production by a weighted "league difficulty number," derived from averaged historical data of players' NHL performance who came out of that league. Basically he's taken, with as many years of player data as possible, the points per game of all players in their NHL rookie years, divided by their immediate prior year in the other league and averaged the whole shebang to give a sense of how, relatively speaking, easy scoring is to come by in that lower league. So out of all the players who came out of the AHL over the years, they wound up averaging 44% fewer points in the NHL than they did in the minor league, so the league's difficulty rating comes in at .44.

To get an easily legible projection of NHL production for a given player, you then take this and multiply it by their AHL points-per-game and then normalize for a full NHL season length by multiplying by 82. Voila, you have a guestimate of a player's NHL scoring potential. So in recap, that's League PPG * League Difficulty # * 82 = NHLe. MATH!

Below is a list of league difficulty ratings to give you a sense of how the leagues compare to the NHL.

League	NHLe Rating
NHL	1.0
Russian Super League*	.83
Swedish Elite League	.78
KHL	.65
Czech Extraliga	.74
Finnish SM-lliga	.54
AHL	.44
NCAA	.33
OHL	.30
QMJHL/WHL	.28

As you can see, there's a few Euro leagues ahead of the AHL in terms of its relation to NHL competition. In active leagues, Sweden presently offers the closest approximation of NHL caliber play based on the performance of players who have been imported from that league. Perhaps Soderberg was right to stay abroad, he's come up against much tougher competition than the North American minors.

*Of (side) note, the former Russian Super League seems to be far closer to parity with the NHL than the current KHL. This may only be because of insufficient data, since fewer players have come over from the KHL to date given its short lifespan. Also, Russian players are increasingly likely in recent years to come up in North American Major Junior, rather than Russian development leagues.

Side Effects

As always, I have to slap on my warning label to caution you about the side effects of the week's stat du semaine, and this time there are many. First and foremost, we're looking at an averaged figure that includes guys who broke out immediately alongside guys who could barely hack it and went right back down. By nature, some players are going to outperform and some aren't, that's how we got the rating figure in the first place. Secondly, we're looking only at raw scoring stats. There's been no adjustment or weighting for ice time, competition or special teams output, the last of which will surely drop when a player is plugged into an NHL team and doesn't crack the top PP unit like he did in Junior. Since it's just scoring we're looking at, it's not going to be terribly useful in evaluating most defensemen unless their primary asset is their scoring ability.

There are some built-in assumptions about players' relative age within their leagues. The OHL number, for instance, is comprised only of players below age 19, so we're not seeing the uber-dominant over-agers included. However, age factors in another way, in that players reach their development peak fairly young, typically around 23. If they've not cracked the bigs by say age 27, their production in the lower league may be misleading in that they're less likely to see growth between their last low-league year and rookie NHL season. That growth assumption is implicitly baked into the NHLe number, if reductively, because of each component player's progression.

Another issue is that the league difficulty number itself is only as good as the amount of data feeding into it. Which is to say, how many players have entered the NHL directly from that league. With European leagues, we have fewer players on which to gauge how easy scoring is to come by in those leagues. As a result, our number for Carl Soderberg is going to be less predictive than a number for Ryan Spooner coming out of the OHL and playing in the AHL, since those leagues produce more NHLers. As you see above, the SEL alleges to offer the closest thing to NHL quality, but since they've not produced a mountain of NHL players, the figure is less trustworthy than other leagues.

NHLe is meant to give a ballpark estimate based on the relative competition level in other leagues, but it's constructed of a basic, low-event counting stats due largely to the dearth of data recorded outside the NHL. Like I said, we're scrimmaging this week, we talkin bout practice. NHLe's not the be all, end all - scouts aren't exactly out of jobs.

Bruins Findings: Out of Sight or Schizo(polis)

So, now that I'm done yammering, let's go to the board. Taking a look at a few other Bruins, maybe we can get a glimpse of how predictive this metric can be:

Player	Feeder Pts	Feeder GP	Feeder PPG	NHL Pts	NHL GP	NHL PPG	NHLe	NHL Pts/ 82
Krejci	74	69	1.07	27	56	.48	38.7	39.5
Lucic	68	70	.97	27	77	.35	23.9	28.7
Horton	68	54	1.26	22	55	.40	30.9	32.8
Marchand	32	34	.94	41	77	.53	34.0	43.7
Bergeron	73	70	1.04	39	71	.55	23.9	45.0
Seguin	106	63	1.68	22	74	.30	38.6	24.4
Peverley	54	45	1.2	44	66	.67	43.3	54.7
Kelly	60	77	.78	30	82	.36	28.1	30
Paille	21	29	.72	11	29	.38	26.1	31.0
Campbell	28	70	.4	9	64	.14	14.4	11.5
Bourque*	64	73	.88	8	51	.16	31.6	12.8

For a couple guys, it's dead on. David Krejci's 74 in 69 AHL season projected him for 38 points. What he got, adjusted for 82 games was 39. Lucic's early career was also pretty close to his WHL projection, missing by 4 whole points (24 and 28 respectively). Nathan Horton? His OHL projection pretty much nailed his rookie pace, ditto Chris Kelly's AHL rate.

Brad Marchand on the other hand proved much better than this metric predicted, putting him a full ten points ahead of the curve. Bergeron's QMJHL projection that was way off, placing him at a 24 point rookie. So much for that. Seguin projected higher than his actual production, but one should keep in mind his light workload in his first year under Claude.

* Because of Bourque's yo-yo career, I took his career AHL numbers, rather than single season, and applied them to the formula, resulting in 31pts NHLe. His career NHL tallies over 82: 12pts. Given the array of ages included in the data, it wasn't entirely likely we could read much from his projections. For what it's worth, while his last AHL year before his NHL debut had him at a 31pt player as well, his NCAA projection fell closer to reality, projecting an 18 point guy.

Carl Soderberg: The Good German Swede

Now that we've seen NHLe in action on our own roster and have some idea of where the projections are landing, lets take a gander at our incoming Swede.

Player	SEL Pts	SEL GP	SEL PPG	NHLe
Carl Soderberg 2012/13	60	54	1.11	71.1
Carl Soderberg Career	132	158	.84	53.4

Soderberg, just looking at his most recent season at age 27, has 60 points in 54 games. With the difficulty rating of his league at .78, he's projected to score a whopping 71 points over the course of an 82 game season. If we take his SEL career numbers, that only drops him to 53 points. The SEL has a tendency to break rookies in very slowly, giving them minimal ice time in early seasons, so Soderberg's first two seasons skews this career number down by 10 points. All told though, the projections place him somewhere in the realm of a second line scoring forward. Such output, even mitigated by lesser ice time, would be an absolute boon to this team.

However, while I'm not going to tell you to knock off the salivating over these figures, it's important to keep in mind that caveat about age. This is a player who has long passed his developmental peak and appears to be at his prime, but as a result he's likely performing as an outlier to the SEL difficulty rating data. Without a look at the age ranges of the SEL players who've come over to do time in the NHL, I can't provide any sense of how close this is to the norm in terms of translating production at his age. Most drafted players come over earlier in their careers, so I have to assume a lot of the data that feeds into the SEL's rating is for players with growth yet to come.

While age presents some accuracy concerns, this player has the scoring touch, a big NHL body, and NHLe points to him being a productive player given the tough competition in his league. He'll likely have some struggles with adapting to the speed and space of the North American game, but he's got the toolkit and has encouraging projections. The assumptions above are probably high, but we know we're getting a player better suited to putting the puck in the net than the one we just sent away.

Oh wait, and did I mention there was videotape?:

Erin Brokovich

With a sense of Soderberg, lets take a quick note of other coming reinforcements, both in the AHL and the OHL and see how they stack up in comparison.

Player	Feeder Pts	Feeder GP	Feeder PPG	NHLe
Spooner	40	44	.91	32.8
Camper	35	41	.85	30.8
Khokhlachev	42	25	1.68	41.3
Griffith	82	54	1.52	37.4

For immediate help, both Spooner and Camper project to provide reasonable third line production, though the size of both players would likely have a negative impact on their adjustment to the game. Same goes for the OHL boys, though Khoklachev clearly shows more offensive upside than the rest of the gang.

It bears noting that the above figures could be underestimating our AHL kids. Based on a recent article by Eric T. of Broad Street Hockey and NHL Numbers, the AHL competition got tougher thanks to the lockout year, with several NHL-caliber players slumming it. Based on data from the prior lockout, he projected the AHL difficulty rating to somewhere around .55. Assuming their points were produced in a tougher league for part of this season, Spooner projects exactly in line with Khokhlachev, and Camper in parity with Griffith, oddly enough.

Presently, Soderberg presents a better option than any of the young guns in our system. We'll have to wait a bit - the first round of the SEL playoffs doesn't start until the 12th and even in a sweep Soderberg's team won't be done until the 19th - but it looks like the wait will be worthwhile as he's the best available player we don't have to give anything up to acquire. Freebie! Thanks Hannu!