2018/19 SHL Watchability Rankings
Back when I used to watch more baseball I’d frequently consult Fangraphs’ NERD scores to figure out which game was the best to watch. NERD was a metric developed by staff writer Carson Cistulli to determine how exciting or watchable each game would be on a given day, dependent upon on-field factors like the starting pitcher’s velocity, the lineup’s home run rate, defensive runs saved, and off-field factors like average age and payroll, along with many other metrics (full methodology here).
I’ve always thought it would be fun to recreate a version of NERD for hockey, so I went ahead a took a crack at ranking the watchability of all 14 teams in the Swedish Hockey League.
The Methodology
Cistulli’s methodology was based on calculating z-scores* across all relevant metrics, summing them along with a constant and voilà — each team now has a watchability score on a scale of 0–10.
*I couldn’t have told you what a z-score was before I did this, but in this case it represents how many standard deviations from the mean a team is on a given metric relative to the other 13 SHL teams.
For my purposes, I started by thinking of metrics to determine what makes a team exciting to watch. Here’s what I came up with:
On-Ice Factors
- Total goals per game
- Pace (shots for + shots against per game or 60 minutes)
- Goal differential per game (how close are their games?)
Off-Ice Factors
- Expected team quality
- Expected team quality relative to last year’s finish
- Starpower
- Acquisition quality
- NHL prospects
- Average age
To start out, I’m going to ignore all the on-ice factors. Obviously the games haven’t started yet, and there’s too much roster turnover to use last year’s numbers.
So everything right now is based on these off-ice factors, which need more explaining.
Expected team quality (xTmQual) was assessed by taking the average standings from five expert season previews (Uffe Bodin, MrMadhawk, Sanny Lindström, Mattias Ek, and Petter Carnbro). For example, Djurgården’s expected finish is a 2.6 based on the 3rd, 4th, 2nd, 2nd, and 2nd-place finish from the five experts above, respectively. Good teams are usually more fun to watch than bad teams, so lower is better here.
Expected team quality relative to last year’s finish (TmQualΔ) was assessed by taking each team’s actual place in the standings from last year and subtracting the metric above. For example, Djurgården finished 2nd last year, so their raw value for this metric would be -0.6. Teams on the rise are usually more fun than teams on the decline, so higher is better here. As a league newcomer without a 17/18 finish, I gave Timrå a slightly positive value here.
Starpower (Star) is measured by how many games of NHL, KHL, and International experience are on each team’s roster (with the NHL games weighted more heavily). This is imperfect in many ways. Players with NHL experience might be very little fun to watch (i.e. Alexander Urbom) and players with International experience might have racked up that experience with inferior hockey nations (i.e. Morten Madsen has played in the World Championships 13 years in a row for Denmark.) So it’s imperfect, but a quantitative attempt to measure starpower — if you have a better idea let me know! Stars are usually fun to watch, so higher is better here.
This is really the combination of two metrics. One for NHL GP (NHLGP) and another for KHL + International GP (KHL+IntGP).
Acquisition quality (AcqQual) refers to how exciting of an offseason the team had, and was assessed by how many slots a team filled on three “best offseason acquisition lists” from three experts (Uffe Bodin, Sanny Lindström, and Måns Karlsson). For example, Broc Little appeared on all three lists, so that gives Linköping three of the 35 available slots across the three lists. Newcomers are fun, so higher is better here.
NHL prospects (NHLPr) refers to how many drafted prospects are on each roster, with first-round draft picks being weighted more heavily. This benefits Luleå handsomely, as they are the only team with two first round picks on the roster (Isac Lundeström and Nils Lundkvist). There are flaws with metric as well, as this says Lucas Elvenes are Adam Ginning are equally exciting. Players who have been drafted, but went unsigned or no longer have their rights held by an NHL team do not count (i.e. Gustav Possler and Robin Kovacs). NHL prospects are fun, so higher is better here.
This is also really the combination of two metrics. One for first-round picks (Rd1) and another for all picks (DraftPr). Teams aren’t penalized for not having 1st-round picks.
Average age (Age) is straightforward. For purposes of this ranking, younger is better and teams aren’t penalized for having older rosters. Teams only gain from having younger rosters.
I applied a constant value of 5 so the range of watchability scores fit naturally on a 0–10 scale.
The final formula ended up looking like this:
(zxTmQual/2) + (zTmQualΔ/2) + (zNHLGP/2) +(zKHL+IntGP/3) + (zAcqQual/2) + (zRd1/2) + (zDrPr/3)+ (zAge/3) + 5
The Results
Here’s what that spits out:
Overall I think this turned out alright. Roughly speaking the most interesting teams are at the top half and the least interesting teams are in the bottom half. This is where I think it misses the mark the most, though:
- Rögle is too low. Sanny Lindström’s season preview for Rögle called them the most interesting team in the league this year, and it’s easy to see why. However they’re getting dinged here for still being a middle-of-the-road team (despite all that excitement), having just two drafted prospects on the roster (despite those two being among the best in the entire league), a lack of starpower (as measured by this model), and minimal representation on best offseason acquisitions lists.
- Färjestad is too low. In addition to being a serious Le Mat contender, most would consider Färjestad an interesting team to watch. But they get knocked down here for a lack of NHL/KHL/International experience and for a below average number of NHL drafted prospects.
- Linköping is too high. You can see their NHLGP z-score is 1.5, the most by a lot. Derek Roy and his 738 games of NHL experience hold way too much weight here.
- Timrå isn’t necessarily too low, but I was surprised by their negative score for NHL prospects. Jacob Olofsson and Filip Hållander are the only two drafted prospects on the roster, though.
The Future
My plan is to maintain these rankings throughout the season, adding and dropping metrics as either data becomes available or current metrics no longer matter.
Maybe by the first international break, I’ll add in the on-ice metrics currently being ignored, I’ll swap out preseason predicted standings for actual standings, and acquisition quality gets dropped. Maybe I’ll figure out a better way to measure starpower or account for prospect quality.
My hope is to post these scores every gameday to help you choose which matchup is most enticing in a slate full of games.
I’d love to hear any thoughts or feedback you have!
Thanks to eliteprospects.com and the experts at hockeysverige.se, hockeynews.se, and Expressen. I couldn’t have built this without your data and content.
