In an up and down season, Nashville Predators fans are hotly debating the team's potential based on results versus analytics. How can one fan base be so divided?
In an up and down season, Nashville Predators fans are hotly debating the team's potential based on results versus analytics. How can one fan base be so divided?
The relationship between analytics and the hockey "eye test" (watching a game and taking a final score at face value) is a complicated one. After the Predators' 7-1-1 recent run, there may be no fan base currently more divided along the results versus analytics lines than the one in Nashville. After a close win like the game over the New Jersey Devils or the November 23 shut out loss to the Detroit Red Wings, social media lights up with hot takes based on the eye test or charts, and rarely have the two sides agreed recently. Why do the two camps struggle to find middle ground?
Analytics have been an integral part of other major sports leagues for awhile now. The 2011 sports drama Moneyball, based on the 2002 Oakland A's MLB season, brought the relationship between sports and analytics to a wider audience. The film highlighted the value of statistical analysis to the world of sports. The NHL has been slower to embrace the role statistics can play within the hockey community, but it is making strides.
The slower growth of analytics in hockey may not be based as much on league resistance as it is due to the nature of the sport, according to Bryan Bastin, analytics aficionado and writer/editor at ontheforecheck.com. Bastin, who just returned home from SEAHAC, a hockey analytics conference hosted this past weekend by the Seattle Kraken, explained bluntly why that is.
"Hockey is chaotic," Bastin points out.
In football and baseball, there are distinct plays with stops and starts that make tracking metrics easier and eventually quantifying that data more straightforward. Hockey is a fast game with quick turnovers and transitions. Even with all the strides in hockey analytics, there are still many split seconds of potential game changing action that aren't being tracked.
SEAHAC 2022 brought together some of the best minds in hockey analytics to talk about the challenges and advances in this growing field. The beauty of the conference was its wide appeal. Everyone from NHL GMs to franchise analytics staffs to analysts leading the way in the field to curious fans was welcomed to Climate Pledge Arena. SEAHAC is for anyone and everyone with an interest in hockey analytics.
That interest is growing. The rise of websites like Natural Stat Trick, Evolving-Hockey, Hockey Viz, and plenty of others have opened a door at the fan level to the wide world of statistics and analytics. Fans, writers, and even broadcasters are leaning on these sites and many others like them to help tell a more complete story of any given hockey game.
And it is storytelling. While Bastin creates graphs and charts and quantifies data about the Nashville Predators, he knows his role goes deeper than the numbers.
"It is data driven storytelling," he shares. "I can't just spit out numbers. I have to tell a story."
It would seem like data - numbers and charts and graphs - would fall into the non-fiction genre. But as Bastin points out, what may seem to be true numerically is just one part of a hockey story. Like any good thriller, analytics storytelling has it plot twists, too.
The loss to the Detroit Red Wings is a great example. The Predators lost that game 3-0. Looking at a few key game stats tells one narrative about those sixty minutes. The Preds generated 14 high danger chances (there is a detailed definition that makes a chance "high danger", but in its simplest form these are chances that are most likely to result in a goal) while Detroit only had 6 high danger chances. Comparing another statistic from that game - Corsi (which tracks shot attempts and serves as a proxy for possession) - shows that the Predators controlled possession 60.92% of the game.
According to the statistics, Nashville was the better team for three periods. So why did they lose?
That is where the "eye test" comes in. Nashville lost that game mostly because they ran into a very hot goalie in Ville Husso, something those numbers don't necessarily reveal. The Preds generated more good offensive chances and possessed the puck more than Detroit, but Husso played lights out and earned the shutout.
Games like these - where the statistics don't match the eye test and the emotional highs and lows based on game results - is where the dissonance occurs. After almost every Preds game, social media lights up with those who are thrilled with a win and those who look at analytics to point out why the result doesn't mean what others think a win means.
The eye test/results fans say, "We beat the New York Islanders!" The analytics fans say, "Don't be fooled! The Predators weren't the better team!"
Can these two sides ever come together to truly evaluate this Nashville Predators team?
Bastin points out that effective storytelling with data can marry the two sides. The first step is acknowledging that both have their strengths and their weaknesses.
The strength of the eye test/final result is that ultimately the only stat that the league really cares about is wins and losses. Bad teams steal wins, good teams lose games. While there is a lot of action packed into a hockey game, fans can get a feel for puck possession and shot quality and intangibles like energy and momentum swings.
A great example of this is one Bastin heard from Micah Blake McCurdy of HockeyViz. Micah explains that expected goals (which measures shot quality) is just a numerical way to quantify crowd reaction. When the crowd at Bridgestone Arena stands up, leans in, and gasps loudly as Matt Duchene circles the net and takes a shot, that means they likely just saw a high danger shot.
The downside of the eye test/results based viewpoint is that there is so much action going on, and it is hard to see everything that happens. It is a little like skimming a page, especially when it comes to defense and goalies. That is where analytics come in handy.
Most people don't keep track of chance suppression, bounces, and defensive pairings on ice as they watch a game, but there is a statistic for that. Expected goals is a model-based metric that evaluates chance creation and chance suppression taking into account a number of key factors. To understand how well a defense plays, expected goals against gives a clearer picture than the eye test often can.
The numbers have their weaknesses, too. Bastin points out that while the field of hockey analytics is growing, it is still in its adolescent years. The best minds in the field are working on ways to track all of the nuance of how a goal was scored - screening, passing, angles. It is a hard job to collect all that information, but SEAHAC showed there are plenty of folks investing in improving the statistics available.
Analytics fill in the spots that the eye test may miss while the eye test can discern unmeasurable action. And that ultimately is where Bastin feels the two viewpoints can meet and perhaps work together to understand this Predators' season better. Evaluating the team's potential isn't either/or - it takes both.
Seattle GM Ron Francis knows that. Bryan tweeted what Francis had to say at SEAHAC last weekend about analytics and the eye test:
The Nashville Predators have gone 7-2-1 in their last ten games. Are they winning big games against teams like the Devils and Islanders? Yes. Are the Predators the better team in those big games they've won? Sometimes. Are there areas of concern beyond the wins? Of course.
Analytics and the eye test may always be a little at odds, especially when it comes to an up and down team like the Predators. But perhaps admitting the weaknesses of both sides and using the strengths of each is a good place to start.