What is going on with the NHL’s play-by-play data?
It’s something I’ve grown increasingly curious about over the past week or so, and a game on Wednesday night between the Toronto Maple Leafs and Montreal Canadiens offered a corroborating data point.
An astute observer noticed that the play-by-play data — the foundation for all historical counting measures and advanced measurements alike – was not matching the run of play. This data is collected for each game by individuals (or trackers) in home arenas as part of the NHL’s Real Time Scoring System.
In this particular case, Leafs centre Auston Matthews won a defensive zone draw, starting the transition up ice. Toronto defenceman Morgan Rielly attempted a shot in the offensive zone, and a full minute later after the faceoff was won, Canadiens goalie Carey Price froze the puck. But Rielly was never credited with a shot of any kind. Also of note: Montreal forward Nick Suzuki’s hit on Leafs defenceman T.J. Brodie was not recorded, nor was Suzuki’s earlier turnover.
Humans, not robots, are tracking the ultra-fast action in real-time. There is no expectation of perfection – in fact, the birth of hockey analytics was married to the reality of making adjustments to the data sets to improve quality. One obvious example? Combating home recording bias.
But there’s something unique about this season that makes me wonder if the issue has manifested in a bigger way. It wouldn’t be the first time that an issue was identified in this area. Just last season shot location data suddenly (and quietly) changed. When the broader stats community identified the issue, the NHL rallied to fix the problem and return shot location data to baseline.
Let’s take a look at the NHL this season compared to the beginning of the NHL’s Real Time Scoring System data, this time with a focus on offensive shot volume. (I’ve also added giveaway and taking data, notoriously less accurate than some of the other data we rely on, but I think it illustrates an important point).
I broke the data sets out by game state because I think there are peculiarities in the tracking when it comes to even strength play, though I’m not smart enough to figure out why.
If you look at the data sets, you see strong collinearity with increasing offensive production over the years – teams are taking more shots, they’re taking those shots from more dangerous areas, and they’re scoring more frequently. (It’s a big reason why league save percentages are down, year after year).
This season, teams are still scoring in line with prior years at even strength, and quite a bit more on the power play. But take a look at shot generation at even strength, which is down five per cent year-over-year and at a 12-year low. Consequently, shooting percentage is at a 12-year high.
That alone means little – it’s certainly possible teams have suddenly become more selective about where they generate these shots at even strength, trading quantity in for heaps of quality. But there are reasons to question that data. Among other things: teams appear to be wholly abandoning that strategy when it comes to the power play, generating shots in line with prior years and scoring goals at a considerably higher rate.
Other measures that have to carry a high level of accuracy, like penalties for and against, remain right in line with historical averages. And perhaps most notably, secondary measures – like even-strength takeaways – are down at such a considerable rate (25 per cent year-over-year) that they raise more eyebrows.
It’s not clear what the root cause of this is, or why there seems to be differentiation between even-strength data (which has deviated from recent years in competing directions) and power-play data (which is in line with recent years). But there has either been a considerable shift in offensive strategy around the league, or the work product of the trackers has changed.
In the grand scheme of things, this doesn’t appear a significant issue – there is still heaps of data being captured, and our understanding of what’s going on at the player and team level through these trackers remains strong.
But it does appear that we are losing data in some capacity. To that end, I invite my readers to send over any theory they have.
Data via NHL.com, Hockey Reference, Evolving Hockey
Vancouver Canucks winger Dakota Joshua is set to make his season debut Thursday after missing time for cancer treatment.
Head coach Rick Tocchet says Joshua will slot into the lineup Thursday when Vancouver (8-3-3) hosts the New York Islanders.
The 28-year-old from Dearborn, Mich., was diagnosed with testicular cancer this summer and underwent surgery in early September.
He spoke earlier this month about his recovery, saying it had been “very hard to go through” and that he was thankful for support from his friends, family, teammates and fans.
“That was a scary time but I am very thankful and just happy to be in this position still and be able to go out there and play,,” Joshua said following Thursday’s morning skate.
The cancer diagnosis followed a career season where Joshua contributed 18 goals and 14 assists across 63 regular-season games, then added four goals and four assists in the playoffs.
Now, he’s ready to focus on contributing again.
“I expect to be good, I don’t expect a grace period. I’ve been putting the work in so I expect to come out there and make an impact as soon as possible,” he said.
“I don’t know if it’s going to be perfect right from the get-go, but it’s about putting your best foot forward and working your way to a point of perfection.”
The six-foot-three, 206-pound Joshua signed a four-year, US$13-million contract extension at the end of June.
This report by The Canadian Press was first published Nov. 14, 2024.
TORONTO – Reigning PWHL MVP and scoring champ Natalie Spooner will miss the start of the regular season for the Toronto Sceptres, general manager Gina Kingsbury announced Tuesday on the first day of training camp.
The 33-year-old Spooner had knee surgery on her left anterior cruciate ligament (ACL) after she was checked into the boards by Minnesota’s Grace Zumwinkle in Game 3 of their best-of-five semifinal series on May 13.
She had a goal and an assist in three playoff games but did not finish the series. Toronto was up 2-1 in the semifinal at that time and eventually fell 3-2 in the series.
Spooner led the PWHL with 27 points in 24 games. Her 20 goals, including five game-winners, were nine more than the closest skater.
Kingsbury said there is no timeline, as the team wants the Toronto native at 100 per cent, but added that “she is doing really well” in her recovery.
The Sceptres open the PWHL season on Nov. 30 when they host the Boston Fleet.
This report by The Canadian Press was first published Nov. 12, 2024.
LAHORE, Pakistan (AP) — A top official of the Pakistan Cricket Board declined Friday to confirm media reports that India has decided against playing any games in host Pakistan during next year’s Champions Trophy.
“My view is if there’s any problems, they (India) should tell us in writing,” PCB chairman Mohsin Naqvi told reporters in Lahore. “I’ll share that with the media as well as with the government as soon as I get such a letter.”
Indian media reported Friday that the Board of Control for Cricket in India (BCCI) has communicated its concerns to all the Champions Trophy stakeholders, including the PCB, over the Feb. 19-March 9 tournament and would not play in arch-rival Pakistan.
The Times of India said that “Dubai is a strong candidate to host the fixtures involving the Men in Blue” for the 50-over tournament.
Such a solution would see Pakistan having to travel to a neutral venue to play India in a group match, with another potential meeting later in the tournament if both teams advanced from their group. The final is scheduled for March 9 in Pakistan with the specific venue not yet decided.
“Our stance is clear,” Naqvi said. “They need to give us in writing any objections they may have. Until now, no discussion of the hybrid model has happened, nor are we prepared to accept one.”
Political tensions have stopped bilateral cricket between the two nations since 2008 and they have competed in only multi-nation tournaments, including ICC World Cups.
“Cricket should be free of politics,” Naqvi said. “Any sport should not be entangled with politics. Our preparations for the Champions Trophy will continue unabated, and this will be a successful event.”
The PCB has already spent millions of dollars on the upgrade of stadiums in Karachi, Lahore and Rawalpindi which are due to host 15 Champions Trophy games. Naqvi hoped all the three stadiums will be ready over the next two months.
“Almost every country wants the Champions Trophy to be played here (in Pakistan),” Naqvi said. “I don’t think anyone should make this a political matter, and I don’t expect they will. I expect the tournament will be held at the home of the official hosts.”
Eight countries – Pakistan, India, Bangladesh, England, Australia, South Africa, New Zealand and Afghanistan – are due to compete in the tournament, the schedule of which is yet to be announced by the International Cricket Council.
“Normally the ICC announces the schedule of any major tournament 100 days before the event, and I hope they will announce it very soon,” Naqvi said.
