Washington Capitals Coach Barry Trotz and General Manager Brian MacLellan stopped by the Washington Post newsroom Monday for a wide-ranging interview with reporters and editors about the upcoming season. One topic broached was the team’s use of analytics and what tools are used to evaluate players’ performances.
“I think you have to use everything, every resource,” Trotz said. “I think the stats are good. They’re not a bad thing. Stats are good for illumination.”
Two years ago, the Capitals hired Tim Barnes as an analytics consultant, bringing aboard one of the stat community’s early pioneers and most legendary innovators. Barnes ran the now-offline website timeonice.com, which compiled data for advanced statistics such as Corsi and Fenwick – both puck possession metrics – zone development and shifts. Trotz said he occasionally consults the data from some current websites that track those statistics.
Trotz said the team tracks its own chances for and against, but it also adjusts those for situation, like zone starts or quality of competition (if a center is matched-up against an opponent’s top forward line all game). The Capitals rate their players after every game, and those ratings are compiled based on film, statistics and if the player performed his role systematically and according to the team’s protocols. They then compile all of that to decide if a player had a good game, or was he not as good as they thought, even if he scored. That information is then presented to the player as a way to help him improve.
Trotz said the “eyeball test” can be revealing, using the example that a center might appear that he’s playing well defensively, but in reality, the top blue-line pairing on the ice with him is doing most of the defensive work.
MacLellan pointed out that some metrics can be subjective, as there can be human error involved when judging what’s considered a shot attempt, a hit and other statistics. As an example, the chances the staff watching from the press box record might differ from the ones the coaches on the bench have tracked, and both might not reflect what the official score sheet says. Trotz said he would “go on record” that there are certain teams in the league that pad their shots on goal.
“I think we take all of the stats with a little bit of a grain of salt because I don’t know that data is 100 percent accurate,” MacLellan said. “I mean, we can grab a stat sheet at the end of the game and go, ‘Those shots aren’t right,’ or ‘Those hits aren’t right. Those blocked shots aren’t right.’ I mean, it’s not accurate data. I think it’s important to get a good data source that’s measuring these things accurately throughout the league before you can even go to the stats numbers.”
Trotz used an example from Sunday’s preseason finale against the New York Islanders, when a player had a shot from the red line and it counted as a shot on goal. That would then impact both team’s players’ Corsi rating, which measures the number of shots directed at net while a player was on the ice at even strength.
“I looked at the board, and I’m like, ‘Seriously?’ ” Trotz said. “Analytically, the five [Capitals] guys on the ice got a minus because they did a great job and he threw it in from the red line. He basically had no where to go, and he tried to get in without icing, and it went on net and he got a shot on net.”
Trotz said he grades the quality of scoring chances, more concerned with the high-danger ones. MacLellan said the popularity of Corsi has potentially influenced some players.
“I think you’ve got players gaming the shot data now,” he said. “Just throwing pucks to the net. … I think it becomes ingrained. You run out of space, there’s nothing going on, why dump it in? Why wouldn’t you shoot it at the net? You get a shot attempt for your Corsi.”
“I think the point is, it’s not as easy as going to the website,” MacLellan added.