Saturday night in Pittsburgh, Toronto Maple Leafs fans were treated to a rare sight. No, not just another convincing win against the Penguins, their second in as many weeks in a building that has not historically been friendly to the team, but an Auston Matthews five-on-five goal.

It was just his fourth on the season, in a little over 355 minutes. Per 60 minutes, Matthews has scored just 0.7 goals, his lowest personal rate in the first 23 games of any season in his career.

It might be easy to look at Matthews’ individual shooting stats and note that his rate of shots on goal per hour, expected goals, shot attempts, and scoring chances (via Natural Stat Trick) are all close to a career best, having established career highs in all of those individual categories last season:

Matthews’ six percent shooting percentage at 5v5 is less than half of his career shooting percentage of 15 percent, and closer to a third of his shooting percentage over the previous three seasons, at 16 percent. One could be forgiven for suggesting that Matthews will go back to normal if he keeps shooting, since eventually some of these shots will fall.

Having watched every Maple Leafs game (prior to Saturday’s contest against Pittsburgh), I am not as confident.

When I watch Leafs games, I’m also counting events that happen throughout the game, including every offensive zone entry attempt, defensive zone touch and shot attempt. The reason I track shot attempts is to provide some contextual data to each shot, whether for accurate time stamping (the NHL scorekeepers are quite bad at this), whether the shot was set up by a teammate, and whether a defensive player pressured the shooter.

The reason I track this is that while NHL teams have access to this data, whether through a tracking company such as Sportlogiq or through the league’s own player- and puck-tracking software, fans don’t have access to the same resources that NHL teams have, and I see it as an opportunity to track the data for myself and give fans a small peek at the sort of stuff a team would be able to work with.

Now, my shot tracking is relatively new, and with just six weeks’ worth of data, I don’t have a very accurate measure for expected goals based on the factors I track. However, I am able to categorize whether a shot is a scoring chance based on certain criteria, such as distance from the net, angle, rebound, pre-shot passing, defensive pressure, or whether the shot came off the rush (within six seconds of a zone entry) or after a turnover. Since I’m looking at every shot individually, I can take into account more factors than what’s by the official NHL data. This method has drawbacks, in that, I can’t compare data leaguewide, nor can I compare to previous seasons. It also takes a very long time to track.