Beyond my tracking of entries and exits which I discussed last week, I have also been trying to keep up with tracking the Bruins shot assists at 5-on-5. I did a similar project during the 2018-19 season on Stanley Cup of Chowder. However, this time, I am tracking the location of passes! This is all made possible by An Nguyen’s app Shot-Plotter.
Given some of the difficulties I have with blackout restrictions, and the fact I fell behind in tracking at the beginning of the season, this project has been on my backburner. But now I have caught up thru December 31, which gives me a sample of 1,233 unblocked shot attempts over the course of 1,679 minutes in 36 games. Ironically, as I write this, that is exactly half of the season so far. Game 73 on Sunday against the Hurricanes ruins that symmetry.
What is a shot assist?
While every shot that is taken in the NHL has a shooter recorded, the only time we know the player who passed the puck to that player is when that shot is a goal. Currently, Connor McDavid leads the league in primary assists at 54. That means he was the last player to touch the puck before his teammate scored a goal. Take his 395 unblocked shot attempts and 79 total assists (25 secondary assists) away from the 1,556 unblocked shot attempts he has been on the ice for, and that is 1,082 unblocked shots in which we have no insight into how McDavid contributed. That’s almost 70%! We have no record of the one-timers that he’s set up that his teammates have missed wide, or the cross-crease set-up where there was an amazing save by the opposing goaltender. This is why I set out to track this data - for the Bruins of course.
Personally, my definition for a shot assist is: An intentionally completed pass in the offensive zone which is followed by a shot within 3 seconds of the pass completion. This varies from some of the precedent already set in the hockey analytics community. Corey Sznajder has been tracking shot assists for years (for the whole league) and has a different definition than my own. I wanted my definition to more closely reflect basketball’s definition, which also narrows down the focus of the shot assist to its impact on the shot itself, as opposed to an impact on the shot’s creation. For what it's worth, most passes proceeding a shot fall under this definition, but a stretch pass leading to a breakaway gets marked as unassisted this way. Of the 1,233 unblocked shots in my dataset, 664 (54%) were proceeded by a shot assist.
Not only do shot assists add context to how players contribute to shots and goals, but they also have a statistical significance. In this sample, the Bruins scored on 6.92% of the unblocked shots that they attempted when proceeded by a shot assist, and only 6.33% when the unblocked shot attempt was unassisted.
Other variables
Beyond recording the location of the pass (start and finish), and the player who passed the puck, I additionally tracked three additional variables: whether or not the pass crossed the royal road, whether or not the shot was a one-timer, and how the shot was created.
Many know the “royal road” by now. It is a line the goes from the center of the goal crease and extends up to the top of the circle.
While this is a selective and subjective point of reference on the ice, shot attempts proceeded by a pass that crosses the royal road have a very high finishing rate. In many ways, this is a high-danger scoring chance-type of metric. Use it with a grain of salt, but it can provide valuable insight.
For one-timers, I much more replicated the traditional thinking of a slap shot where the shooter doesn’t stop the pass. As hockey evolves, and perhaps it’s already at that point, wrist shots and snaps shots with limited set-up and a quick release should count. I just think it’s a very gray area.
For how the shot was created, that was a bit complex. I came up with three categories for this:
*Rush - A shot that occurs less than 5 seconds following a controlled entry OR a shot that occurs from beyond the offensive blueline
Forecheck - A shot that occurs less than 5 seconds following a change in possession in the offensive zone OR a puck recovery from a clear 50/50 battle (i.e., faceoff)
Cycle - A shot that cannot be categorized as a rush shot OR a forecheck shot
Not included in this definition, but a shot would no longer be categorized as a rush shot if the puck carrier crossed the goal line. For instance, if a player was pushed wide but circled the net for a wraparound, I would categorize that as a cycle shot.
Preliminary findings
One of the first things that I looked at was team-level shots and shooting percentage by location. I was not sure what to expect, but shots from the cycle were converted at a lower rate than off of the rush or the forecheck.
In doing some research following this discovery, I stumbled across some amazing old writing from Jen Lute Costella. I think she describes this phenomenon quite well:
Creating rushes into the offensive zone, i.e. dynamic changes in possession as compared to offensive attacks which start with a controlled breakout or regroup) lead to confusion among the defense, less defending players in the zone and chances to get a dangerous shot on the goalie.
While good cycles tire out defenses, and you should certainly not exit the zone to try to create a rush chance instead, shots off of the rush or following a turnover are more advantageous in general given the defense will be less structured.
This brings us to Jake DeBrusk. Heading into the Winter Classic, the Edmonton-native was tied for second on the Bruins in 5-on-5 goals, alongside Charlie Coyle and Patrice Bergeron. Of regular Bruins at the time, DeBrusk had the highest percentage of unblocked shot attempts unassisted at 62.64%. While this likely limited DeBrusk’s shooting percentage, where he had a stable 8.79% finishing rate that ranked 6th on the Bruins at that time, it was offset by his form of attack being dramatically rush-based. 56.04% of his unblocked shot attempts came off the rush, significantly more than the next-highest regular Bruin at 47.37%. This matches the speed and style of Jake DeBrusk that most people reading this would align him with.
I also deeply regret not keeping up with this throughout the season, because I could’ve identified Pavel Zacha’s breakout in goal scoring. Heading into January, Pavel Zacha had just 2 goals on 59 unblocked shot attempts (3.39%). Since then, Zacha has 11 goals on 50 unblocked shot attempts (22.00%). While one could’ve foreseen Zacha regressing to the mean after a poor start to the season finishing wise, the amount of poor luck is even more pronounced when accounting for his teammates setting him up for shots. He had the lowest percentage of shot attempts that were unassisted among any regular Bruin, and the shot assists he was receiving were quite clearly dangerous - featuring 4 royal road and 8 one-time shot assists.
While Zacha’s distaste for shooting and relatively poor finishing ability can get frustrating at times, he was only going to get so much help from his teammates before he started to capitalize on it.
Shifting our focus to the back end, it’s been impressive how dynamic Charlie McAvoy has been in the offensive zone. Now that he is accompanied by Hampus Lindholm who is having a career season, the elite play of Charlie McAvoy isn’t as noticeable. McAvoy’s ability to create offense for others when activated is a step above the rest of the Bruins defensive core.
6 of his 32 shot assists crossed the royal road, and there were plenty of others sent to the low slot. Hampus Lindholm and Connor Clifton each had 2 royal road shot assists heading into the new year in 13 more games. It’s safe to assume that this trend has continued, as McAvoy ranks 3rd in the NHL in primary assists per 60 among regular defensemen this season. If you want to read more about how the Bruins structurally assist their defensemen in getting more active in the offensive zone, you can read may newsletter from January.
I am very excited to continue to work with this data throughout the next few months, as I continue to build out the dataset. If you want to make sure you see those articles, consider subscribing.