Author: max g

Up until the All-Star break, the NHL scoring race hasn’t gone entirely as expected. As usual, perennial contenders Sidney Crosby, Evgeni Malkin, and Patrick Kane, (I think that we can already include current leader Connor McDavid in this sector) are near the top. However, Cam Atkinson has powered the Blue Jackets by transforming himself into a scoring machine, and Brad Marchand has risen from simply an agitator to a top scorer. What’s in store for the second half of the season?

I decided on a project to predict the scoring leaders for the end of the season. Why? Not only could this be helpful and relevant for fantasy hockey, but it’s also interesting to see who could be holding the Art Ross trophy come June. First, a player’s Projected Goals (ProG) and Projected Assists (ProA) were calculated, culminating in Projected Points (ProP), which is the addition of ProG and ProA. Before getting into the methodology and results, there are a few things that I want to establish: This is not a list of the best players in the NHL. A player with a high number of Projected Points could be benefiting from playing with good line mates, a lot of power play time, or simply puck luck.

             If you wish, you can skip the explanatory paragraphs (which are rather mathematical and complicated) and just go to the later tables with results.

Let’s get started with the Projected Goals method. The initial data needed was a players’ games played, games remaining (assuming they don’t get injured), average time-on-ice per game and current goals. All of this data was taken from hockey-reference. Next, a player’s individual Corsi (shot attempts) For/60 (iCF/60) and career individual Corsi shooting percentage were taken from Corsica.Hockey. I used their career Corsi shooting percentage as an assumption that the player would regress back to how they’ve done throughout their career. Additionally, a players’s individual Royal Road shots per 60 minutes (iRR/60) was taken, courtesy of the Passing Project, spearheaded by Ryan Stimson. These are shots preceded by a pass across the royal road – a line across the center of the ice – which causes lateral movement by the goaltender and has a higher percentage of going in than an average shot (approximately 2-3 times more likely to result in a goal).

Let’s use Phil Kessel of the Pittsburgh Penguins as an example to understand how the Projected Goals are calculated. Currently, Kessel has 15 goals, 34 games remaining and plays an average of 17.53 minutes a night. He has 17.63 iCF/60 and 2.38 iRR/60, so roughly 13.5% of his shots follow a pass across the royal road. The shooting percentage for a royal road shot is about 15.5%, so the estimated royal road goals is 15.5% times the total number of royal road shots. His total royal road shots is 13.5% of his estimated total shot attempts, which is his iCF/60 multiplied by his estimated remaining ice time/60. The estimated remaining ice time is his average time on ice per game times the number of games remaining. This culminated in 3.66 royal road goals. The rest of the goals were calculated by multiplying his career Corsi shooting percentage (which was 6.09%) by his remaining iCF (excluding the ones that were royal road goals). This came to 10.54 goals, which added to his projected royal road and current goals, came to 29.20 Projected Goals for the season.

The top 15 skaters for Projected Goals are listed below:

(Note: Players with an asterisk use royal road data from the 2015-16 season, as the sample size for this season was too small)

The Projected Assists were similarly calculated. Here, the current assists were added to projected royal road assists and projected other assists in order to get a total. However, instead of iCF, a player’s on-ice (not individual) Corsi for per 60 was used. The on-ice Corsi for per 60 was averaged with the league average to account for potential regression during the rest of the season. Also, royal road shot assists were used instead of individual royal road shots, and a player’s total shot assists per 60 was also included. The total shot assists per 60 were divided by the on-ice Corsi for per 60 to get a percentage of total shot attempts that the player got an assist on. This was then multiplied by the current on-ice Corsi shooting percentage, or in the case of royal road assists, 15.5%.

Here are the leaders for Projected Assists:

Here are the top-50 leaders in Projected Points:

While the exact methodology and number may be imperfect (they tend to feel a little low to me), I think that this gives a reasonable idea of how a player will finish the season. Feel free to do your own analysis of the totals. It’s interesting to see the sheer dominance of Sidney Crosby.

All Data comes from Corsica.Hockey, Hockey-Reference or the Passing Project