I’ve referenced the pass tracking project by Ryan Stimson and his crew and used the data from the 2014-15 season in previous posts, one using the passing data to analyze Brian Campbell and another for Brad Boyes.
This past week, Ryan began unveiling some of the current season data using the Toronto Maple Leafs to illustrate the findings in a series of posts on Hockey-Graphs.com in anticipation of the first big data release of the 2015-16 year.
Editor's Note: This is your chance to try FanDuel risk free! Sign up today and if you lose your first contest, you’ll be refunded the entry fee up to $10. Play now
There’s a particular passage in the opening piece of personal curious interest.
When we look at the Leafs forwards and their on-ice passing numbers, or passing Corsi, we are looking at the shot attempts that were preceded by a pass that a player was on the ice for and against. This is a refined measure of possession based on our evidence last season that teams shot at a higher percentage from passing shots than non-passing shots. Like everything in our data set, we’ll continue to test that going forward.
In the link post the first graph suggests completing at least one pass prior to a shooting event had a positive effect on shooting percentage. From that post:
Teams that were able to complete at least a single pass prior to shooting saw their shooting percentage rise from 7.2% to 8.0%; teams that completed multiple passes (at least two) prior to shooting saw that 8.0% rise to 9.6%. In other words, teams whose shots were preceded by multiple passes increased their likelihood of scoring a goal by 2.4%.
That’s the crux of the importance to the overall tracking work here. Shooting metrics provide a vast amount of data points and offshoots; the impacts and some context aren’t captured solely with shooting data via collection methods currently available, flattering the inclusion of supplementary data.
Tracking passing provides another solid layer of context. Further enhancements visualizing the data will only amplify their usage. As an example, Chris Boyle used ex-NHL goaltender’s Steve Valiquette’s concept of the Royal Road to illustrate the importance of passes through the slot area.
According to Valiquette’s findings, 22% of all goals are scored by a pass across Royal Road and below the faceoff circle.
Scoring goals is impacted more by forcing goaltender movement and passing than individual player skill and shooting location and offensive systems should be primarily sculpted for this purpose outside of forechecking. Forcing goaltenders to open up by moving, or taking advantage of openings created by the pass receiver and goaltender’s inability to get square in time are keys to scoring, and the best players can exploit this.
Check this series over at Hockey-Graphs prior to the big data release.
A spinoff of the Stats site Hockeyanalysis.com is David Johnson’s Puckalytics.com. Similar to War-on-Ice, Puckalytics offers a downloadable .csv file from the super Stats page with a variety of filters.
I used this feature to capture some on-ice player percentages in the overall team level of various categories, time on ice, goals for, shots for etc.
Hoffman in particular had great 5v5 scoring success in 2014-15, 22 to be exact, leading the Senators. In 2015-16 he’s already scored 17 goals, 10 at 5v5 and another four in slightly over 56 minutes of 5v4 power play time. He scored once on the power play in 2014-15 used sparingly.
At 5v5 though, when isolating Hoffman’s contributions, he’s had an immense impact, to compare his team level contributions year over year.
We can start with last season’s contributions. Hoffman was on the ice for 34.9% of the 5v5 goals scored by the Senators, third behind Mark Stone (36%) and Kyle Turris (35.1%). High shooting percentages are always a concern and 14.01% for Hoffman and 16.22% for Stone were tells about a possible slow down the following season.
This season, the trio has played 144:12 together according to the Puckalytics Super WOWY tool that could be used to measure variable linemates over a customizable period of time. This is a super tool for specifics on linemates – and opposition – and worth exploring. The site carries a gigantic database, so searches may be a bit slow.
The trio above recorded a Goals For percentage (GF%) of 60% with nine goals scored (and six against), firing a 12.5% on-ice shooting percentage and 56.3% Corsi For percentage (CF%).
Turris and Stone have played 96 minutes together in 2015-16 (50% GF% and 52.4% CF% firing 14.89%).
Hoffman with Turris have skated 333 together at 5v5 and share 30 goals between them. Stone with seven hasn’t found the same offensive spark like his 2014-15 rookie campaign.
Mike Hoffman’s 10 goals this season represent 50% of the on-ice goals scored by the Senators. Stone, Turris and a slight bounce back from Bobby Ryan represent the most significant contributions, while playing more than Hoffman. The trio – without Ryan – represents a great portion of the shots for while on the ice as indicated by the chart below.
Score effects can be eliminated by using 5v5 situations in games close (tied, up or down one goal in the first two periods, or tied in the third period).
Hoffman is running away with the on-ice goals, with 57.6%, with Turris and Stone and Ryan in tow.
In close games, the previous season’s trio of Hoffman, Turris and Stone carried the Senators' 5v5 goals in close score situations. Stone led with 41% of the Senators 5v5 close on-ice goals, with Turris at 37% and Hoffman only a slightly behind at 37.6%.
The individual shooting percentages are still slightly high, especially with Turris and Hoffman using a 20% shooting percentage as a fulcrum, Turris slightly over and Hoffman slightly under at 5v5 close – similarly to straight 5v5.
Tread carefully here where individually owned. A great first half could become a very pedestrian second half very quick.