Rush Shots – Why we care and who generates the most

Embed from Getty Images

Thanks to the great work being done by Emmanuel Perry at Corsica, we now have access to a variety of new pieces of data to help us better understand a hockey game.

One particularly interesting new data is Corsica’s rush shots. These shot types are defined as any shot taken within 4 uninterrupted seconds of a defensive zone event or any giveaway or takeaway.

Why do we care?

In some of her latest work, Jen LC (@Regressed PDO) summed the value of rush shots well:

Creating rushes into the offensive zone, i.e. dynamic changes in possession as compared to offensive attaches 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.

How important is this confusion? In a tracking project, Jen LC found that 57.3% of goals are scored within the first 7 seconds of zone time. I’ve used her data to make this chart:

 

Dashboard 1 (79)

 

Even further, LC’s data shows that ~46% of goals are scored within the first 5 seconds of zone time and ~38% happen within the first 4 seconds.

In short, goals happen most often early in the seconds immediately after a team rushes into the zone.

David Johnson (@hockeyanalysis) has done some research into this as well. Working under a somewhat different definition of a rush shot, Johnson found that the league average shooting percentage on rush shots is ~9.6%, more than two percent better than on non-rush shots (~7.3%).

So, rush shots create opportunities for more goals and we have the data to see which teams and players are best at generating these. Below, I’ll break down some of what we observed in the 2015-16 regular season.

 

Who generates rush shots?

At the team level, we find that rush shots for and against are very correlated – an r^2 of about 0.52. Here’s how teams fared this season:

 

Dashboard 1 (92)

The relationship for teams is evident. Teams that generate rush shots at a higher rate seem to surrender rush shots at a higher rate as well. This may be due to team systems or simply that a failed rush for a team results in a rush opportunity for the opposing team immediately after. One way or the other, rushes for and against are strongly related.

At the team level, some patterns are clear. This season, the Canadiens and Jets traded rushes with the greatest regularity – ~4 rush attempts for and against per 60 minutes at 5v5.

At the other end, the Devils approach to rushes is precisely what you might expect – they generate rush shots for, and allow rush shots against, at the lowest rates of any NHL team.

The Bruins and Sharks have the best ratios, managing to create rushes at a good rate without surrendering an equivalent number in return. The Avalanche and Rangers are the opposite – more rushes against than for.

One last thought at the team level – rush shots aren’t frequent events. The average NHL team takes ~2.5 rush shots per 60 minutes and allows the same in return. In an average game, this might work out to two rushes for and two rushes against at 5v5. Rush shots may carry a higher shooting percentage but they remain infrequent.


 

At the player level, some team effects are pretty clear. Here’s a look at the top rushing forwards who play regular minutes (>1000 minutes at 5v5 last season).

 

Dashboard 1 (93)

*click the image for a link to a larger view or, to toggle the filters for yourself, check here.

The Habs had the greatest rate of rushes for last year and that shows in the list of top forwards. Glachenyuk was the NHL’s best, followed immediately by Plekanec. The two averaged more than 4 rush shots per 60 minutes. Pacioretty makes an appearance near the top as well.

The Jets were the league’s second-most rush-heavy team last year. Wheeler, Scheifele, Stafford, and Ladd (majority of the year in WPG) are in the group of top rushing forwards as well.

A few interesting names pop up – Stone, Pageau, and Hoffman from the Senators, the Leafs’ newly-signed Kadri, Marchand (but not Bergeron), Silfverberg, and renowned speedster Jumbo Joe Thornton. The clumping of players from certain team suggests that team systems play a role here.

 

Dashboard 1 (94)

*click the image for a link to a larger view or, to toggle the filters for yourself, check here.

 

Again, we observe some team groupings. Andrei Markov owned the best rush shot rate among defensemen last season (>1000 minutes). Subban was essentially tied for the same rate. Emelin’s here too, which is a surprise. Byfuglien, Enstrom, and Trouba all make appearances in this top group. A Senators defensemen is here but not the one you might predict. And the excellent John Klingberg finished near the top.

As always, shout out to Paul Martin. Ever-underappreciated, Martin does so many things that analytics help to shine a light on. He’s here too.


Final thoughts for now

Thanks to the data at Corsica, it’s now possible to evaluate teams and players from a new angle. Is the generation of rush shots reliant on zone entries systems at the team level? Are certain players simply better at generating rush shots? Do coaches give specific players the freedom to rush at will while asking others not to do the same?

I’m not sure we can answer all of these, at least, not with data alone. But rush shots are better shots that lead to more goals, so parsing this new information will surely lead to clearer thinking on how teams and players generate offense.


Read more…

Playoff Predictions using Corsi Differential and xGF%

NHL Team Corsi Differentials – Atlantic Division

Advertisements

3 thoughts on “Rush Shots – Why we care and who generates the most

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s