Anaheim DucksArizona CoyotesBoston BruinsBuffalo SabresCalgary FlamesCarolina HurricanesChicago BlackhawksColorado AvalancheColumbus Blue JacketsDallas StarsDetroit Red WingsEdmonton OilersFlorida PanthersLos Angeles KingsMinnesota WildMontréal CanadiensNashville PredatorsNew Jersey DevilsNew York IslandersNew York RangersOttawa SenatorsPhiladelphia FlyersPittsburgh PenguinsSt Louis BluesSan Jose SharksSeattle KrakenTampa Bay LightningToronto Maple LeafsVancouver CanucksVegas Golden KnightsWashington CapitalsWinnipeg Jets

Corsi & Fenwick Game-by-Game 2012-13

Link back to Data Visualization
Original Hockey Stats Visualization

Over the summer, working on the McKeen’s Hockey Yearbook, I have been tweeting images of players and teams game-by-game Corsi broken down by the individual components comprising the metric. Along with Corsi, I’ve presented some team cumulative Fenwick rates as the season progressed to a close.

An example for players Corsi is shown below via Coyotes defenseman Oliver Ekman-Larsson.

Pavel Datsyuk is on display here, highlighting his on-ice even strength goals for and goals against.


This is a team Fenwick Chart of the San Jose Sharks 2012-13 season.

This is the Blue Jackets Game-by-Game Corsi:


Housed in excel, capturing the images and sharing over twitter as a whole for all available player seemed like a daunting task (just ask Ryan Dadoun ( who was helping me do this), and until recently, another solution had not presented itself.

Tableau Software made some changes resulting in utilizing the aggregated data to create a proper, interactive data visualization entitling users with the ability to select any combination of input parameters.

The data for the visualizations was aggregated through two sources, the valuable and (Fenwick numbers captured in a timely interval.

Metrics available are for teams, individual players, even strength goals against and then the cumulative Fenwick by date throughout the season - starting from February 18, 2013.

These stats are created through actual NHL play-by-play game sheets and when parsed and calculated is a useful proxy for puck possession, with the logic that if the team/player did not have possession of the puck, these events would not have presented themselves and captured in game sheets.

At the top-left of the visualization appears a tab with a downward arrow. Click on this tab to produce a drop down menu that lists all the available pages and a credits tab. Users can also scroll through the tabs at the top of the ‘viz’ to load a particular page.

Before describing the individual tabs, a note on the makeup of Corsi, the dominant metric on these visuals. Volumes have already been written about Corsi, with Habs Eye on the Prize taking a deeper look at a variety of advanced stats in a feature dubbed ‘Fancy Stat Summer School. A step further, there’s a handy glossary of terms.

The makeup of Corsi has three underlying components, and is usually expressed as a rate stat normalized on a per 20 minutes basis. Comprising Corsi are the underlying components of shots on goal, missed shots and blocked shots directed towards the opponents net minus the same components fired towards the teams net. The differential is presented in the charts (that is the calculation is already done)

The viz highlights Corsi’s individual components differential on a game-by-game basis. Hover the cursor over each of the coloured components to reveal the calculated differential for that particular base stat for that game.
An important point to remember on these charts is score effects as defined in the glossary link above.

Score effects - the term given to the natural tendency of teams to stop pushing the offense when they have a lead, which allows a trailing team to outshoot their opponent. This phenomena occurs with every team in the NHL, some worse than others.

Data visualizations include overall team and players.
The other tab is a game-by-game visual of the even strength goals for/goals against while the player was on the ice.
The final tab is by date along the horizontal axis with a Fenwick chart. This is calculated similarly to Corsi, but omits blocked shots (with the logic that blocked shots would not have reached the net).

Fenwick is the differential of shots on goal and missed shots towards the opposition goal minus the same criteria at the team’s own goal.

The feature here is set as a default to the situation as ‘close’ meaning these values are calculated when teams are within one goal through first two periods, or tied in the third. Among the statistical community, Fenwick Close is one of the best metrics to assess team’s performance.

The ‘close’ situation effectively takes away the score effects mentioned above disregarding wide drifts in scores to skew the numbers.

The situation tab also provides different values for various game situations. Multiple conditions clear up a muddled bigger picture while parsing through game situations over the season.

The ‘sparkline’ shows this value relative to a playoff threshold. The value sits in some situations, (down one goal for instance), teams should be driving the play harder and should, in theory, present values that exceed the playoff threshold. Taking the Fenwick with teams up or down by two goals simulates some of the score effects inherent in the Corsi charts.

At the bottom of the visualizations, there are social media links as well as the ability to export images to be used in blog posts.

With the new season just around the corner, a lot of preview articles are being prepped. We hope this visualization gives some background to digest the shortened 2012-13 season.


McKeen's 2013-14 Yearbook ---- coming soon.