Boucher Scouting tracks puck possession events. The events tracked include puck-battles, loose-puck recoveries, passes, dekes, shots, blocked shots, blocked passes, dump-ins, and dump-outs. Each of these specific events are recorded as successful of unsuccessful.
Even-strength risk/reward is a calculation based on these events. It is used to determine how much more a player is helping his team by producing successful events than he is hurting his team by producing failed events. Simply put, risk/reward represents how many more successful plays than unsuccessful plays a player is making per-minute of ice-time.
Fenwick for % is the percentage of all Fenwick events that occur while a player is on the ice that are considered positive; a positive Fenwick event is defined as an attempted shot on net that is not blocked, while a negative Fenwick event is defined as an attempted shot on a player's own net while they are on the ice.
Corsi for % is the percentage of all Corsi events that occur while a player is on the ice that are considered positive; a positive Corsi event is defined as an attempted shot on net (including blocked), while a negative Corsi event is an attempted shot on a player's own net while they are on the ice.
For the purpose of this graph only, each player's risk/reward has been divided by 3.
*My thanks to stats.hockeyanalysis.com for the Corsi, Fenwick and zone-start numbers*
Among Montreal defensemen, PK Subban had the highest even-strength risk/reward rating, as well as the best FF%, and CF%. The lowest risk/reward rating among defensemen was produced by Alexei Emelin, who had the second-lowest FF%, and the third-lowest CF%.
Raphael Diaz had the second-best risk/reward rating, but both the lowest CF% and FF%. This exception can be explained by zone-starts. Diaz has the highest-percentage of defensive-zone starts among Montreal defensmen. Starting in the d-zone this often made it difficult for Diaz to produce as many positive Corsi or Fenwick events. To further the point, Diaz was involved in the second-most defensive-zone events per-minute played among d-men. Also, the drop-off in Corsi compared to Fenwick can be explained by the fact that Diaz blocked the second-most shots among Habs d-men; the same drop-off can be seen in Josh Gorges CF%.
Among forwards, Max Pacioretty had the best FF% and CF%, as well as the second-best risk/reward rating. Eller had the top risk/reward rating, but fell into the middle of the pack as far as FF% and CF%. Again, this can be explained by zone-starts. Pacioretty had the second-highest percentage of offensive-zone starts at 40.1%. Eller however, had an o-zone start percentage of only 31.3%. Conversely, Eller had a d-zone start percentage of 32.1%, while Pacioretty's was 22.5%.
Armstong is another example of zone-starts impacting his value. Armstong had the third-best risk/reward rating among Montreal forwards, but the fact that 40.2% of his zone-starts occurred in the defensive-zone hurt his CF% and FF%.
The opposite impact occured with Brendan Gallagher and David Desharnais. Both players' CF% and FF% were among the team-leaders, while their risk/reward ratings had them in the middle of the pack. Only 19.7% of Gallagher's zone-starts occurred in the defensive-zone, while the percentage for Desharnais is 23.3%; far-and-away the lowest among centres.
What separates risk/reward from Corsi or Fenwick is that risk/reward is produced by each player's own successes or failures; it does not rely on the play of teammates, or on zone starts. Corsi and Fenwick quantify the result, whereas risk/reward focuses on the process.
The puck-possession events that are used to calculate risk/reward rating can also be broken down to evaluate each player's ability within that specific event; for example a players' passing success-rate, or success-rate when attempting to remove puck-possession from the opposition by way of a stick or body check.
This is not an attempt to put one system above another. In fact, the use of both systems together are necessary for the process of finding the questions worth answering.
