Using puck-possession tendencies and success-rates from individual offensive-zone puck-possession plays
This post will focus on each Montreal Canadiens player's offensive-zone expected goal (EGV), and expected shot value (ESV) above replacement. O-zone EGV and ESV represents the value of plays originating from the offensive-zone. I am able to perform these calculation due to the data available through my data-generated player tracking system. In this system, every puck-possession events that takes place on the ice is tracked as either successful or unsuccessful.
As described here, every possession play that occurs on the ice has both an expected goal value (EGV) and an expected shot value (ESV). The goal value is represented by the chance of that specific puck-possession play eventually producing a goal. The shot value is the chance of that specific play eventually producing a shot on net.
Expected goal and shot value can be calculated by using the success-rate of that specific play (cycle pass/pass to the slot/open-ice deke, etc.), as well as the success-rates of all succeeding plays. The goal of this is to determine how successful that play has been to either produce a shot-on-net, or a goal.
Conditional Probability
For example, we know that for a pass to the slot to result in a shot, the player passing the puck to the slot has to be able to succeed with his pass. Also, the player receiving the pass has to be able to get his shot through to the net.
Using 100 attempts (to simplify the explanation), if David Desharnais is successful with 39% (tops among forwards) of his attempted passes to the slot, we know that the player receiving the pass will have 39 shot attempts as a result of those 100 passes (100*0.39=39).
If the player taking the shot is Max Pacioretty, and he has been able to get 53% of his attempted shots through to the net, we can then calculate the expected shot value of that initial pass by calculating how many of those 38 attempted shots would result in a shot-on-net (0.53*0.39) *100)=20.67). This tells us that a pass to the slot from Desharnias to the Pacioretty has an expected shot value of 20.67; as it has succeeded in the past at a rate of 20.67%.
The expected goal value would be calculated using Desharnais pass to the slot success-rate combined with Pacioretty's true-shooting-percentage (percentage of attempted shots that result in a goal). For example, Pacioretty's TSP was 6.6% (0.066*0.39)*100=2.57). So a pass to the slot from Desharnais to Pacioretty has an EGV of 2.57, as that play was successful in the past at a rate of 2.57%.
This system will allow us to potentially determine the goal or shot value of any play on the ice. The key is to include every potential succeeding play as well.
For the sake of simplicity, I used the average success-rates of Montreal Canadiens players to calculate all succeeding plays. In other words, the average success-rate of Habs forwards within each succeeding play was used in the calculation. That said, the success-rate of any combination of players can be calculated using this model. The success-rates used in this post where determined using the data from every Montreal Canadiens 2013-14 regular season and 2014 playoff game. A list of all the events I track can be found here.
As with any model based on success-rates, sample size has a direct impact on the results. Validating standard deviation of the data used here has shown that once a player hits 150 even-strength minutes played, the season-over-season (4 years) deviation has a consistent SD of 4.2 to 4.5 for grades, 0.43 to 0.45 for ratios, 0.42 to 0.46 for events per-minute, and 0.45 to 0.47 for ratings. In other words, there is a year-to-year consistency in the data I'm tracking.
This post will also incorporate player tendencies into those values. Tendencies are described as the rate at which a player chose a specific play over another option. For example, a player in the offensive-zone has 7 options (that I tracked last season). New Metrics will be available in 2014-15.
- He can attempt a shot on net
- He can attempt an east/west pass across the offensive-zone
- He can attempt a cycle pass along the boards
- He can attempt to beat an opposing player 1on1 (open-ice deke)
- He can attempt to beat an opposing player 1on1 along the boards (wall deke)
- He can make a pass off of the rush
- He can attempt to pass the puck to a teammate who is positioned in the slot
Out of 100 possession-plays (to make the numbers easier to understand); if a player attempted 18 shots than his tendency to shoot is 18%. If he attempted 4 passes to the slot out of 100 possession-plays than his tendency to pass to the slot is 4%.
We can incorporate player tendencies into each plays expected goal/shot value simply by multiplying each play's expected value (EGV/EPV) by the player's tendency to attempt that specific play; in the example given the tendency for a shot would be 0.18 or 18%. The tendency for a pass to the slot would be 0.04 or 4%.
The resulting number for each possible event gives us (based on past success-rates) a real-time indication of the likelihood of each of the 7 possible puck-possession plays resulting in either a shot on net or a goal. Think of the result as a map to the most successful or unsuccessful route.
EXPECTED SHOT VALUE ABOVE-REPLACEMENT (OFFENSIVE-ZONE)
The higher the ESVAR, the more successful that player has been at helping to produce a shot on net from a play that originates in the offensive-zone. The use of tendencies within the calculation substantially impacts the numbers. If a player has a tendency to attempt plays that he has a low success-rate in, than his ESV will be lower than a player who has a tendency to attempt a play he is successful with more often.
Each Habs forward's offensive-zone puck-possession tendencies can be found here.
Among wingers, Max Pacioretty, Michael Bournival, Daniel Briere, and Brian Gionta had the highest o-zone ESVAR. Each of these player's high ESVAR is a product of of both a high tendency to attempt a shot on net while in possession of the puck in the offensive-zone, as well as a relatively high success-rate getting their attempted shots through to the net. Expressed more clearly, 18.3% of the time Pacioretty attempted an o-zone play with possession he successfully shot the puck on net. Briere accomplished the same thing at a rate of 17.7%, Bournival 17.1%, and Gionta 17.7%. It's important to remember that a high ESVAR represents a player's ability to produce a shot on net, and does not necessarily reflect scoring-chances, or points.
At the opposite end of the spectrum, Weise successfully shot the puck on net only 10.8% of the time he had possession, while Vanek successfully shot the puck on net 13.5% of the time he had possession. Weise made a successful cycle pass (which has a lower eventual ESV) 30.3% of the time he had possession of the puck in the o-zone.
Among centres, Ryan Whites he ESVAR is also the product of a high tendency to attempt a shot on net, while Lars Eller's lower ESVAR is the result of a tendency to attempt and complete high-risk open-ice dekes at nearly the same rate (12.4%) as he completed a successful shot on net (14.5%).
Among defensemen, it's not surprising that PK Subban, and Andrei Markov had the top expected shot values above replacement. An interesting note is that at even-strength Subban completed a successful open-ice deke in the o-zone (18.8%) more often than he was able to get a shot on net (14.8). Markov, on the other had successfully beat an opposing player 1on1 (deke) in the offensive-zone only 4.5% of the time he made a possession-play; while getting a shot through to the net at a rate of 18.7%.
EXPECTED GOAL VALUE ABOVE-REPLACEMENT (OFFENSIVE-ZONE)
EGVAR numbers get slightly noisy when we look at totals among defensemen, as well as forwards with smaller sample sizes. Defensemen EGVAR's are particularly difficult to calculate among defensive defencemen, as the number of offensive-zone puck-possession plays among these players are particularly limited. Players such as Mike Weaver and Francis Bouillon for example, have possession of the puck in the offensive-zone only 10.3 and 12.2 percent of the time they are on the ice; respectively.
Statistical anomalies in high true-shooting-percentages (goals per-attempted shots) also create produce noise. This impact can be seen in the EGVAR's of Dale Weise, Mike Weaver, and Francis Bouillon; as each of the 3 had TSP's well above average among players playing their position. Weise's TSP of 13.3% was well-above the average TSP of 4.8% among Habs wingers.
The only real surprise among wingers is the low EGVAR produced by Brendan Gallagher. Once again, Gallagher is the victim of his low TSP, as he scored with only 3.1% of his attempted shots; Pacioretty on the other had scored on 6.6% of his attempted shots, while Desharnais had a TSP of 7.8%. Desharnais was also helped by the fact that he had the top success-rate among forwards when attempting a pass to a teammate positioned in the slot.
Gallagher, like Michael Bournival also attempted a higher percentage of wall-dekes than other wingers. Wall dekes have low EGV's and ESV's as two other plays (pass to the slot/shot on net) are generally required to produce anything resembling a scoring-chance.
Among defensemen, Markov's ESVAR was significantly higher than Subban's. Yet, their EGVAR's are much closer. This is because, while Markov had a higher success-rate getting shots through to the net, both defensemen have nearly identical TSP's.
While this post focused on expected goal values and expected shot values above replacement using plays originating in the offensive-zone. Similar results exist for plays originating from anywhere on the ice.
