I've set out to answer a question for myself. What does a scoring chance tell us about a player, a line, a team? Why bother tracking them? What is a scoring chance really? (We're getting philosophical up in here!)
On the surface the answer seems simple. Its a high quality opportunity in the offensive zone, or in the defensive zone if you are giving one up to the opposition. But what information is really contained in a scoring chance? I think a quick film study should answer these questions sufficiently.
I've chosen the San Jose game, specifically the performance of the Schenn line.
Using Gamecenter and Microsoft Encoder I broke down each scoring chance registered for this line into 30 second videos for review.(I really needed an excuse to play with Encoder this weekend) We'll start with the offensive opportunities.
Flyers Chance 1
1. The sharks start with possession and attempt to move through the neutral zone.
2. Schenn intercepts and sends a quick up to Simmonds
3. Simmonds transitions smoothly and enters the Sharks zone with control
4. He passes off to Lecavalier who moves the puck cross ice unsuccesfully.
5. Gustafsson pinches in and tips the puck deep.
6. Schenn pressures the Sharks d-man and forces a turnover.
7. He passes to an open Simmonds in the circle, and Simmonds is able to get a dangerous shot off from inside the dots.
Almost every Flyer on the ice makes a positive contribution towards generating the chance.
Flyers Chance 2
1. Again, the sharks start with possession and enter the Flyers zone.
2. Schenn pressures Pavelski and forces a turn over to Simmonds
3. Simmonds moves up ice and enters the Sharks zone with control
4. Drop pass to Schenn, whose already sneaked a peek to see who is trailing the play.
5. Cross ice backhand saucer feed to the hash marks.
6. Gustafsson jumps into the play and gets a shot off but misses the net.
Flyers Chance 3
1. Sharks dump the puck into the Flyers zone.
2. Mason clears the puck back out and forces the Sharks to retreat.
3. Simmonds gets on the forecheck and forces a turnover. He passes to a streaking Schenn
4. Schenn gets a shot from the inside has marks.
Flyers Chance 4
1. Sharks have possession, Meszaros forces a neutral zone turnover.
2. Mez dumps the puck into the sharks zone
3. The sharks recover but Schenn steals the puck again the in the neutral zone
4. He enters the zone with control
5. Pulls up and waits for help, Lecavalier goes to the net
6. Schenn sends the puck to the net for a tip play by Lecavalier.
Sharks Chance 1
1. This time around, the Flyers start with possession
2. B.Schenn takes a weak shot which results in an offensive zone turn over.
3. Sharks transition and enter the Flyers zone with control
4. Nieto turns on the jets and goes wide on Streit
5. Nieto gets a chance on net from inside the dots
6. Rebound chance for Thorton in the the slot.
Sharks Chance 2
1. Recurring theme here, Flyers start with possession.
2. Dump in by Simmonds.
3. Corner scrum lost involving Schenn, Simmonds, and Lecavalier is lost.
4. Sharks transition quickly with a hail mary lob pass up ice.
5. Marleau beats Streit wide and gets a shot from the dot.
Sharks Chance 3
1. Lecavalier loses a draw in the Sharks zone.
2. Sharks break out cleanly and enter the Flyers zone with control.
3. Gus blocks a shot and breaks up the initial play
4. Lecavalier steals and starts a breakout to Simmonds
5. Simmonds passes to Schenn, who turns the puck over at the Flyers blue line. Uh oh!
6. Sharks just miss on a cross ice play.
7. Lecavalier loses a board battle and turns the puck over.
8. Stuart misses a shot wide from the dot.
Sharks Chance 4
1. Flyers have control, Lecavalier dumps the puck into the Sharks zone
2. Sharks immediately regain possession and start a break out.
3. Burns tips the puck into the Flyers zone.
4. Flyers gain possession back, but Coburn hesitates with the puck in the corner and turns it over.
5. Lecavalier misses on a clear out.
6. Sharks execute a tip play to Nieto. Schenn has Nieto covered but misses the puck.
Scoring Chances and Sample Size
Scoring chances reveal much more about a player or a line than you might have initially thought. The 8 events I've tracked above actually account for almost 50 individual plays on the ice. Every scoring chance that gets recorded for or against is actually an aggregate of an entire shifts worth of plays. 3, 4, or even all 5 guys on the ice can play a direct part in creating a scoring play. Or conversely, 5 different mistakes can add up into a chance against. This single point of data actually measures d-zone play, neutral zone transition, breakout efficiency, offensive creativity, and more.
The overall point I'm trying to make here is that this single statistical event captures a wealth of information about how this line, or the team as a whole is playing. This is what makes statistical scouting such a valuable tool. It gathers together in a neat, easy to interpret package 3 zones of play. It also illustrates why sample size is such an important factor in statistical evaluations.
Say you were to only evaluate talent by goals scored and points produced. There's about 2 or 3 goals scored in an NHL game on average. Based on the analysis above, you'd only be looking at 10 or 15 individual plays tops, and maybe a cumulative minute of actual ice time. If you're evaluating talent, is it really fair to cut out 59 minutes of the game? Goals are rare events in hockey, and looking at such a limited sample of ice time allows the influence luck to dominate the equation.
So we increase the sample size by looking at scoring chances. Between both teams you're now looking at 30 or 40 events and a cumulative 20ish minutes of ice time. This is a much more representative sample of a player or team's play than goals scored. Its not perfect, but we've minimized the role of luck and gotten a better picture of talent.
Lets say we increase the sample size even further to shot attempts. The same factors are at play here. Every shot attempt for or against has a entire shift's worth of action that leads up to the attempt. A single NHL game has something like 100 shot attempts between both teams. Now you're analyzing 45 to 50 minutes of the action per game and you get the truest account of a NHL player's or team's ability to affect how the game is played. We've sacrificed a bit of shot quality by increasing the sample size this far, but we've gathered a much larger pool of data in return. And we've minimized the impact of luck as best we can. Its no coincidence that Corsi is the dominant mode of evaluating player talent in the analytical community. Its also no coincidence that Zone entries and exits are the new frontier in hockey analytics, considering they account for 150+ events per game.
If you're a fancy stat skeptic, I hope this post has illustrated what a valuable tool of reference they can be. Fancy stats can account for a large portion of games and offer concrete evidence of what a player is or isn't doing on the ice. The average hockey fan doesn't have 2 or 3 hours to sit around breaking down every minute of film. Even professional hockey scouts cannot feasibly watch every game; played by every player, on every team, in every single professional hockey league. Of course they aren't perfect, but they capture the gist of the flow of play on the ice.
P.S. Let me know if you like the videos. They might be low quality but they were easy and quick to put together. I may start including clips in my post game chance summaries so you guys can see exactly what I see when I'm breaking down a game.