A few weeks back, I posed a question to the JR community about potential statistical analysis projects I could tackle in the coming season. I'm a longtime lurker, moderately statistically inclined, and interested in baseball sabermetrics - so I figured this might be a natural way to add to our collected knowledge about hockey.
Since then, I've been trying to settle on a project. And after some thought, I think I want to look at outlet pass effectiveness, since I think this is an underrated contributor to offensive and defensive success. The first pass out of the d-zone often doesn't show up on scoresheets as an assist, and poor outlet passes can lead to stalled rushes, offsides, icing and turnovers. A lot of team success rides on something that we don't know a lot about (at least that I can tell).
So, I'd like to look at outlet passers, outlet pass recipients, and the events that happen in the wake of each outlet pass the Caps attempt. More beyond the jump:
First, I'd define an outlet pass as a pass a) from behind the d-zone blueline b) that crosses the blueline and c) that is clearly intended for a player on the other side of the blueline. That means that clearing efforts, passes that stay in the team's d-zone, etc aren't included.
For each attempted outlet pass, I'd like to record a few things: first, the originator of the pass and his location on the ice, and second, the intended recipient of the pass and his location on the ice. I'd base "location on ice" on this 30-zone system described by Alz Well That Ends Well. Third, I'd note the time the pass was attempted; and then I'd like to record up to 4 subsequent events.
The "events" would be arranged in a sort of causation tree, with the idea that 4 "events" would be enough to reach a terminal event, one that ends the sequence of plays started by the outlet pass. So, for instance, the first layer of events would be simply "pass received" or "pass not received", and then I'd code subsequent events:
The idea is that an event in the leftmost column must result in one of the events in the column to the right, and so forth. For outlet passes that result in D or O-zone time, I'd record that, as well. The "causation tree" is a little rough, and I'd love edits/ideas on other events to record.
If this system works and doesn't prove to be terribly time-consuming, it would provide a wealth of data on outlet passing and the Caps' breakout strategies. I'm totally open to suggestions on improving, changing, or re-scoping things - let me know what you think!