Starting with the Philadelphia Flyers and recently adding the Minnesota Wild, we have learned quite a bit about the nature of zone entries as they relate to offense generated in the NHL. We now know that for both top- and bottom-scoring teams, uncontrolled entries (dump-in and tip-in) generate on average 0.3 Fenwick shots per entry, while controlled entries (carry-in and pass-in) generate 0.6 shots per. We have also observed that there is almost no difference in the shots generated by third line grinders and top line snipers across the different entry types.
However, while cracking into this area of analysis has produced some very remarkable findings, there are many more questions to be answered about zone entries and how they translate to offense. Here are some specific questions for the Wild that I hope to answer as I continue to research this part of the game, and preliminary ideas of how to tackle them:
Team Zone Entry Distribution Over Time
We know that the dump-and-chase game is less productive than the straightforward zone entry, and we have seen that the Flyers utilize the dump-in less than the Wild. But to what extent does a different coaching staff dictate the proportion of zone entries? With this past season being the first with Mike Yeo behind the bench, I would want to look at the 2010-11 season under Todd Richards to see if a mostly similar roster showed a similar controlled/uncontrolled ratio, and if the shot output was the same. For that matter, it would be interesting to see how closely Dany Heatley and Devin Setoguchi’s entry ratios compare from year-to-year, and whether and how they differed playing under different coaches.
Team Zone Entry Variance Within Season
To this point, we have looked at season-aggregated data and average shots per entry. Research using Flyers data has revealed that some things will take more than a season’s worth of data. But I want to know what the amount of variance there is within-season for number of zone entries and controlled/uncontrolled entry ratio, potentially giving us a glimpse at the strategy and game planning.
Zone Entry Ratio Between Opponents
With just one season of data, I am thinking about looking at just division opponents to see if there are patterns in how the Wild play teams like the Canucks or Flames, and if they differ significantly from the overall season data. I do worry about small sample size, with just three home and three road games per division opponent per year.
Big Lead/Big Trail Statistics
One of the main dismissals of advanced stats in hockey goes something like, “Corsi is stupid because when teams are winning the game they sit back and let the other team take more shots.” Researchers have answered this by examining close game situations, big leads and big trails, and I want to dive into zone entries the same way. When the Wild get up by two goals, to what extent do they shift their strategy to a dump-and-chase? This will be great to look across teams, because it will give us a better idea of coaching philosophies of the teams.
Within-Game Zone Entries
This is still a half-formed idea, but it seems to me from the eyeball test while watching the games that in the early minutes of a period or a game, both teams are much more likely to execute a dump-in. Score effects will likely trump time effects, but this may be something to look into.
Bottom Line: Not Enough Hours in the Day
Ultimately, the answers to these questions will come with more data: going forward we can continue to track games (if and when we ever get them back) and going back to previous seasons would give us a lot more context. We could see how teams change their strategy over time, and how players develop their skills as they progress through their careers. We could see if patterns emerge as to how teams play their division opponents, and if they are able to adapt their strategy (entry ratio) within a season. However, for now the only way this data can be collected is by a human being putting two eyeballs on a screen, and at 90-minutes a pop the data will be slow to come in. I am wondering how many games it would take from a previous season to get a representative sample. There is pretty much no way to get a computer to code zone entries, and going forward it could be done with a series of cameras in the arena, but that’s just not going to happen any time soon, though perhaps if they have success with it in the NBA it may transfer over. When I get my genius grant, maybe I’ll work on training rhesus monkeys to code zone entries, but for now, I’ll just have to keep using that elbow grease.
If you have additional ideas or research questions, please feel free to shoot them at me on Twitter, @Hashtag_Hockey or in the comments section below. Thanks for reading!
