This summer, the perpetually rebuilding Edmonton Oilers hired analytics writer and team critic Tyler Dellow. Almost halfway through the 2014-15 season now, the Oilers have once again been a disaster, in contention for the first overall pick yet again, firing their head coach yet again, and yet again trading players for draft picks. It doesn’t seem as though massive improvement is imminent. Those who oppose analytics are all too willing to point to this as an example of why the hiring was a mistake, or of why relying on the practice doesn’t work. But of course increasing one’s use of analytics can’t in one season overcome a poorly constructed roster, or any number of other problems. We can’t truly know whether or not Dellow has helped the Oilers because we don’t know what work he has done since joining the team on an official basis, or how much of it Dallas Eakins, Craig MacTavish, and most recently Todd Nelson have embraced. But we can look for clues; we can get a pretty good idea of whether at least some of what Dellow has championed in the past has been applied.
The most obvious change with the Oilers this season has been their improved possession numbers. Last season’s edition stumbled to a 43.6% Score-Adjusted Fenwick (28th), while this year has brought significant improvements. They are by no means dominant, but this season’s Oilers are at 47.7% (22nd). It’s not good, but it’s a pretty solid year-to-year improvement for a roster that remained largely the same. But an improvement in possession numbers alone isn’t enough to assure us that Dellow has had a significant impact. We need to look for something more Dellow-specific. Luckily, he gave us a specific example last spring of an aspect of the Oilers’ strategy that he felt was having a negative impact on their numbers. The posts are no longer available, but I can still summarize the gist of it from what I remember, and from some of the work that I’ve reproduced.
Dellow found that there are a certain number of seconds following faceoffs before the impact of those faceoffs is washed out. While there is still work to be done in determining the repeatability of these micro-statistics or their association with winning, they can still safely be used for descriptive purposes. Dellow last spring noticed that the Oilers had adjusted their forecheck following offensive zone faceoff losses. Around Christmas (or maybe the Olympic break, I can’t remember), the Oilers went from consistently seconding two men in deep following offensive zone faceoff losses, to only seconding one, still a common practice but a far more conservative one. He cited Bruce Boudreau’s Anaheim Ducks — consistently at the top of the league in terms of shot differentials following offensive zone faceoff losses — as an example of an aggressive team in those situations that is worthy of emulating. While the Ducks don’t always play this 2-1-2, here’s a good example of them doing so.
I took a look back at some games last season to double-check what Dellow hypothesized. Here’s some video from early in the 2013-14 season.
And here are a couple of instances of how that changed later in the year.
While I don’t have the same statistical breakdown that Dellow had in terms of seeing how that strategy impacted their numbers, here’s the breakdown of team Corsi in the 37 seconds following offensive zone faceoff losses last season, which I also used in previous articles here and here.
So it’s clear that overall the Oilers struggled in such situations; whatever changes Eakins made certainly didn’t help. The question then becomes whether Eakins, learning from Dellow’s analysis face-to-face, switched back to his prior tactics in those situations. If he did, then it’s fair to say that the decision-makers on the Oilers have in fact utilized Dellow, and that he wasn’t simply hired as propaganda. The following video features games from both the Eakins and Nelson eras; it’s a small sample, it’s fairly representative of the team’s tactics, according to the hundred or so draws I watched.
(clock to enlarge)
The Oilers don’t send two men following every offensive zone loss — it seems to be something of an option play based on where the puck ends up and the chances at retrieving it — but it’s safe to say that Dellow has encouraged the coaching staff to be more aggressive in these situations, and the staff has listened. How has this impacted results?It’s interesting to note that much of the team’s improvement has come in neutral zone situations. This may have to do with zone entries (both carrying the puck in more and preventing controlled entries), but I don’t have those numbers to be sure. Anyway, it is clear that the numbers following offensive zone faceoff losses have improved. We don’t know how much of that improvement is a result of this change in strategy, but considering what Dellow found last year, it might be a large portion of it. So congrats to the Oilers on being progressive, and to Dellow on being heard. Finding a way to tighten up defensively and finding better goaltending (internally or externally) are the next steps forward.