Tracking Oilers Zone Exits

This is my first article for The Oilers Rig. Since recruitment to this fine blog, I’ve been working on a project tracking zone transitions over the Oilers’ blue line. By that I mean “zone exits” and “zone entries against”. Five games in, this post summarizes the initial data on zone exits. My next post will do the same on zone entries against.

*****

Ten seasons. It’s been 10 years since the Oilers have been in the playoffs. Fully 8 of those years have been spent entirely in the Oilers’ defensive zone. (Okay, not really, but it sure feels like it.)

It’s also been 10 years since the Oilers iced a Real NHL Defense™. Those two items are more than a little related.

Possession

Hockey is, at its core, a game of possession. Get the puck, keep the puck, score. If you give up the puck, it’s only with the intention of getting it back before the other guys score.

Hockey analytics has come a long way when it comes to measuring possession. NHL data (flawed though it is) still gives you data on shots, missed shots, blocks, shot locations, shifts, and a myriad of other things. Smarter men and women than I have used this data to come up with some impressive and intricate analyses of the game. What NHL data doesn’t tell you is how the puck moves when it’s not being shot. It doesn’t tell you how the puck gets out of your zone and into the offensive zone.

The “into the offensive zone” part of it is important, but somehow I’m more interested in the “getting it out of the defensive zone” part. Hall, McDavid, Eberle, Nugent-Hopkins… if you give those guys the puck enough times, they’ll score. They can’t do that spending entire shifts in the defensive zone. That, combined with the absurd lack of Real NHL Defense™, is why I’m interested in tracking zone exits.

Zone Exits

I first got interested in the idea when Jonathan Willis and Bruce McCurdy at Cult of Hockey tracked zone exits and entries for a year back in 2012/13 (sample here). The data they came up with was invaluable and eye-opening to me. Since then, I came across the work of Jen Lute Costella (@regressedpdo) when she was tracking zone transitions for the Chicago Blackhawsk in the playoffs (sample here). Now, I know there are many others doing similar projects around the interwebs, but these two are the ones I have drawn the most inspiration from when it comes to my current project.

Why is tracking and recording this stuff important? Hockey is a fast game. There are zone transitions multiple times a minute. The human brain can’t keep up with all that information. We tend to discard most of those events as meaningless, focusing and remembering the big play or the blatant give-away. Yet, these small events, easily overlooked, add up to dictate the overall flow of the play. Tracking them gives us the chance to shine a light on one small but extremely important aspect to game play. Knowing who gets the puck out and how is valuable information in judging a player’s impact on the game.

The “how” is essential. Getting the puck out is great. Getting out with possession… that’s the key. I’ll go back to Jen Lute Costella to quantify it. She notes that getting the puck out of the zone with possession (carrying or passing it out) leads to your team going on the offensive 88% of the time. Dumping the puck out leads to the other team going right back on the attack ~70% of time. (Source)

Think about that for a moment! That is an enormous swing in possession depending on an appropriate breakout. There’s a reason modern NHL tactics are moving away from “off the glass and out” and we’re seeing the rise of the “puck-moving defensemen”. This doesn’t mean that dumping it out is always the wrong play, but it does mean that whenever possible, you want to maintain possession.

My Project

For the last five games (Mar 14 – 22, 2016) I’ve tracked zone exits by the Oilers’ defense at even-strength. I’ve specifically noted “Controlled” versus “Uncontrolled” exits. A few limitations on the data:

• I’m only tracking 5×5 data, no special teams, no 4×4 and no data with the goalie pulled
• I’m only tracking defenseman

The reasons for those limitations ultimately come down to time constraints and trying to get the most valuable information for the time I invest. This is purely a hobby for me, so I’m finding time outside of my job, my marriage, and my parenting. I tried to focus on what I see as the Oilers’ main area of weakness: the Real NHL Defense™ thing. I intend to finish out this year. Next year, I’m going to go recruiting (see: asking nicely) to find some help to acquire a full season’s worth of data (plus playoffs… dare to dream!).

The Raw Data

First, the raw data. I’ll post this with each update somewhere near the bottom of the article for information in case others want to use the data. However, the raw data by itself isn’t all that useful, so I’ll break it down in just a moment.

• Carry: Defenseman carries the puck over the blue-line maintaining possession
• Pass: Defenseman passes the puck over the blue-line to another player, maintaining possession. Note: I allow some lee-way here. Sometimes a forward will receive the puck within a foot or two of the blue-line in flight, no one in the way and the puck is carried out of the zone immediately. I count that as a pass.
• Dump: Defenseman gets the puck out of the zone with no clear forward target in the neutral zone, leading to a loss of possession.
• Missed pass: Defenseman appears to try to pass the puck to a forward in the neutral zone, but the pass is incomplete, resulting in a loss of possession.
• Assist: A defenseman passes the puck to a forward who gets it out of the zone.

I’ve coded assists separately because there’s another factor at play… notably the forward who actually gets it out of the zone. If Sekera passes it to Korpikoski who flubs a pass leading to an uncontrolled exit, that’s not really Sekera’s fault, yet it’ll be coded as an “uncontrolled exit”. Likewise, if Sekera kicks it to McDavid, who dekes around two players in his own end before hauling it out of the zone… well Sekera didn’t do much there but he still gets credit for a “controlled exit”. However, on average, if the D-man gets it to the forward in a good spot, the forward is able to get it out with control. If the D-man gets it to the forward with a poor pass or in a poor spot, it often leads to an uncontrolled exit. So I’m analyzing the data both with and without the “assist”.

Controlled Exits

So let’s look at that data in terms “how” the puck is getting out of the zone, sorted by what percentage is controlled.

Oesterle is incredibly efficient at getting the puck out with control! Far and away the best Oilers defenseman over the last 5 games in doing so. Reinhart is also (surprisingly perhaps given his reputation) quite good at this as well. Further, when you look at their “total exits”, that pairing is sharing the load when it comes to zone exits compared to say, Sekera and Fayne, where Sekera takes a disproportionate amount of the zone exit work.

Now, it may seem like Sekera and Fayne don’t come out looking good in this graph, and that’s true, but there’s some context to consider here. Time on ice and competition.

Competition

In every game tracked, I took the top opposition even-strength point-producer and looked at what percentage each defenseman’s own TOI was against that one forward. For the games tracked, that includes James Neal, Vladimir Tarasenko, Daniel Sedin, Gabriel Landeskog (since Duchene & McKinnon were injured), and Martin Hanzal. These are “hard minutes” for defensemen… or at least the hardest minutes that opposition has to offer.

Sekera and Fayne clearly play against the hardest opposition the most. Meanwhile, Clendening and Nikitin (during his one game) were sheltered away from the top opposing forward. That’s important context to consider when seeing that Sekera and Fayne struggle at times in the efficiency of their zone exits. It’s not easy playing against Tarasenko!

Time On Ice

Okay, last piece of (extremely) important information. Obviously you’d expect a defenseman to get more zone exits the longer he’s on the ice.

This last chart is the zone exits as a rate of time on ice (per 60 minutes). It’s sorted by Total Controlled Exits / 60.

• Primary Controlled Exits / 60 = Passes and Carries per 60 minutes of 5×5 ice-time
• Total Controlled Exits / 60 = Above plus Assists per 60 min of 5×5 ice-time
• Primary Uncontrolled Exits / 60 = Dumps & Missed Passes per 60 min of 5×5 ice-time
• Total Uncontrolled Exits / 60 = Above plus Assists per 60 min of 5×5 ice-time

Oesterle and Sekera show their strength here. Oesterle leads the time in Total Controlled Exits as a function of ice-time and it’s not close. However, if you take away assists, Sekera edges into the lead (and doing so with tougher competition). Sekera, however, also gets the puck out via dumps and missed passes at a higher rate than anyone else, leading to more total exits.

Oesterle is amazingly efficient, actually having the least uncontrolled exits of anyone on the team. If Oesterle gets the puck out, it’s with control!

The Take Home

I’m going to dig into these numbers (and add to them) in future posts, especially once the year ends and we have about 10 or so games to analyze. However, for now, here’s a short bit on each player:

• Oesterle is a zone exit master. When he gets the puck out of the zone, it’s with control much more efficiently than any other Oilers defenseman… and it’s not close. Maybe that’s why McLellan is playing him over 25 minutes a night.
• Sekera plays tough competition and is the primary puck mover on his pairing with Fayne. That probably gives him a harder load to carry when it comes to zone exits, which may explain why he has so many uncontrolled exits. As a percentage, he doesn’t look strong, but when factoring in competition and time on ice, he’s not doing bad.
• Fayne definitely doesn’t take as much zone exit responsibility in his pairing and is clearly not as strong at it as Sekera. Of course, puck-moving was never Fayne’s calling card. (Spoiler alert: The zone entry against data should play more to Fayne’s strengths)
• Reinhart is surprisingly adept at puck-moving. He’s no Oesterle, but he holds his own on his pairing.
• Nurse looks good here, but what’s not reflected in the data is that he was quite poor in the first two games tracked and much better in the last two. Now, Colorado and Arizona played a more passive forecheck (especially when they got the lead) making zone exits easier, so perhaps that helped Nurse out. Perhaps I happened to catch two anomalous games, but I don’t know which two were the anomaly. I’m going to give it a few more games and see what comes out.
• Clendening has not shown an ability to get the puck out with control, which is particularly galling given that his reputation is as a puck-moving defenseman. He simply has not been good enough. A better puck-mover may help out Nurse and take some of the pressure off of him. McLellan clearly recognizes Clendening’s troubles as he is getting the least time on ice and is getting sheltered from the opponent’s top forward more than any other D.
• Nikitin only played one game so it’s hard to make any judgement here.

Next Steps

Zone Entries Against! I’ll hopefully have a post up about that some time this long weekend! Also, I’ll keep tracking the remaining games for the year to increase the sample size a bit.