I wrote a post the other day in which I argued that there was good reason to think that the Oilers were more like an 85 point team last year than they were a 74 point team if you looked at their goal difference, frequency of OT games and success in OT games. There’s good reason to think that their record wasn’t really a fair reflection of their ability to play hockey.
As I mentioned in that post, that was sort of what I think of as the first level of the analysis, looking to see whether the goal difference supports the record. The next stage, for me anyway, is looking into whether or not the underlying numbers support the goal difference. Here are the Oilers’ 5v5 numbers for the past three years:
So the Oilers have gone from -53 to -20 at 5v5 over the past three years. This is positive. If you look at why that’s happened, it’s pretty clear: it’s got almost nothing to do with their shots ratio (which has barely moved), a little bit to do with a jump in their shooting percentage and a lot to do with improved puck stopping.
In 2009-10, the Oilers took 46.2% of the 5v5 shots. In 2011-12, they took 47.1%. If we assume, for the sake of discussion, that a shot has probability of 0.08 of going in, a team that plays 3850 5v5 minutes and takes 46.2% of the shots is going to get outshot 2071-1779 and outscored 166-143. Improving to 47.1% of the shots belong to your team narrows things slightly. You’ll get outshot 2037-1813 and outscored 163-145. From -23 to -18. So, it’s something but not huge.
The shooting percentage was up slightly but it’s well within the range of the reasonable. I’ve talked a lot about Jordan Eberle’s shooting and on-ice shooting percentage coming down but, team wide, some guys have good years, some guys have bad years. Given that the shooting percentage is pretty normal on a team wide level, there’s not much point in being excited about Eberle’s sky-high year at ES (except to the extent that it impacts on his situation in particular). Eberle will likely have fewer shots go in this year; other guys will likely have more shots go in. Not much to see here.
The big improvement for the Oilers last year over the preceding two years was in the save percentage. I wrote about this in January.
So at 5v5, this year’s team has a shockingly similar share of the shots to last year’s team. Why are they -7 instead of the more traditional -24 or so at this point in the year? PDO. The 5v5 goaltending’s produced a better save percentage which has kept the goals against down and a few more pucks are going in.
A point on PDO – while we say it reverts towards 1000, it actually regresses towards whatever the true talent is. A good rule of thumb for 5v5 is that the goalies will post a .920 and the players will score on 8% of shots; it’s not true with bad goalies. I’m not sure that I’d be inclined to bet on Edmonton’s PDO ending the year above 1000. Who’d you rather bet on from here – Carey Price or Nikolai Khabibulin?
In any event, if you were asked, “Is this edition of the Oilers better at 5v5 than the 30th overall teams?” you’d have a hard time saying yes, on this data. They also don’t compare particularly well to the teams around them in draft lottery contention. It’s all percentages separating them from those teams.
I wrote that on January 6, 2012. To that point in the season, Khabibulin had played 25 games and Devan Dubnyk had played 17 games. From that point forward, Dubnyk appeared in 30 games to Khabibulin’s 15 games. Khabby put up an .890 save percentage in those games, so my instinct that he was going to regress appears to have been a good one. What I whiffed on was that the Oilers decided to go with Dubnyk a lot more from that point forward. Dubnyk played pretty well too – he put up a .921 save percentage (overall; it would have been higher at 5v5) from that point forward, which meant that the Oilers didn’t tumble into a pit of regression.
Did Dubnyk play over his head? Tough to say. His track record entering the 2011-12 season wasn’t that great – he’d faced 1316 ES shots and posted a .914 ES SV%, which is pretty bad. His record was kind of polluted by the fact that he got lit up in his rookie season – 45 goals on 458 shots – and he’s actually had an averageish 2010-11, with a .921 ES SV%. For his career, Dubnyk’s now a .920 guy on 2461 ES shots. He doesn’t really have much of an AHL record but he played on some horrible AHL teams and I suspect that the further you get from the NHL, the more shot quality becomes a thing. All of which is to say that Dubnyk being an average goalie, a .920 ESSV% guy in the NHL, doesn’t seem outlandish to me.
That doesn’t mean that I thought that they should have paid him what they did, for what it’s worth. If you’ve paid attention to the details that have trickled out about the CBA, you’ll have noticed that one of the changes is that teams can no longer walk away from deals that are under $3.5MM. It struck me as odd last summer that the Oilers didn’t use the threat of arbitration walkaway to try and chisel down the amount that they’re paying Dubnyk or, if they did, that it was so ineffective. I guess the NHLPA was worried some other teams might think of this or be good at doing it.
In any event, the only thing that significantly changed at ES for the Oilers last year was the ES save percentage. It wasn’t spectacular or something that was so high that it’s almost certainly an outlier, like Jordan Eberle’s shooting percentage. It was out of keeping with Dubnyk’s career as a whole, if not his last season, but it’s certainly possible that Dubnyk is an average starting goalie in the NHL. It would have been nice if the Oilers had done something about the backup goalie position but I’ll talk more about that later on.
Nothing from the team level ES data last year makes me think that the Oilers didn’t get fair results in terms of goals for/against for their ES efforts and that, in thinking about what this year’s team might do, they were significantly better or worse at ES last year than the goals for/against would suggest. Next up: special teams.Email Tyler Dellow at firstname.lastname@example.org