Photo Credit: © Rick Osentoski-USA TODAY Sports

Proprietary Data Indicates Markstrom Has Played Well; It’s The Defensive Environment That Has Gotten Worse

The Canucks hot start to the season built them a lot of capital both in the standings and with their fanbase. However, yet again the month of November is proving to be their greatest adversary.

After leading the entire NHL in goal differential through the first few weeks, the Canucks have felt the sting of regression in terms of both their goals for (which have dried up) and goals against (which are peaking), which is generally not a recipe for success.

Before the season began, most agreed that if the Canucks were going to take a step forward and push for the playoffs, they’d not only have to improve in several areas, they couldn’t afford to fall back in the areas that they were good in last season, the most notable being goaltending. It’s well chronicled in Vancouver that last season from December onward, Jacob Markstrom was among the best goalies in the league. Many have credited the arrival of new goaltending coach Ian Clark, and subsequently Clark has been used as evidence that Markstrom had turned a corner and would be able to sustain this new higher level play.

That certainly looked to be the case early on this season, as the Canucks led the NHL in save percentage and goals against average through most of October.

Advertisement - Continue Reading Below

However, goals have started to go in, and in high volume. So has Markstrom’s run of high end play finally come to an end? Has the clock struck midnight and Markstrom is turning back into the pumpkin he was for the first six seasons of his NHL career?

Not so fast! Accordingly to goalie guru Kevin Woodley, a contributor to NHL.com and InGoal Magazine as well as a weekly guest on TSN 1040, Markstrom is doing just fine – it’s the environment around him that’s going south.

We can see well enough just based on the scoring chances and expected goals being surrendered on a rolling basis that the games are getting tougher for Canucks netminders, with both well above league average of late.

Advertisement - Continue Reading Below

But yesterday morning on TSN 1040’s morning show, Woodley went much deeper than that. Anyone who’s listened to Woodley’s hits on 1040 should know his fondness for private goaltending analytics, most often cited from Clear Sight Analytics, a company founded by former NHL goaltender-turned-media analyst Steve Valiquette.

While public data like scoring chances and expected goals try to account for shot quality through coordinates on the ice, accounting for the finishing quality of the shooter and other factors like rush and rebound chances (and do a reasonably good job), Clear Sight uses tracking to dig deep on other factors that influence the difficulty of each individual save, including pre-shot movement, layered screens and so on.

Clear Sight’s data is proprietary and thus unavailable for public consumption and scrutiny (as a wise man once said: “if you’re good at something, never do it for free”), but Woodley often drops nuggets during his radio hits, and he did so yesterday morning, beginning with the premise that the defensive environment has gotten worse in front of the Canucks goalies compared to last year, and the goalies have gotten better.

“Overall this season, the narrative has been that they’re better defensively, and they’re not,” Woodley told Mike Halford and Jason Brough. “Even when they were winning by scoring a lot of goals, the chances that they were seeing were still better quality than most give up.”

Clear Sight tracks 34 different points of data for each shot taken to create an extensive database. Each individual shot against then has a value assigned to it based on that data, and cumulatively each goalie has an expected save percentage based on all those shots and all those different factors.

This stuff can get a little heavy, and Halford had to ask Woodley to dumb it down a little. Woodley was more than happy to oblige, and laid everything out in a simple, easy to grasp manner. The data paints a picture of the defensive environment that a goalie is playing in. Teams that permit high volumes of slot line passes, screened shots, and breakaways for example, would present as much tougher environments for goalies to play in.

Advertisement - Continue Reading Below

“[Markstrom] is playing behind the sixth toughest environment in the National Hockey League based on the shot quality he has faced, and that doesn’t even include [the Dallas game],” said Woodley. “The only guys with harder environments in the starter category are Sergei Bobrovsky, Martin Jones, Henrik Lundqvist, Philip Grubauer, and Carter Hart.

“I don’t think a lot of Canucks fans, based on the start of this season, even as poorly as November has gone, would have lumped the Canucks in defensively in with the Florida Panthers, the San Jose Sharks and New York Rangers. They are the sixth worst environment for Markstrom. Those are the types of teams that they’re playing near.”

The fact that the Canucks still have a positive goal differential (+5) and are still above .500 as a team (in terms of points percentage at least) is at least partly due to Markstrom (and Demko) playing above calculated expectation.

“We can talk about all those different things: slot line, layered screens, defensive screens, breakaways. At the end of the day, when you add it all up, it creates an expected number. Jacob is outperforming that expected number. He’s ninth in the NHL [in terms of] the gap in which he’s outperforming [his expected] save percentage. Cumulatively he’s saved them 6.9 goals compared to what’s expected based on all that data, which again ranks in the top ten in the NHL.”

Woodley summed it up succinctly: “He’s playing well; the environment has gotten worse.”

Interestingly, without the added context from Clear Sight, the Canucks’ defensive zone coverage looks pretty decent. The HockeyViz heat map below suggests that Vancouver’s Threat in their own zone is -3%, meaning it’s slightly below league average, taking into account the varying levels of danger across the zone.

Advertisement - Continue Reading Below

To go from roughly league average (according to HockeyViz) to sixth worst defensive environment among starters (according to Clear Sight) is a pretty big jump. The major differences between the HockeyViz chart (which is probably the public leader in contextualizing hockey data) and what Clear Sight has brought to the table is the added context that shot location can’t account for: pre-shot movement, odd-man rushes and goalie sight lines.

The inference that one could make here is that the Canucks have actually been pretty good when it comes to limiting overall volume of shots and shots from dangerous locations, but they’ve been poor at limiting some of those additional factors that usually go unmeasured.

Clear Sight’s data corroborates that inference. Their data provides a window into where goalies have struggled, and as good as he’s been, Markstrom has areas in which he’s been sub-par: primarily screened shots and breakaways.

“16 breakaways (before the Dallas game), eight goals against,” Woodley pointed out. “That’s like 3.6 more than you would expect. League average on breakaways is 28% goals go in, he’s at 50%.”

Advertisement - Continue Reading Below

In fairness to Markstrom, perhaps the Canucks should consider giving up breakaways to less talented players – Dominik Hasek himself wasn’t stopping that Kyle Connor breakaway. But there have been a fair number of less impressive names lighting the lamp after streaking in alone.

“He’s actually lower than average (against breakaways), maybe [that’s an area where] you can see some improvements,” Woodley suggested.

“The other area where he struggles, and this is one area that might not be his fault, is screens and screen deflections,” Woodley continued. “Eight screen goals, seven have been screened deflections, a lot of it defensive screens. That has a better chance of going in, across the board around league, than a breakaway.”

“That’s one where you can look at it and go, ‘you know what Jacob, this is what’s hurting you, but I don’t know if you need to do anything differently. Maybe there’s something we need to do defensively differently, in terms of tying up sticks. If we’re gonna take away your sight lines, let’s not also let the opponent get a tip on the puck.'”

Advertisement - Continue Reading Below

Given that Markstrom has struggled with breakaways and screened deflections, but is still one of the league’s top ten starters in terms of save percentage above expectation, it’s probably safe to assume that he’s particularly good at making saves on other high difficulty shot types, such as shot following high-danger passes, like those that cross the slot line or the end line, forcing the goalie to move and open up. Once again, this seems to indicate that Markstrom’s save percentage could benefit more from the Canucks getting better at maintaining Markstrom’s sight lines, rather than requiring Markstrom himself to make significant changes to his playing style.

As for Thatcher Demko, it turns out that he’s doing even better than Markstrom relative to his peers. Granted, the data is split between starters and back ups, so Markstrom’s peers are considerably better than Demko’s, but it’s encouraging nonetheless.

“For Demko, on the back up side of things, he’s saved 2.83 goals [versus expectation] in roughly half the starts than Markstrom. Amongst the back ups he’s first in that expected goal differential and second in expected save percentage differential.”

Woodley also discussed how the statistics developed by Clear Sight have an application beyond goaltending: new, efficient ways of generating offence. It makes sense intuitively, given that the inverse of saving goals is creating them.

“It may have started with the premise of intuitively understanding what creates more goals, like what gives goaltenders trouble and measuring that, but it’s morphed into something that also measures offence. The Washington Capitals, for example, are a team that uses them, and when they won a Cup, they allowed their goaltending coach, Scott Murray, to play a strong role in the types of shots they look for.”

Advertisement - Continue Reading Below

The applications of both sides of the puck are fascinating. I’m sure that the Canucks hockey ops staff are at least aware of this type of data and likely have something in-house that measures this kind of metric for them. Clear Sight’s indictment of their defensive environment may just be a personnel issue, or it may be a matter of execution.

“Both these goaltenders are outperforming an environment that is probably worse [than] what most Canucks fans probably think they’re playing behind. Maybe not after watching [the Dallas game], but over the course of the whole season right up to [that game], they haven’t been a great defensive team, and these goaltenders have really kept them afloat.”

Whether or not this is sustainable going forward remains to be seen, but it is encouraging to note that goaltending doesn’t seem to be falling apart. In fact, the Canucks find themselves in a pretty enviable situation with two capable netminders. Now it’s about finding ways to help those netminders out a little bit more, so that the team can get back to its winning ways.