Vanecek, Samsonov and Team Defense

A look at the Caps’ play in front of their two goalies

The Boston Bruins may be the best team in the NHL when it comes to five-on-five defense. No team has a lower expected goals rate, allows fewer scoring chances or high-danger chances against. They allow the second-fewest shot attempts and shots on goal, and do it regardless of the game score (their expected goals-against rate when trailing is the lowest in the League).

Here’s a pareidolic heat map of what they’re allowing at fives:

via HockeyViz

Nada. Nil. Zilch.

And here’s what the Caps are allowing at five-on-five with Vitek Vanecek in net:

via HockeyViz

It’s a much meltier face, but what’s really interesting is that xGA/60 number – the Caps, with Vanecek in net, are nearly the Bruins in terms of five-on-five defense. To wit, here are the Bruins and the Vanecek Caps at fives:

data via NatStatTrick

Aside from the high-danger chances, the two map fairly closely; the xGA rate is different here because NatStatTrick and HockeyViz have different models, but they’re plenty close enough to make the point, and that point is that the Caps have played terrific defense in front of Vitek Vanecek… and he’s outperforming that:

via JFresh

That’s a pretty good deal, of course, but this post isn’t about how good Vitek Vanecek has been (this one is, though, so go read it). Rather, it’s about how the Caps have played in front of Vanecek… and how it’s different than how they’ve played in front of Ilya Samsonov. First up, the heat maps:

via HockeyViz

With Vanecek in net at fives, the Caps have played truly great defense; with Samsonov, they’re below League average (like, Edmonton “below average”).

That’s weird.

Sure enough, throw Sammy in goal and the Caps’ shots-, chances- and expected goals-against metrics rise across the board (and that’s before we even get into the goalie’s performance). So let’s dig in on that. (And, no, they don’t “hate playing for Sammy” or whatever.)

Maybe it’s score effects. We know that teams that trail tend to allow more goals as they open up a bit while playing catch-up, and the Caps actually do see a decent spike in expected goals-against down a goal:

via HockeyViz

So maybe Samsonov is playing a lot more at fives down one relative to Vanecek?

data via NatStatTrick

Okay, so there’s a little something there, but not much (and, of course, some of the responsibility for being down a goal falls on, y’know, the goalie).

Maybe it’s Samsonov’s rebound control that’s the difference? It wouldn’t be hard to convince yourself that the two netminders are facing similar initial shots, but by mishandling them, Samsonov is creating higher-danger follow-ups that Vanecek isn’t having to deal with. Here’s some (five-on-five) data on rebounds and frozen pucks:

via MoneyPuck

The expectations (xRebounds Per Save) are the same, but the actual rebounds are quite different, with Samsonov allowing more than double what Vanecek is giving up (owing mostly to Vanecek being way better than League average and Samsonov being slightly worse). This jibes with the data at NatStatTrick, which tells us that among fifty goalies with 750 minutes played at five-on-five, no one has faced rebound attempts at a lower rate than Vanecek… or a higher rate than Samsonov. Now, this is counting rebound attempts (i.e. follow-up shots) and not rebounds (i.e. what’s bouncing off the goalie), but it’s probably a pretty decent proxy when we’re comparing two goalies playing behind the exact same defense.

And what’s the difference between being really good at controlling rebounds and being relatively poor at it? Per NatStatTrick, the raw number of rebound attempts Vanecek has faced is 43. For Samsonov, that number is 104, or 142 percent more (in around 17 percent more minutes). Now think about where most of those rebound attempts are coming from.

Feels like we’re getting somewhere. Samsonov is facing roughly one-and-two-thirds more rebound attempts per night at five-on-five than Vanecek, and giving up about a quarter of an expected goal more per game. Couple that with a little more “down-1” time, and we’ve probably answered most of our question.

Here’s one last piece of the puzzle: Vanecek has faced 587 shots at five-on-five, Samsonov 685, and the difference between the average shot distance they’ve faced is just six inches (Vanecek 34.4 feet, Samsonov 34.9). Sure looks like they’re getting a similar workload that Samsonov is simply making more difficult than it needs to be.

If you want a bottom line, it’s probably this: the Caps are a very good defensive team, and one of their goalies has taken advantage of that, while the other… has gone the other way with it. The good news is that if Ilya Samsonov can clean up his rebounds the way that Vitek Vanecek has, he can probably enjoy some of the same successes that Vanecek has. But until then, the Caps should be fine being a very good defensive team in front of a goalie that makes them look like it.