Skip to content
Join our Newsletter

Numbers Never Lie: The Canucks are not terrible

While numbers never lie, it’s sometimes difficult to figure out exactly what truth they’re telling.
Numbers Never Lie Banner

While numbers never lie, it’s sometimes difficult to figure out exactly what truth they’re telling. To help you out, Numbers Never Lie is a weekly look at the Canucks’ advanced statistics, and figuring out exactly what they have to say about the Canucks’ season and players.

Ten games into the 2017-18 season, the Canucks are in a place that no one expected them to be: top-5 in the Western Conference.

It’s worth keeping in mind, of course, that win-loss records mean very little this early on. Certainly, it’s better to have a winning record, but a hot start is no guarantee of success throughout the rest of the season. For instance, the team directly ahead of the Canucks in the West is the Las Vegas Golden Knights, who have a 7-1-0 record. Does anyone, other than their social media manager, actually think that’s going to last?

 

 

We can get a bit better of an idea of how the Canucks are doing through ten games by dipping our toes into the world of analytics. I’m going to be specifically looking at corsi and expected goals.

For some, those words send them running for cover, but there’s no need to be antagonistic towards analytics or scared of them. While analytics are not the be-all-end-all of understanding hockey, they can be illuminating.

Corsi is simply shot attempt differential at 5-on-5. You take the total number of shot attempts directed at the opponent’s net, subtract the total number of shot attempts directed at your own net, and you have your corsi. It frequently is expressed as a percentage: 50% means the shot attempts were equal at both ends of the rink, which means that a number above 50% is preferable as it indicates you out-shot your opponent.

In this case, this early in the season, corsi is useful as it increases our sample size. A win-loss record right now for the Canucks gives us a sample size of 10. Looking at 5-on-5 goals increases the sample size to 36. Expanding to 5-on-5 shots increases the sample size to 430. Expanding all the way to all shot attempts at 5-on-5 gives us a sample size of 765 individual on-ice events.

By increasing the sample size, we reduce the impact of randomness. In the sample size of 10 wins and losses, a string of good or bad luck can obscure the true talent level of a team. In a sample size of 765 shot attempts, a string of good or bad luck has a smaller impact.

I’ll get into expected goals in a moment; for now, let’s look at the Canucks’ corsi.

After their very successful road trip and return to Rogers Arena, the Canucks’ underlying numbers are perfectly average. At 5-on-5, the Canucks are currently in 18th in the NHL in raw corsi. Their corsi percentage of 48.85% indicates they have been out-shot, but not by an egregious amount.

The Canucks look better when you go by adjusted corsi. We know that teams who are behind in the score during games take more shot attempts, while teams that are ahead tend to sit back and defend their lead, taking fewer shot attempts. We also know certain rinks in the NHL under or over-count shot attempts. We can adjust for these factors, or rather, we can let people with a background in statistics calculate these adjustments for us.

When you run the adjusted numbers on Corsica.Hockey, the Canucks jump up to 15th in the NHL at 49.34%: still average, but considering how terrible they were last year, average is a massive improvement.

But the Canucks fare a little better than average when you look at expected goals.

Expected goals (xG) is a model created by Emmanuel Perry that attempts to account for shot quality. The model does this by estimating the probability that a shot attempt will become a goal according to a series of factors based on seven seasons of shot data:

  • Shot type (Wrist shot, slap shot, deflection, etc.)
  • Shot distance (Adjusted distance from net)
  • Shot angle (Angle in absolute degrees from the central line normal to the goal line)
  • Rebounds (Boolean – Whether or not the shot was a rebound)
  • Rush shots (Boolean – Whether or not the shot was a rush shot)
  • Strength state (Boolean – Whether or not the shot was taken on the powerplay)

So, when Bo Horvat takes a wrist shot off the rush from an angle of 53 degrees and a distance of 25 feet at 5-on-5, Perry’s model can calculate the probability of that type of shot, on average, going into the net. While there are a few other factors that go into an individual shot attempt becoming a goal — shot velocity, an individual shooter’s deception and accuracy, whether there was a pass just prior to the shot, etc. — this model accounts for quite a bit.

If you’re interested in knowing more about xG and Perry’s other statistical models and projects, I highly recommend reading the Corsica.Hockey blog.

The Canucks’ 5-on-5 adjusted expected goals for percentage (xGF%) is 51.34%, good for 13th in the NHL. This suggests that the Canucks are outperforming their corsi by creating higher quality chances for themselves and/or preventing higher quality chances for their opposition.

Corsi Adjusted Corsi Expected Goals Adjusted Expected Goals
Percentage Rank Percentage Rank Percentage Rank Percentage Rank
48.85% 18 49.34% 15 51.07% 13 51.34% 13


For the Canucks, it’s mostly prevention.

If the Canucks are successfully forcing shots to come from the outside, from areas where a shot’s probability of becoming a goal are lower, we would expect to see that reflected in their expected goals against (xGA).

Sure enough, the Canucks are third in the NHL in xGA per 60 minutes of ice time. They give up few shots against and the ones that they do give up have a lower probability of becoming goals. In fact, their expected goals against and actual goals against are nearly identical. The Canucks’ xGA is 15.16; they’ve actually given up 15 goals at 5-on-5.

One issue is that they don’t get many shots of their own. Only the Minnesota Wild have fewer shot attempts per hour of ice time. The Canucks’ expected goals for (xGF) per 60 minutes of ice time is only slightly better at 28th in the NHL. They’ve managed to outperform their xGF of 15.83, scoring 21 goals at 5-on-5.

The chart below looks strictly at shot rates, but it captures a bit of what the Canucks are doing. More than any other team in the NHL, the Canucks are committed to playing low-event hockey. And when those events do happen, the Canucks want them to be as low-quality as possible for the opposition.

Canucks shot rates - numbers never lie


What does this mean overall? First of all, it indicates that the Canucks are not terrible. They may not be good, as they are heavily dependent on suppressing shot attempts and playing low-event hockey, but they’re also not bad.

The Canucks are average.

For those hoping to see the Canucks tank for a high draft pick, that might sound like a death knell, while those optimistic few who see the playoffs in the Canucks future, you might think I’m being cynical. I’m not. In fact, these numbers make me feel quite optimistic. The Canucks are legitimately limiting shot quality against so far this season.

We can see that shot quality at work by looking at a heat map of unblocked shot attempts against, as provided by HockeyViz.com. Compare the two heat maps below. On the left is the Canucks; on the right is the Chicago Blackhawks.

numbers never lie heat map canucks blackhawks


The Blackhawks have the highest xGA per 60 minutes in the league. Looking at the heat map, it’s not hard to see why. The Blackhawks give up a massive amount of shots from directly in the slot and all around the net.

Compare that to the Canucks, who give up a relatively low number of shots from in front and severely limit shots from in tight on the glove side.

So, the Canucks are preventing quantity of shot attempts against and, to a smaller extent, preventing quality of shot attempts against. Combine those two with some decent goaltending to start the season, and it’s no wonder they have allowed the sixth fewest goals against per hour.

In the next Numbers Never Lie, I will be looking at which players, lines, and defence pairings are driving this for the Canucks.