One of the requests from readers, when I took over, was to have some guest posts, to allow people an avenue to provide one-time content or allow people a chance to spread their wings.

With that in mind, I started the “From The Community” series a few weeks ago with Stephan Roget breaking down some players who could rebound or comedown this upcoming season. The response was extremely positive and generated some great discussion, so we will continue it.

It’s important to remind everyone that the point of this series is built on people taking a leap of faith. Please keep your comments to being respectful and constructive. You can not agree with what they are trying to talk about but this series relies on people reaching out to me and doing the work. I would like that to continue all season long. Obviously, Brandon Sutter is a trigger point in the comment section and thus it has to be stressed heavily that this is a post from someone who reached out to us.

For this week, Matthew Dolmage reached out to me for his chance in the spotlight. He provided this quick blurb about himself:

Without further ado, let’s dive into the second post in the series.

In a recent “What Would You Do Wednesday”, many CanucksArmy readers requested some introduction to statistics and further explanation of statistical concepts, to help them understand the advanced stats that are often used by CanucksArmy writers. There have also been a number of heated debates in the comments section over the past few weeks about Brandon Sutter’s defensive abilities, and whether Sutter’s elevated on-ice save percentage is the result of a genuine effect Sutter’s play is having on the ice or is the result of random chance.

In this piece, I’m going to discuss some simple but extremely important concepts in statistics – significance, noise, and randomness – and then use these tools to investigate Sutter’s on-ice save percentage to determine what (if anything) Sutter is doing to elevate the play of the goalies behind him.

Statistical significance is a concept that comes up a lot in sports analytics. When an analyst asks “is this information statistically significant”, they’re essentially asking “is it more likely that this information is meaningful, or just the result of random chance?”.

For a result to be statistically significant, it must deviate noticeably from the result that was expected, and it must be more likely that this deviation is a result of a genuine effect, and not a sampling error or random chance. For example, if you flip a coin ten times, you might get a result of seven heads and three tails. This is a deviation from what was expected – five and five – but with such a small sample, it’s very likely this is just the result of chance. However, if you flip a coin a thousand times and get 700 heads and 300 tails, you might reasonably think there’s something going on with your coin that’s causing it to land on heads so much more frequently than on tails.

My favourite example of the importance of sample size in sports is that of Jeff “The Bra-barian” Cowan. Jeff Cowan was a fourth-line grinder who played a little over 400 games in the NHL between 2000 and 2008. Cowan came to the Canucks halfway through the 2006-2007 season, and shortly after had an incredible run where he scored six goals in four games. If you had only taken a ten game sample of Cowan’s season around when he scored those goals, you might think Cowan was one of the league’s top goal-scorers. But on the season, Cowan scored seven goals – bang-on his career average. Goals are “noisy” – they are rare and are distributed unevenly – so to get a sense of a player’s true talent level – to know if a result is really statistically significant and not just the result of chance – you need a very large sample size. This is why analytics has tended to rely much more on shot attempts than on shots or goals. Shot attempts occur much more frequently than goals, or even shots on goal, and give analysts a much larger sample of data to work with, which leads to a more reliable analysis of past results, and better predictions of future success.

But how does all this apply to Brandon Sutter?

As CanucksArmy commenter Killer Marmot has pointed out on several occasions, Brandon Sutter’s on-ice save percentage is consistently higher than average. Goalies simply stop more pucks when Brandon Sutter is on the ice than when his teammates are on the ice. Since the 2012-2013 lockout, Brandon Sutter’s on-ice save percentage has been average or above average in five of six seasons.

We’ve checked the first box in our two-part test – these results deviate noticeably from what we might expect. So Sutter must be doing something to improve his goalies’ play or suppress his opponent’s shooting ability when he’s on the ice, right? Not so fast. We have to determine if Sutter’s results are actually significant, or are just the result of random chance playing tricks on our brains.

Here’s a breakdown of the relevant statistics from Sutter’s post-lockout seasons – we’ll be looking at team even strength save percentage and Sutter’s on-ice save percentage, shots against, expected goals against, and actual goals against:

So what can we learn from all these numbers?

Since the lockout, goalies have saved eight more goals than one might otherwise expect when Sutter was on the ice, or about 1.3 goals per season above average. In just his time in Vancouver, Vancouver’s goalies have saved two more goals than expected when Sutter was on the ice, or 0.67 goals per season above average.

Forwards tend to be on the ice for far fewer shots against than goaltenders are – consider that in his

entire time in Vancouver, Brandon Sutter has been on the ice for fewer even strength shots against than Jacob Markstrom was in just sixty games last season. These results fall far below the threshold for statistical significance. And even if Sutter was doing something to subtly influence his goaltender’s save percentage, this is far from a useful skill – a goal saved above average in a season is unlikely to result in even a single additional win. None of this is to say that Sutter isn’t a valuable player or a useful defensive forward, but it’s much more likely that a goal here and there is the result of a lucky bounce than it is that Sutter is doing anything to influence the play of the goaltender behind him.