While the Bruins sent their fanbase on a third straight roller coaster ride of a season as it pertained to their chances of making the Stanley Cup Playoffs, their possession metrics at 5-on-5 were very strong.

One question that seems to come up often when "fancy stats" such as corsi are brought up, and one I remember Mike Felger asking on Felger and Mazz in the middle of the season, is simply, "why should I care?". After all, if a team wins a game because they dominated possession, or on the back of a great goaltending performance, the result is the same. What does it profit a team if they gain all of the corsis, and yet lose the game?

The answer, in short, is that possession metrics such as corsi seem to be more able to net a team a reliable edge over another, particularly in the context of a small sample like a playoff series, than metrics such as shooting percentage. And in their series against the Ottawa Senators, the Bruins seems to have that reliable edge.

To illustrate this, the website Corsica has a great tool which (with some help at times from Plotly, the site Corsica uses for their generated visuals) allows one to obtain a graph showing a rolling average for a team (or a player) for numerous different statistics. This allows us to look at how reliable an advantage, or disadvantage, the Bruins will have against the Senators in a given stat over a 7-game average, or the maximum length of a playoff series.

For starters, one of the most basic of metrics we can use is simply goals for percentage, or the percentage of goals scored in a game or series of games that are scored by a given team.

As you may know, the Bruins have a considerable edge over the Senators in goal differential, and that is seen in this graph, where it appears that the Bruins have had edges over the Senators for slightly longer and to slightly greater margins than they have had dips below Ottawa. Given that goal differential highly correlates to win percentage, this seems to give the Bruins an edge right off the bat. When you dig deeper, that advantage seems to get more clearer.

A basic way of looking at this is the "components" involved in determining how goals are scored and allowed. One division that can be made is between 5-on-5 hockey and special teams; a further division within 5-on-5 hockey is between possession, shooting and goaltending. Once these divisions have been made, we can use Corsica's tool to examine each of them.

We'll start with 5v5 corsi, the most talked about "fancy stat" in hockey. I'll make adjustments for both venue (home vs. away) and score effects (the disposition of teams to control less of the play when they are ahead and more of it when they are behind), and see how the 7-game averages have moved throughout the season:

The chart speaks for itself: the Bruins simply own this category against Ottawa, and only at the very end of the season were the Senators ever ahead of the Bruins in the seven-game rolling average. To put this in perspective, compare the chart you just saw to the 7-game rolling average in simple goals for percentage between the best team in the league in that department, the Washington Capitals, and the worst, the Colorado Avalanche:

When the discrepancy in 5-on-5 possession is so clear, the Senators have to make up that ground in some other area if they hope to win the series.

One category where it is thought that the Bruins could be weak is shot quality. What's the point of taking a bunch more shots than the other team, after all, if you can't get any shots in good scoring areas, while being generous in giving up such shots to the opposing team? Fortunately, Corsica offers a statistic to attempt to incorporate shot quality, expected goals. And unfortunately for the Senators, shot quality only bridges the gap between these teams minimally:

The Bruins were behind Ottawa for brief stretches late in October and early in March, but they still put a fair bit of space between themselves and their first-round opponents for the bulk of the year, albeit a bit less so under Bruce Cassidy.

So, what gives? How in the world did the Senators finish ahead of the Bruins, and how did they even get within 24 goals in goal differential? The answer lies mostly in shooting percentage slumps for the Bruins under Claude Julien, as well as a dip in performance from Tuukka Rask, which, as I see it, was partially attributable to the heavy workload of Rask, necessitated by the horrendous play of the backup goaltending of the club for much of the year. These two stats, shooting percentage and save percentage, are called PDO when combined, which has long been seen as indicators of fortunes bound to change one way or another. Why, you may ask? Well, one can see with the power of these moving average charts.

Here is such a chart for 5-on-5 shooting percentage:

The Senators certainly had a sustained run of being ahead of Boston in the middle of the season, but it wasn't quite as spread out as the difference Boston had in the possession metrics, and the turnaround the Bruins had in that department under Cassidy goes to show that shooting percentages are fickle; if a team only has an advantage over another in a category about 2/3 of the time, it is difficult to rely on it going forward. The same basic moral is present for goaltending:

The Sens have a slight edge for most of the season in 5-on-5 save percentage, and for full disclosure, Craig Anderson seems to be pretty competitive with Rask in terms of 7-game averages. This seems to be an area where the Senators can offset Boston's advantages elsewhere, but how much? Certainly an advantage of a couple of percent in stopping pucks would help a fair bit, but it doesn't seem particularly reliable that Ottawa will enjoy such an edge.

The Bruins' apparent advantage at even strength hockey is supplemented by the league's best penalty kill in terms of percentage of failed power plays against (as opposed to Ottawa's 22nd-ranked unit), and the league's 7th best power play in terms of percentage of successful power plays (as opposed to Ottawa's 23rd rank). But, how does this translate into a moving-average chart, when goals for per 60 minutes on 5-on-4 and goals against per 60 minutes of 4-on-5 (reminder that less is better) are considered?

As with shooting percentage and save percentage, special teams can only help so much as far as how a team can be predicted to do in a playoff series. You will doubtless hear of the Bruins' strengths both a man down and with the man advantage during the course of this series, but when you do, keep in mind the very real possibility that those strengths won't end up being a factor in the series against Ottawa.

In summary, I am by no means saying that the Bruins have this series in the bag. You can never be sure of the result of a playoff series in the NHL, and injury news regarding both Brandon Carlo and the difficult to replace (as far as corsi is concerned) Torey Krug, including the possibility that the latter may not even see the ice in Round 1, concern me a fair bit. But as a baseline, the Bruins should be favored against Ottawa, and by a reasonable margin (the linked chart, though, not accounting for the aforementioned Krug news). And based on the nature of hockey injuries, that the missing ice time can be "absorbed" by the position group as a whole, somewhat limiting the drop-off, and given that there appears to be a bit of a gap between these teams as a baseline, I don't think the injury news takes away Boston's favored status. (Consider that the Penguins' chances against Columbus only moved a single percentage point in the previously-linked model from hockeyviz.com before and after the Kris Letang injury was accounted for, although it appears the expected home ice team did change)

And if the Bruins do advance and get their players healthy, they are in a position to possibly surprise a fair number of those who would fool-heartedly ignore the fancy stats.