April 12 2017 12:12PM
It's the most wonderful time of the year. The regular season is over and five Canadian teams made the playoffs. Quite the change from last year, when none made it. Just like last year, however, the Washington Capitals won the President's Trophy and are a hot pick to win the Stanley Cup.
Do the numbers have a different prediction? In this article I'll look at shot and goal differentials to separate the pretenders from contenders.
Here is a link to my analysis of last season's numbers for reference. You will notice it was written after the playoffs, so this will be this model's first real trial.
Here is the 2016-2017 shot and goal differentials (all situations) in a scatterplot:
A quick refresher on the chart’s quadrants:
Top left - Negative shot differential but a positive goal differential (Quadrant 1)
Top right - Positive shot and goal differentials (Quadrant 2)
Bottom right - Positive shot differential and negative goal differential (Quadrant 3)
Bottom left - Negative shot and goal differentials (Quadrant 4)
Teams that made the playoffs have orange plot-points, teams who did not are blue. Historically speaking almost all Stanley Cup Winners are quadrant 2 teams. Unsurprisingly and predictably the vast majority of playoff teams find themselves exactly there. Yet there are exceptions. Notable absent are Anaheim and Chicago, the Pacific and Central division winners respectively. Both these teams had dominant regular seasons, yet both teams got outshot on a consistent basis. Anaheim practically broke even on the shot clock, but cup winners need more than that. Chicago is always dangerous and Vegas has them pegged with the best odds to win the cup. History suggests that it is very unlikely that a team with a negative shot differential goes all the way. A quick look at the Stanley Cup winners in recent years and their shot differentials:
2016 - Pittsburgh - 2nd
2015 - Chicago - 4th
2014 - Los Angeles - 3rd
2013 - Chicago - 2nd
2012 - Los Angeles - 6th
This trend hold true for the last 20+ seasons. Only Boston in 2011 and Pittsburgh in 2009 won the cup while being outside the top 10 in shot differential (source). For the 2016-2017 season Dallas was the cut off for top 10. Looking at the scatterplot we can eliminate any teams to its left. That leaves Boston, Washington, San Jose, Edmonton, Nashville and Pittsburgh (4 non-playoff teams are on the list).
A pretty clear picture. Now the same for goal differentials:
2016 - Pittsburgh - 2nd
2015 - Chicago - 5th
2014 - Los Angeles - 7th
2013 - Chicago - 1st
2012 - Los Angeles - 9th
Edmonton is the only team from the West that made both lists. Yet this should not be interpreted as a guarantee for them to make it to the final. History simply suggests that it is very likely that one of these teams are going to win the cup. Unfortunately Washington and Pittsburgh are set to clash in the second round and that is a shame. They both deserve to get past the second round.
One of the reasons scatterplots are so useful is because they can draw attention to data points that would otherwise go unnoticed, in this case Washington, Boston and Los Angeles. A quick note on each.
- LA did not make the playoffs despite having the second best shot differential. This goes to show the limits of this type of analysis. As anyone who has seen an LA game can tell you, they shoot the puck from any and everywhere. Highly skilled teams like San Jose and Edmonton can turn these shots into a deflection or a tip, but LA unfortunately lacks the skilled players for this (shooting percentage of 7.79% and PDO of 98.3).
- Washington is on the other end of the spectrum. They have a fantastic shot differential and a ridiculous goal differential. Unfortunately for them their shot differential (again which is fantastic) does not line up with their goal differential. They have a lot more goals in their favour than they should and their league leading PDO of 102.7 speaks to that. The key here is sustainability as PDO regresses to 100 very heavily in the long run. Looking at past cup winners their PDO at the end of the regular season:
- 2016 - Pittsburgh -100.9 (9th)
- 2015 - Chicago - 100.5 (12th)
- 2014 - Los Angeles - 99.9 (16th)
- 2013 - Chicago - 102.3 (4th)
- 2012 - Los Angeles - 99.9 (18th)
- Boston sported the best shot differential. Unlike LA however they were able to convert on their chances (shooting at 8.49% and PDO of 99). If they can keep this up and if Rask returns to form they should be considered a serious threat.
If history is any indicator of what will happen this season, Washington, Pittsburgh, Boston or Edmonton will hoist the cup this year.
- Unfortunately for Washington, their brand of hockey is does not usually lead to a cup. High PDO in the regular season does not translate into high PDO in the playoffs. Pretender, not contender.
- Pittsburgh has terrific shot and goal numbers and their PDO of 101.5, while high, is within the normal range. Contender.
- Boston is has sustainable numbers and are under-performing. Contender.
- Edmonton is finally back in the playoffs. Their shot and goal numbers are well within the normal, sustainable range and fans can rest easy. This is not a mirage (*cough* 2014 Colorado). The majority of the team does not have any playoff experience, and Talbot has already played a league leading 73 games. The numbers don't lie though. Contender.
- Chicago (Quadrant 1) vs Nashville (Quadrant 2) - Chicago
- Minnesota (Quadrant 2) vs St Louis (Quadrant 2) - Minnesota
- Anaheim (Quadrant 1) vs Calgary (Quadrant 2) - Anaheim
- Edmonton (Quadrant 2) vs San Jose (Quadrant 2) - Edmonton
- Montreal (Quadrant 2) vs New York Rangers (Quadrant 1) - Montreal
- Ottawa (Quadrant 4) vs Boston (Quadrant 2) - Boston
- Washington (Quadrant 2) vs Toronto (Quadrant 1) - Washington
- Pittsburgh (Quadrant 2) vs Columbus (Quadrant 2) - Pittsburgh