Those articles used regular season goal differential, regular season estimated Fenwick close and “higher seed” to make predictions each round, weighing those results against my own predictions.
The final results for the 2015 WHL Playoffs were as follows, with the team with the higher regular season goal differential winning more often than when compared to the other strategies:
What about a bigger sample size through analyzing year’s past? Let’s take a look after combing through the information from 2013-14, 2012-13 and 2011-12 as well, giving us four full WHL postseasons to work with:
When I did my first piece on this topic during the winter, I only analyzed the 2013-14 playoff information and that obviously left me thinking that the estimated Fenwick close data was the most effective (as it won 12 series compared to 11 for both goal differential and seed).
That obviously isn’t the case when you amp up the sample size. While the team with the higher regular season estimated Fenwick close has still won 75% of all playoff series the past four years, it doesn’t come close to the other two strategies.
In fact, it doesn’t even out-predict simply taking the higher seed when picking a winner in a playoff series as that higher seed has won 49 of the past 60 WHL playoff series’.
Then we look at the big winner, regular season goal differential. The team with the better regular season goal differential has won an impressive 52 of the past 60 WHL playoff matchups - an 86.6% clip - including going a perfect 15-0 in the 2015 WHL playoffs.
Ultimately, I still think the best way to make a prediction is to weigh all the factors you can, gathering as much information as possible. My personal predictions this year went 14-1 in the WHL playoffs, with my only incorrect pick being Brandon over Kelowna in the final series.
Before making my predictions throughout the postseason I looked at all the data. Which team was favoured based on the above-mentioned categories, plus things like overall health of the their lineup, goaltending, special teams and anything else I could think of. I’m not saying that I would have the ability to pick at a 93% clip every year but analyzing as much data as possible needs to be your strategy if you expect to predict things accurately.
Thanks for reading!