Right after the 2016 season, I published my findings on Pythagorean Expectation forecasting the 2016 CFL season. Long story short: Pythagorean Expectation (PE) is a way to predict how many wins a team should have earned for a given season based on point differential.
(Hey, here’s a really ancient 2-minute video explanation. It’s for baseball, but still. –Ed.)
PE is often accurate, but sometimes, as in the case of the 2016 Hamilton Tiger-Cats, PE is not on target. I decided to dig in further to see if I could find out why teams miss expectations. It went in quite a different direction than I expected, but it was a fun piece to research.
I gathered five years’ worth of data from CFLstats.ca, including wins, points for, points against, yards for, yards against, turnovers for, and turnovers against (which includes fumbles, interceptions, and loss of downs, anything where the a team gave the ball up unwillingly is being noted here). I then calculated the differences of each, and then ran a series of correlations (that is, I wanted to see how much point differential, yardage differential, and turnover differential correlated with wins).
A perfect positive correlation (in which a rise in one piece of data indicates a rise of data in the other piece of data) is registered as 1, and a perfect negative correlation (in which a rise in one piece of data indicates a drop in the other piece) is a -1. If there’s no correlation whatsoever between each data set, it would be a 0.
As expected, point differential correlates to wins more closely than the other two, registering a .894. Yard differential (yards for minus yards against) is still somewhat strong, coming in at a .757. Turnover differential (turnovers for minus turnovers against) was a bit lower at .713.
Let’s go through the 2016 CFL season team by team and see why they met (or didn’t meet) their Pythagorean Expectation.
The Stampeders finished 15-2-1, exceeding their Expectation by almost exactly 1.5 wins. Their yardage differential was 873, which is the 6th highest over the past five years, and their turnover differential was 18, the fourth highest in the past 5 years. All in all, this team was the best over the past 5 years in the regular season.
The Lions slightly exceeded their Expectation at 12 wins, and were one of the few teams in the past five years that had a better yardage differential than the Stamps, but they had a negative turnover differential at -8. They only created 29 turnovers, the third lowest in the past five years, but mitigated that somewhat by being around average in giveaways.
Winnipeg Blue Bombers
Exceeding their Expectation by about a win, winning 11 games, and having the largest turnover differential in the past 5 years by a long shot make things seem like all is right in Bomberland.
However, Winnipeg’s massive turnover differential of plus-32 screams outlier to me, especially when the next highest was at 19. That’s not likely to repeat itself, and the -850 yard differential would normally put a team anywhere from 3 to 7 wins. 20 teams over the past five years had a negative yardage differential, only 4 had winning records, and among those, only one other, the 2013 Toronto Argonauts, had a positive PE (meaning the other two had fewer than 9 wins expected, but exceeded their PE).
Based on what I’m seeing, if the turnovers regress to the mean as is statistically likely, the Bombers could be poised to plummet in the standings.
Even with all the transition after the Grey Cup win and the personnel attrition that comes with it, the Eskimos barely missed their PE, had a plus 727 yardage difference, and a -5 turnover differential. The turnover differential appears to be tugging down the yardage’s effect on points slightly, so there were really no big surprises here.
This version of the Riders exceeded their expectation by over half a win, but that’s about as positive as it gets. Yes, they jumped from 3 wins to 5, however, they went from being massively unlucky based off their expectation in 2015 to being slightly lucky this year. Their PE actually fell from 2015 to 2016 by a win and a half, the team had nearly a 50-point fall in point differential, and also suffered a fall in yardage differential and turnover differential. This is a bad team, and despite the real win increase, everything else is trending the wrong way.
Despite the Grey Cup victory, this was an average team (points-wise, anyway) in 2016. They were actually one of the best in the league with yardage differential with 851, but a negative 12 turnover differential really hurt them. The 2016 team had the highest positive yardage differential for a team that finished under .500, with the next best at 338. I had been looking closely at only their point differential before (which is about middle of the road), but I may have underrated them over the past few years based off their yardage differentials.
The Ti-Cats under performed their PE last year by over 2 wins, having an PE over 9 and a true win total of 7. Their PE was about as close as it can get, in my mind, because they had a slightly positive yardage differential and a plus four turnover differential. If they put up the same numbers this year, I’d expect a 9-9 season.
The Alouettes fell short of their PE by over a game. Turnovers were nearly even, but the yardage differential was -1077, one of the worst over the past five years. It’s a pretty open-and-shut case here.
Believe it or not, the Argos exceeded their PE, but a -916 yardage differential and a -13 turnover differential tell the tale here.
Coming soon: Predictions for the 2017 CFL season based on these mathematical discoveries and adjust for what has transpired this offseason.