by Anthony Reinhard
The XFL season will begin this Saturday and will end a period of speculation about which of the eight newborn franchises will emerge as the team to beat in the XFL’s inaugural season. This is always a tricky phase for folks like myself who gravitate more toward empirical evidence and data. While there is very little reliable information available, the absence of information presents an opportunity to test one’s understanding of uncertainty.
Months before the season, our base assumption should be that each team will win 5 of their 10 games on average and have a 12.5% chance (1 out of 8) of winning the 2020 XFL championship. As rosters start to take shape we can begin to adjust these assumptions. The challenging part is figuring out how much to adjust. I see betting markets as an ideal breeding ground for zeroing in on just how much uncertainty exists in a situation. As far as finding a favorite for the upcoming XFL season, there isn’t much uncertainty at all. Every sportsbook I looked at had Dallas as the favorite to take home the title in April.
I used the information above to back into Elo ratings. Those ratings will be serving as a starting point for my XFL model, which I will be updating each week. Ultimately, I found that the difference between having a 20% chance to win the XFL title at the start of the season and a 7.5% chance doesn’t really translate to as big a gap as one might think. My initial ratings would give Dallas less than a three point edge against Seattle on a neutral field if they played in Week 1. Despite that modest spread, Seattle would need to win by about 18 points (I won’t bore you with the math) to earn a higher Elo rating than Dallas entering Week 2.
My final conclusion is that while there is a meaningful consensus about who the best and worst teams are as we enter the season, we shouldn’t be surprised if this hierarchy is completely toppled come March.
2 thoughts on “Deep Routes: Improbable Probability – XFL Week 1”