UFC 197: Jones vs. St-Preux - Statistical Model Estimates


MAIN CARD

Jon Jones 90.7% Ovince St-Preux 9.3%
Demetrious Johnson 74.4% Henry Cejudo 25.6%
Anthony Pettis 66.3% Edson Barboza 33.7%
Robert Whittaker 73.5% Rafael Natal 26.5%
Andre Fili 55.8% Yair Rodriguez 44.2%

PRELIMS

Sergio Pettis 75.7% Chris Kelades 24.3%
Danny Roberts 52.0% Dominique Steele 48.0%
Carla Esparza 56.4% Juliana Lima 43.6%
Glaico Franca 53.1% James Vick 46.9%
Cody East 69.8% Walt Harris 30.2%
Clint Hester 51.5% Marcos Rogerio de Lima 48.5%
Kevin Lee 61.5% Efrain Escudero 38.5%

I've mentioned before that this model has had trouble accurately forecasting fights featuring very highly ranked fighters. Well, this problem manifested itself big time with Jon Jones vs. Ovince St-Preux. Without any adjustments, the model gives Jones a 98.6% chance of beating St-Preux. That doesn't pass the smell test to me, so I've decided to implement a couple controls to regulate the model's estimates in these extreme cases. The result is a revised estimate of a 90.7% chance of Jones winning. That's still exceptionally high but at least it's a believable forecast.

The model has no such issues with Demetrious Johnson vs. Henry Cejudo. Their statistical profiles are actually somewhat similar; Johnson lands takedowns more often but their striking statistics match up pretty well. The key difference, of course, is that Johnson is much more accomplished and higher ranked as a result.

Underdog picks are: Andre Fili, Glaico Franca, and Clint Hester. The model also sees Danny Roberts vs. Dominique Steele as a very close fight despite Roberts being a -360 favorite on the betting markets.

Reply · Report Post