That’s a wrap for the Divisional Round. Here’s a breakdown of the model’s performance for Week 20:
It was a weekend where all favorites won, so the model posted a 25% accuracy rate on moneylines with 1 win, and 3 losses across all games. As designed, the model targeted value in underdog picks, but unfortunately none of them panned out this week, going 0-3. This brings the overall season moneyline performance to 171-99-12 (63%), with 38 upsets called.
The model had High or Very High confidence in three games this week and went 1-2 for a 33% success rate. This brings the model’s high-confidence season moneyline record to 88-37-1 (70%).
Bills @ Broncos –We knew this game would be close, with the Bills having only a 55% chance to win. The game ended up going to overtime, where the Broncos were able to secure the victory.
Texans @ Patriots – Holy turnovers, Batman! Somehow this game remained close, with C.J. Stroud throwing more passes to the other team than his own. The Texans defense was spectacular, but without offensive support (and with a defensive touchdown from the Patriots), the Texans couldn’t pull it out.
Divisonal Round – Top Bets had a rough week, going 0-5. The model expected most of the games to be close and believed some underdogs could pull off wins. The 49ers game was the only true blowout and the one the model missed most significantly.
Here are the full results:
Parlays:
Moneyline Bets:
Spread Bets:
It was a rough week for predictions and bets, especially compared to last year when the model went 4–0 in the Divisional Round. With only favorites winning this weekend, value opportunities were hard to come by.
With just two games left before the Super Bowl, we’re excited to see how the model rounds out the postseason.
On to the AFC and NFC Championships!