Every pick tells a story, and sometimes the story doesn’t have the ending we wanted. That was the case on October 25th when the Green Bay Packers traveled to Detroit to take on the Lions, and our model locked in on a player prop that looked like a strong value play on paper.
The Pick: Savion Williams Under 24.5 Receiving Yards
The XO Sports AI model projected Savion Williams to finish with just 15.53 receiving yards in this matchup, making the Under 24.5 look like a comfortable target. The model gave this pick a 78.9% probability of hitting, which is a high-confidence signal by any standard. At odds of -113, there was solid value attached to a play the model strongly believed in.
So what made Williams such an appealing under candidate? The model weighed a range of factors including target share, defensive matchup, and recent usage trends. Detroit’s defense had shown the ability to limit production from secondary receivers, and Williams had not been a consistent volume target in Green Bay’s passing game. The numbers pointed toward a quiet afternoon, and the projected output of 15.53 yards reflected that clearly.
How It Played Out
Unfortunately, the result came back as a loss. Williams surpassed the 24.5-yard threshold and delivered more than the model anticipated. Games like this serve as a reminder that even high-probability plays carry risk. A single catch for a big gain, a quick slant that breaks into open space, or a late-game target in garbage time can flip a prop result in seconds. That is the reality of betting on individual player performances in the NFL.
The model was not wrong to like this spot. A near 79% probability is built on real data and real patterns, but the NFL has a way of producing outlier performances when you least expect them. One game does not change the long-term value of a model that consistently identifies edges in the market.
Keep Trusting the Process
This is exactly why tracking results over time matters more than any single outcome. The XO Sports AI prediction engine is designed to find plays where the probability is mispriced relative to the odds. Over a full season of picks, that edge adds up. One loss in a high-confidence spot is part of the variance, not a reason to abandon a smart strategy.
If you want to stay ahead of the market and see which player props and game picks the model is targeting each week, head over to XO Sports and explore the full slate of AI-powered predictions. The tools are built to give you an edge, and transparency on wins and losses is part of how we earn your trust.
Check out today’s picks at XO Sports and let the model do the heavy lifting for you.