Sometimes the model is right and the result still goes the wrong way. That’s the reality of sports betting, and it’s exactly what happened with our player prop on Woody Marks to stay under 14.5 receiving yards in the New York Giants vs. Houston Texans matchup on October 25th, 2026.
The Pick Rationale
The XO Sports AI model projected Marks to finish with roughly 13.14 receiving yards, putting his expected output comfortably below the 14.5-yard line. That projection translated into a model probability of 68.76%, which is a meaningful edge in the player prop market. At odds of -113, this was the kind of bet where the math genuinely favored the under.
The reasoning behind the projection made sense on paper. Marks is primarily a running back whose receiving role tends to be limited and situational. Against a Houston defense that is built to take away easy check-down opportunities out of the backfield, the expectation was that his targets would be minimal and his yardage after the catch would be controlled. The Giants also entered this game with a game plan that did not lean heavily on Marks as a pass-catching option.
How It Played Out
Despite the strong model backing, the pick came back as a loss. Marks cleared the 14.5-yard threshold, which means something broke from the expected script. Whether it was an unexpected target share, a scramble drill that found him in space, or a simple garbage-time reception that padded his stats, the final number landed above the line. It happens, and it is a reminder that even high-confidence player props carry real variance on a play-by-play basis.
This is not a result that undermines the process. A pick with a 68.76% model probability will still lose roughly 31% of the time. Bankroll management and trusting the long-term edge are what separate disciplined bettors from the rest of the field.
The Bigger Picture
One loss does not tell the whole story. The XO Sports AI platform identifies edges across hundreds of player props every week, using advanced data modeling to surface opportunities the market underprices. The goal is not to win every single bet. The goal is to find consistent positive expected value over a full season of picks.
Marks going over the line on this particular Sunday is a single data point. The model’s methodology, its accuracy over time, and the edges it continues to find are what matter most to sharp bettors playing the long game.
Start Finding Your Edge Today
Want to see which player props and game picks the XO Sports AI is targeting this week? Head over to XO Sports and explore our full slate of NFL predictions, complete with model probabilities and projected stat lines. Stop guessing and start betting with data on your side.