Welcome to this Wednesday edition of Props Daily!

We’ve barely had time to process Week 4 and we’re already getting ready to do it all again.

Week 5 starts tomorrow night, but four weeks into the season feels like a good time to look at what our NFL model has been doing.

Because… it had a rough start.

Week 1 finished at -7.2% profit, and Week 2 was even worse at -37.9%. In fairness, the model was basically trying to figure out what the hell was going on.

We had very little 2026 data to work with. There were new quarterbacks, new coaches, injuries, changing roles and teams that looked completely different from what we expected coming into the season. For a model built around player performance and game context, a lot of what it thought it knew was changing very quickly.

Then the picture started getting a little clearer.

Week 3 finished at +1.2%, essentially bringing us back around break-even for the week. Week 4 was the first real jump, finishing at +15.0% profit.

Four weeks is nowhere near enough data to declare victory, but the improvement is encouraging. And we’re learning more than which picks win and lose.

We’ve also been testing when the model makes its picks.

We’ve compared its performance at 15 minutes, 30 minutes, one hour and two hours before kickoff. So far, 30 minutes has produced the highest accuracy of the four windows at 47.7% across 107 picks.

It’s still negative overall, so we’re definitely not calling 30 minutes some magic window.

But this is what we mean when we talk about smarter betting.

It isn’t pretending we know exactly what’s going to happen, but testing the model against what happens, finding where it gets things wrong, looking for patterns in the data and using what we learn to make the next version better.

Week 4 was a good week, now we get to see if the model can build on it.

Dallas will host Tampa Bay to open Week 5, and the two offenses are coming into this in very different places.

The Cowboys are averaging 29.3 points and 330.3 yards per game, while Tampa Bay is averaging just 20.7 points and 279.3 yards. The Bucs are also still looking for their first win of the season and will once again be starting rookie Jalon Daniels in place of the injured Baker Mayfield.

For us, the first player to watch is Dak Prescott.

Dak is averaging 37.5 passing attempts and 26.2 completions per game through four weeks. His passing yardage has also climbed over his last three games, going from 279 yards against Washington to 276 against Baltimore and 335 against Houston last week.

Dak’s volume has been one of the more interesting things for us to track this season, especially because the market hasn’t always been easy to beat.

Last week was a good example. We had him under 35.5 attempts and 24.5 completions against Houston. He finished with 45 attempts and 32 completions, blowing through both lines as Dallas spent much of the game throwing.

It’s also a pretty good example of what we were talking about earlier.

Sometimes the model can have a reasonable read on how a quarterback should be used and still underestimate just how far the actual game moves away from that expectation. A few extra drives, a different game script or a defense that simply can’t get off the field can completely change a passing-volume prop.

Now we get another Dak game tomorrow night.

The lines will tell us where the market has adjusted after last week, and we’ll be watching to see whether there’s still an edge as the model gets another week of data to work with.

We do the modeling; the community sharpens the edge.

Before Thursday’s props go live, you can see which props are gaining attention on Trending Props and where the lines are shifting on the Boosts page.

Also before we go, the NBA preseason has started and we are fine-tuning the model we had last year, you can expect an update on it shortly.

Join the conversation on Discord to share slips, compare reads, and talk through the upcoming slate with the Propsguru community.

— Propsguru