Draft season is basically here which means two things: your group chat is already toxic about draft order, and somebody is about talk themselves into a running back in round two because they liked him in college. I get it. I've been that guy. Before you lock in your cheat sheet, SportsLine just dropped their 2026 projections after running 10,000 simulations of the upcoming NFL season. There's real value in digging through the numbers.

The model has some credibility here. It called Daniel Jones as a breakout last year, which is the kind of call that makes you either trust the system completely or question everything you know about football. Probably a little of both. You don't nail a Daniel Jones prediction by playing it safe.

What I care about with these projection models isn't the overall rankings, because those mostly shake out close to consensus. The gold is in the gaps between ADP and projected finish. That's where sleepers live. A guy going in round eight who the model has finishing at round five value, that's your edge.

Busts matter just as much. Every year some name-brand player gets overdrafted based on what he did two years ago, and every year half your league watches their second-round pick finish outside the top 20 at his position. The model doesn't care about name value or your feelings about a guy's ceiling. It just runs the numbers 10,000 times and tells you what's probably going to happen.

Do I take model output as gospel? No. Deb would remind me that players are human beings, not data points, and she's right about that more often than I'd like to admit. Ray would say you can't quantify heart, which is exactly the kind of thing Ray says, and then I'll run the numbers and sometimes the old man is right anyway. Simulations are a tool. They don't replace watching games and reading injury reports.

But 10,000 runs is a serious sample. And the fact that this model hit on a weird call like Daniel Jones tells you it's finding patterns that draft rooms are missing. That's the whole game. If you're drafting off the same consensus board as everyone else, you're playing for a .500 season.

I'll be tracking my own calls all season like I always do, wins and losses. Last year I had some hits and some misses I'd rather not revisit. The process matters more than any single result, and stacking good process over a full season is how you end up holding a trophy in December.

Go read the full breakdown. Find the two or three names that don't match your gut, and figure out why the model sees something you don't. That tension between instinct and data is exactly where the best draft decisions come from.