NFL Today's Best Bets (September 17): No Model Board for Lions at Bills on Thursday Night

Week 2 of the NFL season opens in Orchard Park, where the Bills host the Lions at 8:15 p.m. ET at Highmark Stadium. The market has made Buffalo a clear favorite: 4.5 points, minus 240 on the moneyline against plus 198 for Detroit, with a total of 54.5. This column exists to set Verdexed's own model probability beside that price and report the gap between them. Thursday it cannot. The verdexed.com prediction, edge, and accuracy pages are blocked from this publishing environment, and per-game win probabilities, confidence tiers, and edge figures were not retrievable through search either. There is no biggest edge to name tonight, and no probability printed below belongs to the model. What follows is a matchup and market read, published as model output and entertainment rather than betting advice, for readers 21 and older located where sports wagering is legal.

How to read this

A normal edition of this column carries two numbers per game. The first is Verdexed's model probability, built from a weighted feature set per matchup (more than a dozen inputs covering recent form, team strength and scoring differential, home-field advantage, head-to-head history, and rest) and resolved through ten thousand simulations of the game. The second is the market's own probability, taken from the current moneyline with the sportsbook's margin stripped out. The edge is the distance between the two, and it means something only when the model half is formed independently of the market half. Tonight only the market half was readable. Verdexed logs every prediction before kickoff and scores it afterward on win accuracy and on Brier score, where lower is better, but the trailing thirty-day figures were not retrievable at publication, so no hit rate is cited here and none should be inferred. The standing caveat holds either way: a calibrated 68 percent side is built to lose roughly one game in three, and no edge is a guarantee.

The model's top edges today

There are none to publish. An edge is the difference between a model probability and a market probability observed at the same moment, and tonight only the second of those was available. A thin slate and a slate whose model numbers could not be retrieved are different findings, and this is the second, continuing the retrieval failure that has run through this week's editions. Nothing below is a pick, a rated play, or a confidence tier.

Thursday night, market side only

**Lions at Bills, 8:15 p.m. ET, Highmark Stadium.** Buffalo is listed at minus 240 and Detroit at plus 198. Removing the sportsbook's margin from that pair puts the market's own probability on the Bills at roughly 68 percent, which is the reference point a model opinion would be measured against rather than a projection in its own right. The spread is 4.5 and the total is 54.5, with the over priced at minus 114 and the under at minus 106, close enough to even money on both sides to suggest the books see no strong lean on scoring.

The clearest reported advantage in this game is structural rather than numerical, and it sits on Detroit's offensive line. The Lions have ruled out tackle Blake Miller (knee) and guard Christian Mahogany (hip), which puts backups into two starting spots against a Buffalo front that gets to rush with a lead if the market is right. Cornerback D.J. Reed is listed as questionable with a foot injury, and questionable is a status rather than an absence: it should be treated as unresolved until Detroit's game-day designation lands. Buffalo's own questionable list runs to safety Cole Bishop (knee and groin), running back Ty Johnson (hamstring), and defensive lineman T.J. Sanders (knee and illness). None of those three has been ruled out, and reading a questionable tag as an absence is one of the more common ways a pregame read goes wrong.

One further caution applies to any model number in Week 2, Verdexed's included once it is readable again. Two weeks into a season a model has almost no current-year signal to work with, so its output leans heavily on preseason priors and on last year's team strength. That is the correct way to build it, but it means early-season probabilities move sharply as real games accumulate, and a confidence tier in September deserves more skepticism than the same tier in November.

What could go wrong

On a one-game slate the risk is concentrated rather than spread out. A late downgrade on any of the four questionable players changes the matchup read, and Thursday games give both teams less practice and less recovery time than a normal week, which tends to show up as sloppier execution. Detroit's two offensive line absences are the specific thing to watch: if Buffalo's pass rush wins that matchup early, the game and the total can move in the same direction quickly. Cutting the other way, a short week and a mid-sized home spread is a familiar setup for a game that stays closer than the price implies without the favorite ever actually losing. And the standing risk tonight is the one named at the top: with no model board retrievable, there is no independent number here to calibrate the market against.

Responsible play

**This is model output, not betting advice.** Verdexed publishes its model's probabilities and edges for informational and entertainment purposes only. Nothing here is a recommendation to place a wager, and no outcome is guaranteed, even the model's highest-confidence picks lose regularly.

**You must be of legal betting age in your jurisdiction** and located where sports wagering is legal to bet. Sports betting is not available everywhere; know and follow the laws where you are. Never bet more than you can afford to lose, and never chase losses.

If gambling is affecting you or someone you know, help is available. In the US, call or text **1-800-GAMBLER** (1-800-426-2537), or visit ncpgambling.org. Please play responsibly.