Dimers calls the under and Broncos ATS in TNF tight overtime win over Seahawks

Thursday's Broncos-Seahawks game was the NFL's fourth overtime finish in five days, decided 17-14 by a Wil Lutz field goal after Seattle rallied from 11 points down. The Dimers NFL predictions model projected a two-point game. It finished as a three-point game in overtime.

Two of our three picks cashed, including a best bet flagged by our Sweet Spot signal, and the third stayed alive into overtime. Denver's two touchdown scorers were also the two Broncos our model rated as most likely to find the end zone. 

Here's how every pick graded, and how the model read the game.

Two of three picks cashed in the thumb-wrestle Dimers projected

Dimers' NFL machine learning model weighs more than 100 data points for every game. Two dedicated models use that data to predict the margin and the total points, and on Thursday they pointed to a two-point game with 40 total points. 

Before kickoff, Seattle was a 2.5- to 3-point road favorite across the books on our NFL odds page, with the total at 40.5.

All three picks were in our Seahawks vs. Broncos preview, and here is how each one was graded.

The model's case for Seattle rested on the Seahawks' defense, their team rating, and Denver's quarterback situation. Seattle's defense was the one that showed up most, and it kept the game as a slugfest.

Broncos +3 was the top pick, backed by a 3.0% edge and an 11.8% ROI Sweet Spot range

Our best bet was Broncos +3.0 at -115 on Hard Rock Bet. The model gave Denver a 56.5% chance to cover at that number, a 3.0% edge over the price.

The play also carried a Sweet Spot signal, our flag for bets that fall inside an edge range and bet type where the model has delivered positive ROI over a sample. This spread fell into a range that has returned 11.8% ROI so far this season.

Denver won outright, so the pick cashed with room to spare. A Seattle win by one or two points, the result the 21-19 projection pointed to, would have cashed it too.

But one winning pick is not a meaningful sample, and Sweet Spot exists for that reason. It points you to the ranges where the model's edge has held up over time, rather than asking you to trust a single result. For this week's picks and Sweet Spot signals, see our NFL best bets.

The under cashed 9 points beneath the line in a tightly priced market

The model's 21-19 projection added up to 40 points, a half-point under the total, and the under was our pick. The market leaned the same way. The best prices, at Novig (50¢) and Polymarket, carried a 0% edge, meaning they matched the model's number exactly. Every other price carried a negative edge.

That's an aggressively priced total, and it's where shopping matters. Anywhere other than those two venues, you were paying more for the under than the model thought it was worth.

The game finished with 31 points, and only 28 by the end of regulation. Seattle kept Denver scoreless from RJ Harvey's 1-yard touchdown late in the second quarter until overtime, forcing six three-and-outs. Denver, playing without Patrick Surtain II and Talanoa Hufanga, held Seattle to three points through three quarters and sacked Sam Darnold five times.

Seahawks moneyline came down to one overtime interception

Our third pick, Seattle at a 58.2% win probability, lasted into overtime. Down 14-3 in the fourth quarter, the Seahawks got a 1-yard touchdown from George Holani and a 2-point pass to Rashid Shaheed, then tied it on Jason Myers' 34-yard field goal with three seconds left.

Seattle opened overtime with the ball, and with 8:25 left our live win probability tracker had the Seahawks at 52.3%. Then McMillian intercepted Darnold. By 6:14, Denver was at 81.0%, and Lutz's kick ended it with 4:53 left. To see every momentum swing as it happened, you can see the in-game probability chart in our Seahawks vs. Broncos prediction.

The TNF knuckle biter was the NFL's fourth overtime game in five days, after the Ravens vs. Falcons, Giants vs. Commanders, and the Bills vs. Rams in Week 5. 

Read more about how in-game probabilities can play out in: Dimers had Ravens and Giants dead in the fourth quarter.

Our top two Broncos TD predictions both found the end zone

J.K. Dobbins (27.9%) and RJ Harvey (25.4%) were the two Broncos the model rated most likely to score, and they accounted for both Denver touchdowns. Each had roughly a one-in-four chance, and both landed.

The miss was Jaxon Smith-Njigba, the highest touchdown probability in the game at 50.4%, who didn't score. He still beat his receiving projection by 39 yards after catching two passes for 11 yards in the first half.

Projections for every Week 6 player are on our NFL player projections. For our model's props edges, see our NFL props.

Next up: Cowboys at Packers on Sunday Night Football

NFL Week 6 continues Sunday, with the Dallas Cowboys at the Green Bay Packers on Sunday Night Football. Win probabilities, picks, and projected scores for every game are on our NFL predictions, and our straight-up winners can be found in our NFL Week 6 picks.

If you're looking for Sunday's tightest games, our model projects Cowboys-Packers and Panthers-Eagles as one-point games and Saints-Giants as a two-point game. Nick Slade breaks them down in the three Week 6 games our model says could go to overtime.

For every best bet, prop, and player projection across the full NFL slate, join Dimers Pro.

Seattle hosts the Kansas City Chiefs on Sunday Night Football on Oct. 25, and Denver hosts the Arizona Cardinals the same day.

Set a budget before you bet and stick to it. Tools and support are available in our responsible gambling hub.