The models project Scheffler to finish 10th. They expect him to play well, contend, and possibly land inside the top 10 cleanly. What they do not see is value in the +550 winner price. The DTM (Difference to Market, the gap between Rithmm's model probability and the implied probability baked into the sportsbook odds) on Scheffler's outright bet is negative 78.5%, and the models have flagged it as outside the value window. Translation: he is the most likely individual winner, but the price is so heavily shaped by public handle that the math does not work.
That is not "fade Scheffler." That is the more precise read most data models make on heavy public favorites: the player is good, the price is wrong. Knowing the difference is what separates a real model from a coin-flip.
Here is what Rithmm's models are actually projecting at Shinnecock, where the real value is hiding, and how a data-driven approach handles the most beatable major championship of the year.
Scheffler's "play well, don't win" read is the most interesting projection on the board, but it is far from the only one. The models are surfacing a tight cluster of mid-tier names where the win probability and DTM both check out.
Jon Rahm at +1500 is the models' top winner pick, with a 7.4% win chance and an 18.1% DTM. The models project him at Finish 2. Rahm's combination of ball-striking, course history, and recent form fits Shinnecock's punishing setup cleanly, and the +1500 price does not reflect that as much as it should.
Matt Fitzpatrick at +2300 is next, with a 5.5% win chance but a 32.9% DTM, also projecting Finish 2. Fitzpatrick is the prototype of a US Open winner profile: precise ball-striker, elite around the green, comfortable on tough setups where mistakes punish severely. The DTM on him is one of the strongest positive signals on the entire board.
Xander Schauffele at +1900 rounds out the models' top tier with a 5.5% win chance and a 10.3% DTM, projecting Finish 10. He is a defensible pick but the value is smaller than on Rahm or Fitzpatrick.
The longshot tier is where the models get more interesting. Bryson DeChambeau at +4000 projects Finish 5. Sam Burns at +4000 projects Finish 4. Chris Gotterup at +4300 carries a 38.6% DTM. Russell Henley at +4000 carries a 25.4% DTM, projecting Finish 8. Viktor Hovland at +5700 carries a 40.4% DTM. Maverick McNealy at +7200 carries a 50.9% DTM, projecting Finish 19. Justin Rose at +7200 has a 65.5% DTM, the largest positive signal on the entire winner board.
That last one is worth pausing on. A 65.5% DTM on a +7200 price means the models see the actual probability as dramatically better than the line implies. Rose has the major championship pedigree, the ball-striking, and the recent form to make that read defensible. Whether he wins is one question. Whether the price is wrong is a different question, and the models are loud about their answer.
The winner market is one of four outright markets Rithmm runs at majors. The models also project Top 5, Top 10, and Top 20 finishes for every player in the field, and the value often lives in those secondary markets rather than the outright winner.
This matters because golf's variance is enormous on a single-week sample. A model that picks a US Open winner straight up at +1500 also picks the wrong winner most of the time, simply because golf is golf. A model that identifies which players are mispriced on top-10 and top-20 markets produces a much higher hit rate because the bet covers a broader range of outcomes.
Scheffler is the clearest example. The models' read on him is "Top 10 Strong, Projects Finish 10." That does not make him a good outright bet at +550. It makes him a much more interesting top-10 or top-20 play, where the price reflects the same projection without the variance tax of needing him to win the entire tournament.
When you look at Rithmm's bet card for any major, the value distribution shifts depending on which market you are in. Some players are values on the outright. Some are values only on the top-10 or top-20. Some are not values anywhere, which is also useful information.
Beyond the pre-tournament outrights, Rithmm has an in-round prediction tool called AI Caddie that produces 1,000 simulations per hole for every golfer in the field. The output is hole-by-hole projections: probability of eagle, birdie, par, bogey, or worse for each player on each hole, compared against the field average.
This is the live-betting layer most golf bettors never get access to. When Fitzpatrick steps to the first tee at Shinnecock Thursday morning, the AI Caddie projects him at 26% birdie probability and 60% par probability on Hole 1, versus a field average of 23% birdie and 53% par. Higher birdie probability, higher par probability, lower bogey risk than the field. That is the kind of hole-level read the public cannot see and the in-round markets often misprice.
Over 72 holes of a major championship, that compounds. Bettors using AI Caddie are working with hole-by-hole probability shifts the books cannot keep up with in real time, which is exactly where in-round prop value lives in golf.
A few things about this week make the models' reads sharper than usual.
Shinnecock is set up to play par 70 with notoriously narrow corridors, deep rough, and treacherous greens. The forecast has stiff winds Thursday and Friday. The USGA is expected to lean into the difficulty all four rounds. Under those conditions, the player who plays a clean, mistake-free week beats the player with more pure talent, which is exactly the dynamic that produced Wyndham Clark at 100-1 in 2023 and J.J. Spaun at 150-1 in 2025.
The models are built to identify that profile of golfer in advance, which is why names like Henley, Conners, Lowry, and Rose are surfacing alongside the bigger names. Course fit at a brutal US Open setup matters more than seasonal form, and the models are reading both.
The public will pile onto Scheffler, McIlroy, and the names they recognize. The books know that and price accordingly. That is the structural setup that creates value across the rest of the field, and Rithmm's models are built to find it.
The cleaner approach for a major championship like this one uses Rithmm's full bet card rather than chasing outright winners. A read on Scheffler's top-10 likelihood, anchored by the models' Finish 10 projection. A position on Rahm or Fitzpatrick at outright winner prices the models see as undervalued. Top-20 positions on the strong-DTM longshots like McNealy, Rose, and Hovland where the value is concentrated. Plus AI Caddie reads on the live in-round markets across the four days of the tournament.
That is not betting the favorite hoping for the best. That is running a major championship as a portfolio of mispriced markets, which is exactly how the bettors who consistently profit on golf actually approach the sport.
Every winner, top 5, top 10, and top 20 market at the US Open is live in the Rithmm app right now, with the models' projected finish, win chance, and DTM on each player. The AI Caddie goes live for in-round play once the tournament starts Thursday morning at 6:45 AM ET, producing hole-by-hole simulations across all 18 holes for every golfer who makes the cut.
If you have been searching for the best AI golf predictions for the US Open and want to see what the models are actually showing instead of what the public is shouting, this is the week.
The full Rithmm subscription is $29.99 a month, and it does not stop at golf. The same AI models powering every Shinnecock projection run across eight sports total: NFL, NBA, WNBA, MLB, PGA golf, World Cup soccer, College Football, and NCAA men's basketball. That is year-round coverage of the biggest betting markets in the world, all in one app, with predictions and DTM signals updating across every slate every day. Props and Game Lines included.
Start the 7-day free trial today, run the models across the full Shinnecock field and you will be ready when the first tee time hits Thursday morning.
