A full Tuesday slate gives the model plenty to chew on, with the Knicks and Rockets drawing the widest gaps on the board. These are morning numbers only — the lock happens about an hour before tip.
These are morning leans, not locked plays. The official pick for each game locks roughly 60 minutes before tip, using whatever line is on the board at that moment. If a number moves between now and then, the locked pick can differ from what's written here, or it can vanish entirely. Only locked picks count toward the published record.
The model has a built-in governor. It only publishes a spread or total when its own number sits within 3 points of the market's. Anything wider than that isn't treated as a bigger opportunity — it's treated as the model missing something, an injury or a lineup wrinkle it doesn't know about. So those bigger gaps get held back instead of published.
The widest gap among published spreads belongs to Knicks-Bulls, where New York sits at -6.5 at -115 and the model's projection (118.78 to 109.35) implies closer to a 9.4-point Knicks margin — a 2.9-point lean in the Knicks' favor. Houston at Milwaukee isn't far behind: HOU -8.5 at -105, with the model projecting Houston by 11.1, a 2.6-point disagreement toward the Rockets.
Cleveland-Washington gives the model a smaller but still real lean. CLE -10.5 at -112 sits half a point shy of the model's projected 10.97-point Cavaliers margin, and the total leans UNDER 239.5 at -115 against a model number of 238.17. Detroit-Charlotte works similarly on both sides: DET -7.5 at -110 against a 9.19-point model margin, with the total leaning OVER 215.5 at -108 off a 217.4 projection.
Orlando at Toronto is about as tight as these leans get — ORL +4.5 at -110 is only 0.1 points off the model's number, and the UNDER 222.5 at -110 is a 0.3-point gap. Denver-New Orleans leans toward the Pelicans getting the points (NOP +4.5 at -110) with the OVER 240.5 at -115 riding a 241.92 projected total. San Antonio at Utah has the model further from the market than most, liking UTA +13.5 at -115 by 2.3 points and the OVER 233.5 at -115 by 1.7. Houston-Milwaukee's total also clears, OVER 225.5 at -110, as does the Golden State-Minnesota total, OVER 227.5 at -112, and the Lakers-Clippers total, OVER 229.5 at -110.
A good chunk of tonight's slate didn't clear the 3-point cap. Miami-Atlanta is held on both spread and total — a 7.7-point gap on the line and 3.3 on the total. Philadelphia-Indiana is held on both sides too, and the total gap there is a wide 8.4 points. New York-Chicago's total sits out even though the spread published, off by 3.4 points. Golden State-Minnesota's spread is held (4.8 points), Oklahoma City-Dallas is held on both spread and total (4.3 and 5.0 points), and the Lakers-Clippers spread stays off the board at 3.5 points. None of that means the model is wrong about those games — it means the model isn't confident enough in its own number to publish it.
Locked picks sit at 20-21, down 2.94 units, with spreads at 9-11 and totals at 11-10, across 56 games of ratings data this season. In the model's own backtesting, no slice of these picks has cleared the 52.4% break-even most books require. That makes this a market-level model, not a demonstrated edge — read it as one system's output, not financial advice.