+10.85. That is the points-of-spread edge our model gives Phoenix at Las Vegas tonight. The market has the Aces -8. The model has them -1.9 at neutral. We took the dog. The same number the Aces opened on the model's board yesterday. The same number the Aces will probably open on tomorrow.
Here is the thing about repeated edges across multiple days. They mean one of two things. Either the market is wrong, in which case we keep betting and we get rich. Or the model is wrong, in which case we keep betting and we get poor. The math does not care which one. The bankroll does.
The Liberty miss yesterday was the same architecture. Model said NYL -3.4. Market said NYL -11. We took CON +11. Liberty won by 31. The reason was specific. Sabally arrived in April. The model's prior was a 2025 Liberty without Sabally. The 2026 Liberty playing in their building on opening night had a fully integrated four-star core that the model could not yet see. We acknowledged this and recommended no STRONG LEANS on Liberty games for two weeks. That recommendation stands.
The Aces have the same problem. Two specific reasons.
First, what the model sees. The 2025 Las Vegas Aces had a net rating of +3.9 (4th in the league). They went 14-5 in clutch situations, the best clutch record in the WNBA. They were a top-five team. The model's prior is built on these numbers and they are accurate.
Second, what the model does not see. The 2026 Las Vegas Aces added Jewell Loyd in April on the wing. They added Chelsea Gray to run the offense. They added Brianna Turner for frontcourt depth. Chennedy Carter is on a training camp deal that will end with her in the rotation by Memorial Day. These are massive offseason additions. The starting five tonight is Wilson, Gray, Loyd, Young, and Turner with a bench that includes Carter. Combined TPV: 342 across the starting five. That is the highest five-player TPV in the league.
The market knows all of this. The market priced Vegas at -8 against Phoenix because the market is looking at the same five players the building is going to see tonight. Our model is looking at last year's Aces and saying they are evenly matched with this year's Mercury. Both can have a real basketball reason. The market's reason is based on a roster the players haven't actually played a game with yet. Our model's reason is based on data that does not include the new roster.
Here is what we are doing about it.
We are still placing the PHO +9 bet, but at half stake. The model has a +10 edge. That is too large to dismiss. The market may be overpricing the Aces brand. They could be a championship-favorite roster on paper that takes a month to integrate, the way the 2024 Phoenix Suns or 2023 Milwaukee Bucks needed real game time before the new pieces fit. If that integration is messy on opening night, Phoenix covers easily. The base case is also possible.
We are publicly downgrading the model's confidence on any Vegas game until at least May 24. By then we will have ten games of 2026 Aces data, which will let the rolling rating catch up to the actual team. Until then, treat any Vegas-related lean from this newsletter as a coin flip with a slight tilt rather than a STRONG LEAN.
We are adding the Aces to the same flagged-priors list as the Liberty. There are likely two or three more teams in this category, and we will identify them as the games come in. Likely candidates: New York, Las Vegas, Atlanta (added Reese in April), Indiana (locked Boston extension and added Cunningham). Anywhere a 2025 prior diverges sharply from the 2026 reality, the model will keep flashing edges that may or may not be real.
The honest part. We bet PHO +9 last night. The game was not on the slate. We did not get the chance to find out. We bet PHO +9 tonight at half stake. We will know how that went by morning. The 2025 backtest hit 57.6 percent over 151 picks, but every season starts with the priors slightly off, and 2026 has more roster turnover than any season I can remember. The first three weeks of the season are a calibration period. We are calibrating in public. The number that matters in October will not be tonight's record. It will be whether the model still says we are right after the priors have caught up.
Tonight's other two STRONG LEANS are different architecture. POR +5.5 is an expansion home dog where neither team has 2025 priors that matter. DAL UNDER 178.2 is a pace-and-volume bet that does not depend on roster integration. Those two we like at full stake. The Vegas-related lean we like at half. That is the discipline.
The picks scoreboard updates at /picks tomorrow morning regardless of how this lands. We do not hide anything. The model does not hide anything. The Vegas pattern is the second specific thing it has shown us in two days, and that is what makes the model worth running. It is also what makes it worth questioning out loud.
[ End Report ]
Share This