0.5. The margin SEA covered the 11.5 spread by last night. The model bet was SEA +12 at the actual line; the picks file captured +10.7 yesterday morning but the line moved before we sized it. Either way the cover hit. Final score IND 89, SEA 78. The Storm lost by 11. SEA +11.5 cleared by half a point. SEA +12 cleared by one.
The bet that mattered tactically was the spread, not the margin. Nina wrote the tactical preview yesterday around the Storm's perimeter defense being the structural advantage the line was not pricing. The Storm held the Fever to 89 points, well below the projected total. Caitlin Clark was a story but not the story; she played a normal Caitlin Clark game and Indiana still won by 11 against a Storm team missing Magbegor.
That is the fifth application of the conviction architecture in 10 days.
Five for five.
PHO at half stake the previous Saturday. ATL at full stake on Tuesday. MIN at full stake on Wednesday. TOR at full stake on Friday. SEA at full stake last night. Same shape every time. Same outcome every time.
The framework is no longer "early signal." It is a real edge. We can plan around it. When the model has the road team favored, the agent has the road team winning the most-relevant late quarters, and the home team has a structural Q4 weakness, we take it. That is the bet. Has been since Saturday May 9.
Now the second big story from last night.
The Toronto Tempo won their first WNBA road game. Outright. By 10 points. Final TOR 106, LAS 96.
The half-stake call on TOR +8.5 cleared by 18.5. The Tempo did not just cover. They walked into the same building they lost in three days earlier and beat the home team by 10. Brondello's 72-hour adjustment was real. Whatever Julie Allemand did on the playmaker question got resolved well enough for the offense to function. The two-game-in-three-nights second-game dynamic that Jordan wrote about yesterday played out exactly as the historical data suggested.
Toronto is 2-2 on the first WNBA road trip. The franchise narrative is exactly what Jordan wrote about. The home crowd on May 22 will be loud. The conversation in the building is "what is the ceiling" not "did we miss our chance."
Now the third part of the recap. The part we missed.
We passed on three STRONG UNDER signals last night per the totals-model discipline framework (model was 4-6 going in, framework rule says do not size up on a model below breakeven). All three unders hit.
UNDER 177.3 SEA at IND. Game went 167. Cleared by 10. We passed. UNDER 170.1 CHI at MIN. Game went 165. Cleared by 5. We passed. UNDER 172.8 LVA at ATL. Game went 169. Cleared by 4. We passed.
If we had sized those three at half stake each, we add 1.5 units of profit to last night's ledger. We did not. The framework cost us those wins. The same way the framework cost us POR plus 11.3 on Monday two weeks ago when the Liberty injury read fired the discipline brake.
Two takeaways.
The pessimist read. The discipline cost money. Three more wins were available. Maya is too cautious with the framework brakes.
The optimist read. We added two full-stake conviction wins (SEA and our TOR cover) by being disciplined. The framework brakes saved us from sizing up on a model that was 4-6 going in. The win rate matters more than the win count. If we had sized up on the totals model when it was 4-6 and gone 4-8 instead, we would be down units. Last night the variance went the right way and the under signal hit. That does not retroactively change the math that said "do not size up on a 40 percent model."
I lean optimist this morning. Three thoughts.
First, the under signal that hit last night was not flat across the model. It hit because three specific games happened to play under by 4 to 10 points each. The CHI at MIN game was 165 vs a 170.1 line. That is the kind of game the model has nailed before. The cleaner read for the totals model is to size half stake on STRONG UNDER signals where we have a multi-quarter agent breakdown that confirms slow tempo (like CHI at MIN had). The blanket "model is 4-6, pass all unders" is too coarse.
Second, the pace correction work is now ahead of schedule. The model went 3-0 on STRONG UNDERS last night. That is statistically significant for a 14-game sample. Total STRONG ledger moves from 4-6 to 7-6, now 54 percent and above breakeven for the first time in two weeks. If tonight's CHI at MIN under signal is right (and the matchup looks right), we can size half stake on STRONG UNDERS going forward when the architecture confirms slow tempo on both sides.
Third, the conviction architecture for spreads is now the dominant signal. 5-for-5 over 10 days. Spread STRONG ledger moves from 7-4 to 9-4 (now 69 percent). When we get a sixth application tonight (WAS at DAL has it), we size full stake without hesitation. The framework is producing.
Cumulative through May 17. Spread STRONG: 9-4 (69 percent). Total STRONG: 7-6 (54 percent). Combined STRONG: 16-10 (62 percent). The model is now profitable on both tiers for the first time this season.
Tonight is small. Two games on the board.
Washington at Dallas. STRONG LEAN WAS plus 4.6. Edge plus 7.7. Same architectural shape as the five wins. DAL has the worst Q4 in the league. WAS just won outright in Indianapolis on Friday. The conviction architecture is here for a sixth application. Full stake. Nina has the tactical breakdown publishing in an hour.
Connecticut at Portland. PASS the spread (model edge inside the threshold). STRONG UNDER 174.8 with a minus 12.71 edge. Connecticut is missing Griner (Day-To-Day), Lacan, and Nelson-Ododa. Portland is a slow team that ranks bottom-five in pace. The matchup is the kind of slow-on-slow read that the totals model has gotten right when the structural conditions are present. Half stake on UNDER 174.8 per the new tier-downgrade reasoning above (totals model is now 54 percent and the matchup specifically favors slow pace).
Tonight's recommended action.
FULL STAKE: WAS plus 4.6 at Dallas. Sixth application of the conviction architecture.
HALF STAKE: UNDER 174.8 (CON at POR). Slow on slow, CON missing key players. Tier-downgrade per discipline.
PASS: CON at POR spread (model edge under the threshold).
Cumulative for the slate. One full, one half. Same shape Maya wrote two weeks ago about how the framework should size.
Talk tomorrow morning.
[ End Report ]
Share This