1. The number of bets we are sizing tonight. Seattle at Dallas, half stake on SEA plus 12.5. The conviction architecture is firing at its cleanest alignment of the season — every condition confirmed, the agent's Q4 read at HIGH tier — and it lands on Dallas three days after we instituted the exclusion rule.
The framework discipline answers the question by taking it to half stake. Not full. Not pass. Half.
The reasoning is the rule itself.
The Dallas exclusion was set on Friday after the LVA-at-DAL full-stake loss made the architecture's Dallas applications 2-2 across four tries. The rule said future calls with Dallas as the home team size down to half or pass. The rule did not say the architecture stops firing at Dallas. It said the variance from Bueckers's individual scoring runs has been large enough to neutralize the structural Q4 edge in roughly half of the Dallas applications.
Tonight is the first call where the rule gets to operate. The architecture wants full stake based on the alignment math (road favorite at neutral, agent confirms all four quarters with Q4 HIGH, home Q4 weakness at minus 3.0). The exclusion rule cuts that to half. Nina has the tactical piece this morning with the specific film and the specific Bueckers concern. The bankroll math here is the rule applied: half stake because the rule says half stake on Dallas applications, regardless of the alignment strength.
If the cover hits, the architecture's non-Dallas record stays 5-0 and the Dallas applications go to 3-2 at the half-stake unit count. If the cover loses, the architecture's non-Dallas record stays 5-0 and the Dallas applications go to 2-3 at the half-stake unit count. Either way, the rule has protected the bankroll from the variance that has historically beaten the architecture at Dallas. That is the entire point of the rule.
Yesterday in honest accounting.
Lone game on Sunday's slate. LVA at GSV. The bot card had the spread at PASS and the total at STRONG LEAN UNDER 167.5 (line moved to 169.6 by tipoff). We passed both per the framework discipline. The line moved against the under by 2.1 points between the morning capture and the actual market price at tip. Final score Las Vegas 91, Golden State 81. Game total 172. The UNDER lost regardless of which line you used (172 went over both 167.5 and 169.6).
Our actual sized exposure: zero units. The framework discipline saved a half-stake total loss that would have cost 0.9 units after juice. Nina's piece yesterday flagged the missing structural confirmation as the reason for the PASS. The framework worked.
The bot's published-bets ledger from yesterday: zero HIGH picks generated, the only graded outcome was the bot's STRONG LEAN total at UNDER which lost. That puts the bot's headline STRONG total record at 15-16, slightly below breakeven on the season.
Our editorial-sized totals: 9-9 still. We did not size the loss. The framework's structural-confirmation filter has now produced PASS calls on three consecutive STRONG UNDERS without confirmation across the last week (PHO at MIN, two on Wednesday, LVA at GSV on Sunday). All three were correct PASS calls. Bot would have gone 0-3 on those if we had sized them.
That is the filter doing the work it was built to do.
Tonight's framework reads.
SEA at DAL. Spread: half stake on SEA plus 12.5 per the Dallas exclusion rule applied to the architecture firing. Total: PASS (small edge per the card, no STRONG label, no structural confirmation either way).
MIN at PHO. Spread: PASS. The bot card has MODEST LEAN PHO at minus 3.4 with edge 2.95, below the STRONG threshold. The agent disagrees with the bot's lean — MIN Q3 plus 3.1 HIGH and MIN Q4 plus 1.7 medium. That is two agent quarter edges on the opposite side from the bot pick, which is below the 3+ HIGH threshold for the formal model-versus-agent conflict rule but close enough to suggest the bot's PHO read is shaky. Plus Napheesa Collier is Out for Minnesota, which thins the road team's star alignment and makes the existing read noisier. PASS by tier (MODEST) reinforced by the injury and the agent disagreement.
Total: MODEST LEAN UNDER 167.3. PASS by the framework default rule that MODEST UNDERS without structural confirmation are PASS.
Combined tonight: one bet sized. Half stake on SEA plus 12.5. Roughly 0.5 units of exposure.
The architecture's standing.
Non-Dallas applications: 5-for-5 (one hundred percent). The headline number that the exclusion rule was built to protect.
Dallas applications: 2-for-4 (fifty percent). The exclusion rule's reason for existing. Tonight's half-stake on SEA either makes that 3-for-5 at one-and-a-half full-equivalent units won or 2-for-5 at one full-equivalent unit lost. Either resolution is acceptable within the exclusion-rule framework.
The plan tomorrow.
Tuesday slate already has four games on the board per the data pull this morning. CHI at WAS, CON at ATL, LVA at LAS, POR at GSV. The agent reports and the picks file will be generated by the bot at 10:00 UTC tomorrow. Worth pulling for the Tuesday morning loop. ATL at home is a non-Dallas candidate for the architecture (their 2025 closing record was middling). Worth watching as the slate develops.
Talk tomorrow morning. The Dallas exclusion rule's first live test happens tonight at 7:30 PM ET. The discipline takes the half stake. The variance does the rest.
[ End Report ]
Share This