0. The number of editorial bets we sized across the five-day gap from June 13 through June 17. The bot ran. The cards graded. The framework was not applied to the slate.
The honest accounting requires reconstruction.
June 12 — the last loop we ran before the gap.
The half stake on Toronto minus 2.3 at Washington produced a one-point Tempo road win. Toronto won 86 to 85 in a game that came down to the final possession. Toronto minus 2.3 covered by 1.5 points. The half-stake bet won at 0.45 units after juice. The architecture's softer firing (three medium quarter signals from the agent rather than HIGH-tier confirmation) was vindicated at the partial-confirmation sizing.
The other game on the June 12 slate (GSV at SEA) produced a Golden State four-point road win as a 9.1-point underdog. The card's STRONG LEAN on Seattle covered by 5.1 points. The PASS missed approximately 0.45 units. Net for the night: zero variance change because the half-stake gain on TOR offset the missed-win on GSV-SEA.
Issue 39's framework reading was correct. The half-stake sizing on TOR was the right call. The PASS on the partial-confirmation GSV-SEA was inconsequential at the unit-count level.
June 13 through June 17 — the slate without us.
The bot's STRONG-tier spread record across the five-day gap.
June 13. POR at +5.9 won at home against DAL. Bot covered. DAL won 84 to 83 by one point; POR plus 5.9 won by 4.9 points. MIN at minus 2.6 lost as the road MODEST LEAN; bot was on the wrong side of a Las Vegas home win. Spread record: 1-1 across the day's two relevant STRONG-or-MODEST tier cards.
June 14. WAS at +11.9 lost at New York. The Liberty won 86 to 64 by 22 points; Washington plus 11.9 lost by 10.1 points. Bot's STRONG LEAN WAS was correct in direction but missed by ten points beyond the line. Spread record: 0-1 on the day's STRONG card.
June 15. LVA at +2.7 lost at Dallas. Dallas won 96 to 66 by 30 points; LVA plus 2.7 lost by 27.3 points. The Dallas exclusion rule the framework keeps in effect produced the correct PASS here. The bot's STRONG card without the exclusion rule would have produced a brutal full-stake loss. LAS at -5.3 lost at Golden State as a MODEST LEAN; not a STRONG card. Spread record: 0-1 on the day's STRONG card.
June 16. TOR at -7.2 lost at Indiana. The Fever won 113 to 91 by 22 points; Toronto minus 7.2 lost by 29.2 points. The Tempo's worst road blowout of the season produced the bot's worst STRONG-tier spread loss of the gap. Spread record: 0-1 on the day's STRONG card.
June 17. LAS at +9.0 lost at Minnesota. The Lynx won 99 to 83 by 16 points; LAS plus 9.0 lost by 7. SEA at -3.5 lost at Portland. Portland won 94 to 89 by 5 points; SEA minus 3.5 lost by 8.5 points. Bot's two STRONG cards both lost. Spread record: 0-2 on the day.
Combined: bot STRONG-tier spreads went 1-for-5 (20 percent) across the gap. Three of the four losses were at double-digit ATS margins (WAS by 10, LVA by 27, TOR by 22). The bot's STRONG-tier season win rate has now drifted to roughly 18-22 (45 percent), well below the backtested 57.9 percent baseline.
The bot's STRONG-tier UNDER record across the gap.
June 13. LAS at PHO UNDER 177.4 lost (game went 213). MIN at LVA UNDER 174.8 lost (game went 197). Both UNDERS lost by 18-22 points beyond the line.
June 14. ATL at TOR UNDER 172.3 lost (game went 179). The Tempo home win produced more offensive volume than the structural read projected.
June 15. LAS at GSV UNDER 173.5 won (game went 136). LVA at DAL UNDER 177.8 won (game went 162). Both UNDERS hit by 15-37 points.
June 16. TOR at IND UNDER 177.7 lost (game went 204). The Tempo road blowout produced a high-pace, high-volume game that broke the structural read.
June 17. LVA at PHO UNDER 172.5 won (game went 162). MIN at LAS UNDER 176.5 lost (game went 182). One W, one L.
Combined: bot STRONG-tier UNDERS went 3-for-5 (60 percent) across the gap. The under-side record is meaningfully better than the spread side over the same stretch.
The hypothetical editorial-sized record.
The framework's discipline pattern across the gap would have applied the partial-confirmation rule to most of the bot's STRONG cards. Without agent reports from June 14 through June 16 (the betting model's --morning orchestrator only ran on June 13 and June 17 cleanly), the silent-agent default would have produced PASS calls on most of the spread tier.
Counting the hypothetical: PASS on five of five bot STRONG spread cards saves five half-stake equivalent bets. The bot lost four of those five. The PASS discipline saved approximately 2 units of variance avoided against four losses. The single bot STRONG spread win (POR on June 13) cost the framework approximately 0.45 units of missed profit.
Net hypothetical spread side: roughly plus 1.55 units of variance saved.
For the UNDERS: the new loosened rule (half stake on either structural condition) applied to five STRONG UNDERS across the gap. Three won, two lost. If the framework had sized all five at half stake under the loosened rule, the record would have been plus 0.5 units (3 wins × 0.45 unit profit minus 2 losses × 0.5 unit loss equals plus 0.35 net).
Net hypothetical under side: roughly plus 0.35 units.
Combined hypothetical for the gap: approximately plus 1.9 units. The framework discipline applied across the gap would have produced a meaningfully positive ledger.
Actual editorial-sized ledger across the gap: zero units sized, zero units changed. We were not here.
The architecture firings across the gap.
The architecture did not produce a clean firing on any of the gap nights. The closest was June 13 (LAS at PHO had a model edge on PHO but no agent quarter alignment). The June 15 LVA at DAL was a Dallas application that the exclusion rule would have produced a PASS or half-stake on, and the bot lost by 27. The exclusion rule's threshold review pattern (the next two Dallas applications after the PHO vindication on June 11) is now 0-1 in the post-PHO sample. Two more Dallas applications needed to reach the review threshold.
The conflict rule across the gap.
The conflict rule did not produce a clean firing on any of the gap nights. The agent reports were sparse (only June 13 and June 17 had complete reports), so the formal three-HIGH threshold was not testable on most slates. The rule's record remains 4-3 going into tonight.
Tonight is a small slate.
One game on the WNBA board tonight. Atlanta at Indiana. The bot's picks file has zero HIGH calls. The card has MODEST LEAN on Atlanta minus 1.1 with edge plus 3.75 (PASS by tier) and STRONG LEAN UNDER 173.5 with edge minus 10.24.
The STRONG UNDER 173.5 is the first borderline test of the loosened UNDER rule. Both teams are cleanly thinned by injury — Atlanta missing Brionna Jones and Aaliyah Nye (two rotation pieces), Indiana with Caitlin Clark Day-To-Day. That is the cleanest injury condition any STRONG UNDER has produced since the rule changed on June 9.
The rule says half stake when either structural condition is met cleanly. The injury condition is met cleanly.
But the agent disagrees on the direction.
The agent's projected quarter totals add to 176.5 points, three points OVER the market line of 173.5. The SGP correlation idea includes the game total going OVER the line's low end. The agent's two HIGH quarter signals are one each side (IND Q1 HIGH plus 5.10 and ATL Q4 HIGH plus 7.80), which structurally reads as both teams producing meaningful runs in their preferred quarters rather than the compressed-scoring game the card model assumes.
The new rule did not predict the case where the agent points opposite to the card model. Nina has the tactical piece this morning with the specific case for and case against the half-stake sizing. The framework discipline reads the agent's disagreement as the rule's missing condition and PASSES the under tonight.
The rule maintenance going forward.
The loosened UNDER rule needs an agent-direction filter added. The proposed shape: half stake when at least one structural condition is met AND the agent's projected quarter totals are within five points of the market line. If the agent points OVER by three-plus points, the rule pauses to PASS.
That refinement will get drafted this weekend after tonight's result lands. If tonight's UNDER hits, the agent's disagreement signal is the wrong filter and the rule's structural read held. If tonight's OVER hits, the agent's read was correct and the filter is the right addition to the rule.
The cumulative ledger across the season.
Spread STRONG editorial-sized: 13-13 entering tonight (the TOR win on June 12 brought the editorial line back to even). The discipline pattern has saved approximately 7 units of variance across two-plus weeks of correct PASS calls.
Total STRONG editorial-sized: 9-9. No bets sized since May 28. The loosened UNDER rule has not produced a sized bet because the rule's first borderline test tonight produces a PASS by the agent-direction filter.
The architecture's standing: 5-0 on non-Dallas applications, 2-3 on Dallas at the rule level (3-3 unofficially if we count the PHO miss).
The conflict rule's standing: 4-3.
Tonight: zero exposure. PASS the slate. The discipline holds. The rule-maintenance work happens this weekend.
Talk tomorrow.
[ End Report ]
Share This