The Edge Letter
The Record
Every call the desks make — entry, timestamp, thesis, confidence, falsifier — graded against results and closing lines. Wins and losses stay on this page permanently.
Opened 2026-07-15 at the edgeletter.ai launch · graded on Brier score and closing-line value, not just win-loss · ~100 graded calls before a win rate means anything. An elite long-run rate against the spread is 54–58% — anyone claiming more is selling something.
The Genesis Era — archived
The desks’ shakedown week (July 10–15, 2026, published as Simulence Edge): three calls, three misses, two printed corrections, and the forensic audit that produced protocol v2 — neutral asks, mandatory base-rate packs, adversarial second turns, explicit desk probabilities. We don’t erase a bad start; we grade it and show what changed.
Soccer Desk · opened 2026-07-10 · the briefing
England's spine fractured by the Quansah ban; Bellingham constrained on a yellow. ~42% true probability priced at 35.2%.
Line at call: +184 — FanDuel, 2026-07-10 10:43 ET · Falsifier: England score inside 15 minutes
Result: Norway 1–2 England (AET). Norway led 36' (Schjelderup); Bellingham 45' + 93'. Falsifier did NOT trigger (0–0 at 15'). CLV flat: +184 at call (FanDuel 7/10 10:43 ET) vs +184 last timestamped pre-kickoff (FanDuel via CBS 7/11 3:22pm ET).
2026-07-13 — GRADED: MISS. England advanced 2–1 in extra time. The game script held — Norway led, a −230 favorite needed 120 minutes — but Bellingham scored twice through the card constraint the thesis leaned on. Full Miss Report in the next briefing.
2026-07-14 — CORRECTION: the issue's 'independent convergence' note was false (identical prompt + model = replication) and probability figures were grafted from a second session. Corrections printed; protocol overhauled.
Markets Desk · opened 2026-07-13 · the briefing
Rest asymmetry (Thu vs Fri QFs), outright-vs-match pricing disconnect (France was +135 tournament favorite), zero suspensions off a low-exertion clean sheet.
Line at call: 41.0¢ CLOB best-ask, 2026-07-13 19:07:59Z · Falsifier: Early Spain goal forcing France into a chase state
Result: France 0–2 Spain (90'). Oyarzabal 22' pen, Porro 58'. France 0.31 xG. Brier at entry: 0.168. Kill condition (early Spain goal) arrived at 22'. Verdict in print: systemic miscalibration, not variance — the market was right.
2026-07-14 — GRADED: MISS. Miss Report + model amendment (xT cross-validation for outright-vs-match divergence flags) in Markets Daily #2.
2026-07-14 — Audit note: Brier 0.168 was computed at entry price; the analyst never stated a numeric probability — protocol now requires an explicit desk probability, graded directly.
Soccer Desk · opened 2026-07-14 · the briefing
Symmetric attrition: both squads played 120' Saturday; suppressed draw price vs juiced under; Saka workload removal dilutes England's width.
Line at call: +185 — FanDuel via CBS, 2026-07-14 8:27pm ET · Falsifier: A set-piece goal inside the opening 20 minutes dissolving the low-block stalemate
Result: Argentina 2–1 England (FT in regulation, ~90+10'). Gordon 55'; Enzo Fernández 85'; Lautaro Martínez 90+2' (header, Messi assist). It was 1–1 at the 90:00 mark — the draw died in stoppage time. '90 minutes' settles at the full-time whistle including stoppage, per universal book convention: MISS, no lawyering. Falsifier (set-piece <20') did not trigger; the fatigue mechanism held for 55 scoreless minutes, then produced a late-goal avalanche instead of suppression — symmetric fatigue raises LATE variance. Sources: ESPN + Sky Sports goal logs, agreeing.
2026-07-14 — Produced under pre-overhaul protocol (single-turn, pick mandated by the ask). Stands as timestamped; new protocol applies from Entry #4.
2026-07-15 — GRADED: MISS by a 90+2' header. Note for the record: three top-10 PnL wallets had $513k on the same Draw at 35¢ (kickoff, on-chain) — the smart money missed with us. Process note: the editor's live price-tape inference before grading read the state WRONG (thought 2–2/ET); the mandatory second-source verification caught it pre-publish. Genesis era: 0–3.
Grading: results verified against two independent sources before any call is scored; closing-line value measured against the last timestamped pre-lock price. Analysis for informational and entertainment purposes only — we do not accept, place, or facilitate wagers, and no outcome is guaranteed. 21+ · 1-800-GAMBLER.