FATE record
Performance — the honest ledger
Live pre-kickoff watches and archive/backtest results are separate ledgers and are never combined into one headline number.
Page served 2026-07-28 12:55 UTC · if this time is old, you are seeing a cached copy — hard-refresh for the live record.
REVIEW ONLY — not betting advice, no edge claimed, 0 PLAYABLE.
Live World Cup AH watches
Pre-kickoff receipts only; graded with exact Asian settlement.
Settled: 42 · Record: 25 full win · 2 half win · 4 push · 0 half loss · 11 full loss
Hit rate (pushes excluded): 71.1% (grade-only hit rate, NOT profit — without verified prices and with negative CLV this is not money)
Profit / loss: Not meaningful yet — 1/42 settled watches have a verified pre-kickoff price. Unpriced = the result is counted, but no exact same-side, same-line price was captured before kickoff, so it is never turned into dollars.
Dollars are computed ONLY over watches with a verified captured pre-kickoff market price for the exact side and line; the rest are counted in the record but never priced.
| Match | Watch | Result | Captured price | P/L per $100 |
|---|---|---|---|---|
| Australia vs Egypt | Australia +0 | push | HK 1.429 | +$0.0 |
Unpriced settled watches (counted above, never dollared): 41
90-minute winner lean (secondary)
Record: 68 won · 36 lost (65.4%)
No verified price ledger for 1X2 leans — no dollars computed.
Archive — Premier League walk-forward backtest
NOT live picks. Simulation on historical data, shown for calibration honesty.
Matches: 1215 · 1X2 accuracy: 66% · Brier: 0.223
ARCHIVE subset, CLV NEGATIVE: +$0.71 per $100 over 207 simulated flat-stake bets — historical simulation, not edge
Simulated flat-stake subset at balanced-odds lines from the walk-forward audit. Closing-line value was NEGATIVE across every tested variant — this is calibration evidence, NOT proven edge and NOT a forward promise.
tournament archive dc_xg50 (walk-forward champion): n=1215, n_bets=207 — same frozen figures the public /yesno receipt grid has shown since sprint 20260428_120729; posture frozen by the value-audit (reference-only, CLV negative).
Best-candidate policy — does it make money?
Honest answer from the policy backtest (live-priced, live-unpriced and archive kept separate; nothing mixed into one hit rate):
Did the odds-floor policy make money? NOT PROVEN. The live priced sample is tiny (1 priced of 42 settled), and the only archive figure is a CLV-negative simulation. There is not enough priced, market-referenced AH history to show the policy is profitable.
By market: Asian Handicap is the only supported candidate. Over/Under FAILED its backtest (unvalidated vs baseline) and is kept out of the candidate pool. 1X2 is a secondary lean only.
Pure model vs market-blend? Insufficient data to prove blending the model line helps — that would be blind tuning. FATE uses market-first candidate selection with a WAIT / MARKET FAR downgrade gate: it anchors to the real market main line and never claims an edge when far, unpriced, or below the floor.
Sample-size warning: these numbers are far too small to prove profit. No BET wording until the truth gate is genuinely passed — 0 PLAYABLE.
When would FATE ever say BET?
This record graduates to public BET wording only when ALL hold: live pre-kickoff published watches; at least 50 settled with verified captured prices; positive profit per $100 over that sample; no retroactive selection; verified closes. Until then every slip stays REVIEW ONLY.
← Back to the board · Main page · World Cup 2026 final report