Latest free predictions are the product output, dated when a signal last cleared the Edge Gate. The scan ledger is secondary proof for matches that did not clear the Edge Gate.
Main public product / Free Pick Window
Latest free predictions
Recent FATE predictions from the free window. The latest free prediction is dated when a signal last cleared the Edge Gate - which can be older than the latest scan, because most scans produce no edge. Full history is in the Prediction Ledger.
Loading latest prediction ledger rows...
$100 per pick simulation
Full Prediction Ledger
A simple flat-stake view of the full Prediction Ledger. Expected profit is what the model projected before the matches. Actual P/L is what happened after results.
Plain English
Scan. Filter. Predict.
FATE is a survivor forecasting engine. It checks football fixtures, compares model probability against the market, and only shows a free prediction when the signal survives the Edge Gate.
01 / Scan
Check the fixture
Teams, market, price, timing, and model probability are scored before anything appears as a prediction.
02 / Filter
Reject weak edges
No Edge and Filtered Out rows are not failures. They show FATE refused a weak or noisy call.
03 / Predict
Show the survivors
Surviving signals enter the Free Prediction Ledger with Hit, Miss, or Pending settlement labels.
Ask FATE
Get the plain-English answer.
Ask why a match was filtered out, what the latest free predictions are, or how the proof archive works.
Try: "Why no pick today?"
WHY NO PICK TODAY?SHOW LATEST FREE PREDICTIONSWHAT IS EDGE?WHAT IS EV?SHOW FLAT-STAKE RESULTSWHY TRUST FATE?WHAT DOES HIT/MISS MEAN?
Free Prediction Ledger
Full free prediction history
Every free prediction row stays labelled by source, result, probability, edge, and proof details.
Loading rows...
Flat-Stake Simulation
Expected profit is what the model projected before the matches. Actual P/L is what happened after results. This is a hypothetical flat-stake simulation, not a claim that bets were placed.
No real bets implied
Rows labelled Model History are proof/history rows, not live-before-kickoff receipts. Public/free rows use Free Prediction Ledger flat-stake simulation. This does not imply real bets were placed.
Loading prediction ledger...
Latest ScanLoading
Matches FATE scanned but filtered out because the edge was not strong enough.
Loading latest scan ledger rows...
Scan Ledger
Full scan history
Matches FATE scanned but filtered out because the edge was not strong enough.
Competitions are shown only with the evidence found locally. Current live scans are Premier League only; broader row-level validation covers the top five leagues. The UEFA Champions League, Europa League, and domestic cups are NOT live-scanned yet - research and validation data only, with no live model - so a major final such as the Champions League final is not scanned. They are never shown as covered until their fixture adapters and Model History rows are wired.
Coverage Status
FATE is not claiming all football coverage. Each competition is labelled by what is actually proven.
Competition
Live scan
Prediction ledger
Historical replay
Calibration proof
Betting edge proven
Status
Premier League
LIVE
YES
HISTORY
VALIDATION
NO EDGE CLAIM
LIVE
La Liga
-
-
HISTORY
VALIDATION
NO EDGE CLAIM
VALIDATION
Bundesliga
-
-
HISTORY
VALIDATION
NO EDGE CLAIM
VALIDATION
Serie A
-
-
HISTORY
VALIDATION
NO EDGE CLAIM
VALIDATION
Ligue 1
-
-
HISTORY
VALIDATION
NO EDGE CLAIM
VALIDATION
World Cup
-
-
CALIBRATION
CALIBRATION ONLY
NO ODDS / NO EDGE
CALIBRATION ONLYowner-preview
Euros
-
-
CALIBRATION
CALIBRATION ONLY
NO ODDS / NO EDGE
CALIBRATION ONLYowner-preview
UEFA Champions League
-
-
-
-
-
NOT BUILT
Europa / Conference League
-
-
-
-
-
NOT BUILT
FA Cup / League (Carabao) Cup
-
-
-
-
-
NOT BUILT
Friendlies / Internationals
-
-
RESEARCH
-
-
NOT BUILTresearch only
LIVE = live daily scan. HISTORY = row-level no-lookahead history. VALIDATION = scored validation archive (not live scan, not in the free ledger). CALIBRATION ONLY = probabilities only, no odds. NOT BUILT = no adapter / no row-level history yet. NO EDGE CLAIM / NO ODDS = no proven betting edge in any competition. Live per-league validation detail is below and at /fate?fate_api=coverage.
Loading coverage...
Proof Archive
Validation, public receipts, and model history.
Open the archive for row-level proof, validation receipts, and historical model evidence. Live-published receipts are archived separately for proof.
Free predictions are public during the Free Pick Window.
Daily automation creates scan reports and candidate drafts. Free-window ledger rows are public to view; live website receipts only mark picks that were posted before kickoff.
Notification fires only when FATE publishes the next edge pick. No newsletter spam, no fake cadence.
How credits work
1 credit = 1 prediction. Missed call? Credits come back.
FATE refunds the credits you spent on any prediction that settles as a miss. Refunds are issued as FATE credits — they are not cash, not USDT, and they cannot be cashed out or transferred. This is forecasting, not betting. Historical validation is not a guarantee of future results.
$1 / credit5-pack discount20-pack volume rateRefund-on-loss in creditsNo cash-out
TRENDING MARKETS
Example market questions. Public FATE picks remain football-first and signal-gated.
MODEL PERFORMANCE
Track record scored against real outcomes. Lower Brier = sharper calibration.
EVOLUTION OF THE ENGINE
FATE wasn't designed. It was selected.
Same engine. Same rules. No lookahead.Models that looked clever in theory died on contact with outcomes. Sharper ones replaced them.What you see is what survived.A prediction isn't a guess. It's what remains after the rest fails.
17
CANDIDATE MODELS
1,215
WALK-FORWARD MATCHES
dc_xg50
CHAMPION (LOWEST BRIER)
207
ACTIONABLE BETS
Tournament receipt — variants tested across V4 (xG-blended Dixon-Coles) and V5 (adaptive mutation). dc_xg50 selected post-hoc by lowest Brier with highest sample. These are historical validation-archive figures: dc_xg50 needs xG data that is not wired into the live app, so it is not currently re-runnable, and the live free-prediction lane uses a simpler Premier-League model. The 1,215 / 66% / 0.223 headline describes this validation archive, not today's live picks.
Open the Proof Archive →
See the 17-model tournament leaderboard, the row-level walk-forward archive (2,288 scored questions), and the hits, misses, and calibration receipts.
MARKET DISCIPLINE
Accuracy without odds context isn't credibility. FATE is scored against real market lines.
207
ACTIONABLE BETS
62.8%
BET-SUBSET WIN RATE
1.67
AVG DECIMAL ODDS
+0.71%
ROI · 10-UNIT STAKES
Across 1,215 walk-forward matches, the model logged 207 actionable bets at near-even market lines. We do not pad headline accuracy with low-payout locks. Drawdown is real and disclosed in the engine docs. This +0.71% is a small historical sample, not a forward promise: closing-line value (CLV) was negative across every tested variant - the model did not beat the closing line - so this is calibration and market-discipline evidence, NOT a proven betting edge and NOT a profit guarantee. FATE does not claim to beat the market.
Target near-even line: decimal ~1.90 (Asian net 0.90). Current bet-subset average: 1.67 — just inside the practical band [1.65, 2.20].
Asian Handicap and Total Goals are the natural next markets. The data is there; the model is being built. Per-match archive is being reconstructed from validation logs (row-level dc_xg50 outputs are not currently persisted).
PAST RESULTS - HISTORICAL VALIDATION
Walk-forward backtesting on real matches. Same engine. Same rules. No lookahead.
Free-window predictions, live website receipts, and historical validation are displayed separately. No fake track record blending.
First Pick Hit · 1 / 1 HIT
The next edge pick is the one to catch.
The live website receipt settled as a hit. Free-window predictions are visible in the Prediction Ledger; get notified when the next live-before-kickoff receipt appears.
You get one shot at every decision.
This engine gets thousands.We run it through history. Break it. Fix it. Calibrate it.
Until weak models die and sharper ones replace them.The market already has an opinion. We test where it is mispriced.That gap is the edge.
FATE AI BOTAsk about model logic, edge, scans, and free predictions.