Direct answer: hit rate = hits ÷ settled games; void games (rainouts, cancellations) are excluded. Check the sample size, evaluation period and league too — past performance doesn't guarantee future results.
🔥 AI Prediction EngineHit rate Loading…Proprietary quant model · auto-settled dailyLocked before kickoff · transparently verified after🔥 AI Prediction EngineHit rate Loading…Proprietary quant model · auto-settled dailyLocked before kickoff · transparently verified after
Overall performance dashboard
Live stats
—
Elite-pick hit rate ★
— hits / — elite games
—
Elite-pick hit rate, last 45 days
— hits / — elite games
—
This week's elite-pick hit rate
— hits / — elite games
—
Current streak
Recent winning streak
Advanced Prediction Accuracy
Auto-settled per game
Beyond the moneyline win rate, the model also outputs correct-score, over/under and spread predictions — auto-settled and accumulated game by game.
—
🎯 Correct Score hit rate
Accumulating
—
📊 Over/Under hit rate
Accumulating
—
📐 Spread hit rate
Accumulating
✅ Verified correct-score hits
The system auto-settles each game; correct-score hits are added automatically. Below are human-verified historical examples (full hit rate shown above).
⚽ Ghana 1-0 Panama
⚽ Mexico 1-0 Korea Republic
⚽ Scotland 0-1 Morocco
⚽ Panama 0-1 Croatia
Last 20 prediction records
Auto-settled
Date
League
Home
Away
Predicted winner
Predicted probability
Actual result
Hit
Settled
Loading…
Monthly elite-pick hit-rate trend
Elite picks · last 3 months
Loading…
Each monthly figure is that month's real settled hit rate for Elite predictions, updated automatically as matches accumulate.
Analyst's fairness pledge
Transparency statement
Fairness pledge
Locked before tip-off:All predictions are published before the game starts, with unalterable timestamps.
Data & model version:Each prediction is generated from the match data available at the time; model updates may change performance across different periods, so results should be read alongside the date and sample size.
Uneditable pledge:Once published, predictions are locked — settled afterward without adjustment, never deleted, with full history preserved.
Honest settlement:Every hit and miss is listed — no cherry-picking, no padding.
Sample note:Covers MLB, NBA and football, excluding void games such as rainouts and forfeits.
All records are uneditable, locked before tip-off
FAQ
FAQ
How accurate are picks168 (EdgePicks) predictions? What's the hit rate?
The hit rate for high-confidence "elite" picks is based on the settled sample shown live on this track-record page; the overall settled hit rate is shown separately by sample period and league. Live figures follow the dashboard at the top of this page, and every record can be verified in the table below. No model can guarantee future results — past hit rates don't guarantee future performance.
How is the hit rate calculated? Are only the good picks counted?
No. Hit rate = (hits ÷ settled games); every hit and miss is included — nothing is cherry-picked or deleted. All predictions are locked before kickoff and can't be edited afterward, and void games such as rainouts or forfeits are excluded.
What is the Brier score?
The Brier score measures the accuracy of probability predictions, ranging from 0 to 1 — lower is better. It also penalizes predictions that are "overconfident but wrong," making it a better reflection of model calibration than raw hit rate alone; it's shown live on this page's dashboard.
Can past predictions be checked record by record?
Yes. This page keeps the full history of predictions — date, matchup, predicted direction, actual score and hit/miss — all public, locked before kickoff and never adjusted afterward. You're welcome to verify every record.