Top 5 Bets · Stability ≥90% ⓘ
Loading…
Recent Form · Last 10 Games
Loading recent form…
📅 This Week & Next
Loading schedule…
👤 Player
⚾ Opposing Pitcher
Career vs This Pitcher
Pitch Mix — adjust if needed
Pitch Mix (approx %)
📅 Game Info
Pick the opponent — Home games track them into PHX, Away games track ARI into their city.
Home Plate Umpire
Enter a game date to look up the umpire.
Lineup Protection (3-4-5 Spots)
Enter a game date to load lineup context.
🌤 Stadium & Weather
Live Conditions at Stadium
Complete the Setup and hit Run Prediction first.
📊 Performance Score
—
/100
🏃 Batter
—
⚾ Pitcher
—
🏃 Batter Factors ▼
⚾ Pitcher Factors ▼
🌤 Conditions ▼
📝 Scouting Report
Run a prediction to generate the scouting report.
⚾ Pitch Mix
🎰
SNAKE SAVANT CorBET
Model-aligned prop recommendations · Odds API
Run a prediction first to generate prop recommendations.
📊
KALSHI EDGES
No-vig prediction-market YES prob vs model · Kalshi exchange
📈 CorBET Record
Units
—
ROI
—
Hit Rate
—
Record
0-0
No bets logged yet. Run predictions and save CorBET picks to build your record.
📊 Model Calibration
Calibration data builds up as you grade bets. New bets save model probability and MC confidence — older bets without these fields are excluded.
LIVE MODEL CORRECTIONS
Auto-applied to new predictions. Each prop fits its own correction once enough graded bets accumulate; below threshold it stays uncorrected (or uses the pooled fit).
CALIBRATION BY PREDICTED PROBABILITY
If the model is well-calibrated, "Avg Pred" should match "Hit Rate" in each bucket.
MC CONFIDENCE THRESHOLD
85%+ is the Top-3 cutoff. Higher MC should mean higher hit rate.
PER-PROP TYPE BREAKDOWN
Identifies which prop markets the model handles well vs poorly.
📈 2026 Splits
Select a player in Setup to load splits.
📊 Advanced Stats · 2026
Slash Line
Plate Discipline
Power & Contact
Runners in Scoring Position
Statcast / Advanced
Loading Statcast data...
Select a player in Setup to load stats.
📋 Pitcher Analysis
Search for a pitcher in Setup to see their full breakdown.
Season Stats
Pitch Mix
Recent Form
Workload
Statcast (Pitcher)
Statcast data loads after pitcher is selected.
⚡ Pending Grades
No predictions waiting to be graded. Run a prediction before a game and come back after to grade it.
🧠 Model Learning
Grade at least 5 games to start seeing model insights.
Each factor's hit rate = % of times the player outperformed baseline when that factor fired. Model auto-adjusts weights after 15+ games.
🎯 Per-Player Accuracy
Grade games to see how accurately the model rates each player.
How well the model's score predicts each player's actual game.
Avg — the player's average prediction score (0–100).
Good% — how often they actually had a good game (perf ≥ 40), across every game.
Bull% — only the games the model was high on (score ≥ 60): how often the player then delivered a good game. Answers "when Carroll gets a 96, does he show up?"
Acc% — across all games, how often the model's call matched reality — this also rewards correctly predicting a bad game, not just a good one.
▸ Tap a player to see which factors most inflated the model's score on their bad games.
Avg — the player's average prediction score (0–100).
Good% — how often they actually had a good game (perf ≥ 40), across every game.
Bull% — only the games the model was high on (score ≥ 60): how often the player then delivered a good game. Answers "when Carroll gets a 96, does he show up?"
Acc% — across all games, how often the model's call matched reality — this also rewards correctly predicting a bad game, not just a good one.
▸ Tap a player to see which factors most inflated the model's score on their bad games.
📈 Prediction Score vs Actual Performance
Grade games to see prediction accuracy over time.
Green = performance score (0–100) · Red dashed = model prediction score · Closer lines = better calibration
📋 Graded Game Log
No graded games yet.
DateScrPlayerActual LinePerfOutcomeModel