Recent
W1I·KC @ BAL+0.56u|L1I·COL @ SF-0.61u|WBATTER_TB·Mariners @ Rays+0.38u|L1I·SEA @ TB-0.61u|WBATTER_TB·Braves @ Cardinals+0.41u|L1I·PHI @ DET-0.61u|WBATTER_TB·Phillies @ Tigers+0.73u|LBATTER_TB·Astros @ Rangers-0.65u|WBATTER_TB·Royals @ Orioles+0.24u|LBATTER_TB·Braves @ Cardinals-0.39u|W1I·KC @ BAL+0.56u|L1I·COL @ SF-0.61u|WBATTER_TB·Mariners @ Rays+0.38u|L1I·SEA @ TB-0.61u|WBATTER_TB·Braves @ Cardinals+0.41u|L1I·PHI @ DET-0.61u|WBATTER_TB·Phillies @ Tigers+0.73u|LBATTER_TB·Astros @ Rangers-0.65u|WBATTER_TB·Royals @ Orioles+0.24u|LBATTER_TB·Braves @ Cardinals-0.39u|
The Edge

Models

Active systems, pipeline markets, and the validation rules behind the card.

View results
Core engine
XGBoost + props sims
Market-specific model stack
Validation
Walk-forward CV
No future data leakage
Launch gate
200 bets
Per-system paper validation
Training Methodology

Models use walk-forward validation: trained on prior seasons and validated on the next season so the model never sees future data.

A pick has to clear model probability, implied probability, edge threshold, and Kelly-positive checks before it reaches the card.

Systems stay in paper mode until they clear the 200-bet gate. Results are tracked transparently before paid access opens.

Active systems
System
What it prices
Metric
Training
Detail
NRFI
v17
Predicts whether the first inning will be scoreless using starter quality, umpire zone, park, and weather context.
0.577
OOS AUC
-9.3%
2021-2025
HR
v6
Prices batter HR probability with barrel rate, launch angle, pitcher HR risk, park factor, and platoon split.
0.630
OOS AUC
-18.2%
2021-2025
F5
v5
Models first-five moneyline outcomes from starter SIERA, recent form, opponent wOBA, umpire, and weather.
0.553
OOS AUC
-9.9%
2022-2025
K
v1
Projects starter strikeout totals with swinging strikes, zone/chase rates, opponent K%, and recent K workload.
1.807
OOS MAE
+4.8%
2021-2025
OUTS
v1
Projects starter outs recorded from pitch efficiency, innings trends, bullpen context, and opponent OBP.
Proxy
IP model
+5.0%
2021-2025
NRFIOOS AUC: 0.577

Predicts whether the first inning will be scoreless using starter quality, umpire zone, park, and weather context.

v17 / 2021-2025-9.3% ROI
HROOS AUC: 0.630

Prices batter HR probability with barrel rate, launch angle, pitcher HR risk, park factor, and platoon split.

v6 / 2021-2025-18.2% ROI
F5OOS AUC: 0.553

Models first-five moneyline outcomes from starter SIERA, recent form, opponent wOBA, umpire, and weather.

v5 / 2022-2025-9.9% ROI
KOOS MAE: 1.807

Projects starter strikeout totals with swinging strikes, zone/chase rates, opponent K%, and recent K workload.

v1 / 2021-2025+4.8% ROI
OUTSIP model: Proxy

Projects starter outs recorded from pitch efficiency, innings trends, bullpen context, and opponent OBP.

v1 / 2021-2025+5.0% ROI
Pipeline models
GAME
Full Game

Moneyline and totals layer that extends F5 context into full-game pricing.

F3
First 3

Early-game starter window for openers, short leashes, and lineup-top exposure.

F1H
First Half

Hybrid innings window for stronger starter splits before bullpen noise dominates.

F7
First 7

Late starter and bridge-relief pricing before full bullpen exposure.

BATTER_K
Batter K

Batter strikeout props using pitcher shape, zone, chase, and batter whiff profile.

BATTER_TB
Total Bases

Batter total-base props with contact quality, matchup, lineup slot, and park.

BATTER_HITS
Hits

Hit props using contact rate, expected average, platoon split, and park/run context.

PITCHER_ER
Pitcher ER

Earned-runs props using starter quality, opponent run creation, weather, and leash.

Data sources
Statcast
Pitch rows, launch angle, exit velocity, spin, and contact quality.
MLB Stats API
Game results, innings, lineups, starters, and settlement.
Open-Meteo
Historical and forecast weather by stadium.
Umpire Scorecards
Zone accuracy, K boost, and run impact by umpire.