⚠ Important: All picks are for informational and entertainment purposes only.
We make no guarantees of profit. Past performance does not predict future results.
Bet responsibly. 21+.
The 6-Layer Analysis System
Every MLB game goes through 6 independent scoring layers. Each layer is scored and normalized.
The final score is a weighted composite of available layers — if a data source is unavailable,
the score is normalized to the layers that did run (no layer is assumed to be positive or negative by default).
Only games above a minimum threshold are classified as picks. Games that score below the threshold
appear as "WATCH" or are excluded from the daily board.
📡
Layer 1 — Market & Odds (0–25 pts)
Real-time odds from 7+ sportsbooks via TheOddsAPI. Tracks line movement, steam, reverse line movement, and pricing inefficiencies. Games with extreme prices (-250+) are capped to prevent value distortion.
🔬
Layer 2 — Sharp / Public (0–20 pts)
Action Network Pro betting data. Analyzes money percentage, ticket percentage, and divergence between sharp professional bettors and the public. If this data is unavailable, the layer shows as "pending" and the score is normalized without it.
⚾
Layer 3 — Pitching (0–20 pts)
Probable starters from MLB Stats API (public). Scores pitcher ERA, WHIP, recent form, and strikeout rate. If no starter is announced, the layer shows as "TBD" and the game is eligible for a lower confidence pick only.
📊
Layer 4 — Situational (0–15 pts)
Recent form (last 10 games), home/away record, bullpen density (days of rest, recent usage), travel schedule, and winning/losing streak. Data is sourced from MLB Stats API.
🌤️
Layer 5 — Weather & Park (0–10 pts)
Open-Meteo API for real-time weather at all 30 MLB stadiums. Wind speed and direction are scored for hitter/pitcher advantage. Temperature and dome/retractable roof status are factored in.
🤖
Layer 6 — AI Consensus (0–10 pts)
GPT-4o, Claude, and Grok each independently review the top-scoring picks. A positive consensus across models adds confidence points. This layer runs on-demand and may not be included in every pick.
Score Classification
- PREMIUM PICK (80–100): Strongest signal across most or all layers. Highest unit suggestion.
- STANDARD PICK (65–79): Good signal across multiple layers. Normal unit suggestion.
- LEAN (50–64): Moderate signal, partial data, or mixed layers. Low unit suggestion.
- WATCH / NO BET (below 50): Insufficient signal. Not published as a pick.
Hard caps apply: games without a confirmed starter are capped at 69/100. Games with heavy price (-250+) are flagged regardless of score.
Units & Bankroll Management
A unit is a standardized bet size relative to your bankroll. We recommend treating 1 unit as 1–2% of your total betting bankroll. This protects you from variance and allows for sustainable long-term play.
- 1u = standard low-confidence play
- 2u = standard confident play
- 3u = high-confidence, rare designation
Never bet more than you can afford to lose. Set daily and weekly limits. Do not chase losses. This service is informational — your betting decisions are entirely your own responsibility.
What We Don't Guarantee
- We do not guarantee any pick will win.
- Odds and lines change — always verify before placing a wager.
- Data may be delayed, incomplete, or inaccurate.
- A high score reflects data signal strength, not a guaranteed outcome.
- Past performance of the system does not predict future results.
Data Sources
We use publicly available and licensed data sources. We are not affiliated with MLB, any MLB team, sportsbooks, or data providers unless explicitly stated. All analysis is our own derivative work.
- MLB Stats API (public, no key)
- TheOddsAPI (licensed)
- Action Network Pro (licensed)
- Open-Meteo (public)
- OpenAI / Anthropic / xAI APIs (licensed)