Diamond AI API v1

Get all predictions for a specific date. Includes Claude AI analysis, model probabilities, confidence scores, value edges, and reasoning.

Parameters

Example Response

{
  "date": "2026-04-06",
  "model_version": "1.0",
  "count": 13,
  "predictions": [
    {
      "game_id": 352,
      "home_team": "COL",
      "away_team": "HOU",
      "predicted_winner": "HOU",
      "predicted_home_score": 5,
      "predicted_away_score": 8,
      "confidence": 64,
      "confidence_level": "medium",
      "model_probability": 0.677,
      "vegas_probability": 0.648,
      "consensus_probability": 0.662,
      "value_edge": 5.5,
      "reasoning": "Houston's rotation advantage...",
      "key_factors": ["Elo gap: +85 HOU", "Coors Field effect"]
    }
  ]
}

Get a raw statistical model prediction for any matchup. Uses current Elo ratings and logistic regression. No Claude AI, pure math. No game needs to exist in the database.

Parameters

Example Response

{
  "matchup": {
    "home": {
      "abbreviation": "NYY",
      "name": "New York Yankees",
      "elo": 1567.3,
      "momentum": 12.4
    },
    "away": {
      "abbreviation": "BOS",
      "name": "Boston Red Sox",
      "elo": 1523.1,
      "momentum": -3.2
    }
  },
  "prediction": {
    "predicted_winner": "NYY",
    "home_win_probability": 0.5892,
    "away_win_probability": 0.4108,
    "confidence": 58.9
  },
  "features": {
    "elo_diff": 68.2,
    "pitcher_diff": 0.0,
    "park_adj": 0.049
  }
}

All 30 MLB teams ranked by Elo rating. Includes peak/floor Elo, recent form (last 5), momentum (last 10 Elo delta), and tier classification.

Parameters

Tier Definitions

Example Response

{
  "season": 2026,
  "count": 30,
  "baseline_elo": 1500,
  "rankings": [
    {
      "rank": 1,
      "team": "LAD",
      "name": "Los Angeles Dodgers",
      "elo": 1598.2,
      "tier": "elite",
      "momentum": 8.3,
      "last_5": "4-1"
    }
  ]
}

Games where Diamond AI's model probability diverges from Vegas lines. Sorted by edge size (best value first). Updated daily with new predictions.

Parameters

Edge Ratings

Example Response

{
  "date": "2026-04-06",
  "count": 5,
  "value_bets": [
    {
      "game_id": 343,
      "home_team": "CLE",
      "away_team": "KC",
      "model_home_probability": 0.638,
      "vegas_home_probability": 0.520,
      "value_side": "home",
      "edge_pct": 11.8,
      "rating": "strong"
    }
  ],
  "disclaimer": "For entertainment and informational purposes only."
}

Model performance metrics: overall accuracy, Brier score, calibration curve, performance by confidence level, and recent daily form. Compare against benchmarks (coin flip, Vegas consensus).

Parameters

Key Metrics

Example Response

{
  "season": 2026,
  "overall": {
    "total_predictions": 156,
    "correct": 91,
    "accuracy_pct": 58.3,
    "brier_score": 0.2318
  },
  "calibration": [
    { "bucket": "55-59", "predictions": 42, "actual_win_pct": 54.8 },
    { "bucket": "60-64", "predictions": 38, "actual_win_pct": 63.2 }
  ],
  "recent_form": [
    { "date": "2026-04-05", "games": 15, "correct": 10, "accuracy_pct": 66.7 }
  ],
  "benchmarks": {
    "coin_flip": 50.0,
    "vegas_consensus": 56.0
  }
}

Public health check (no API key required). Returns liveness, model freshness, last grading time, and last data sync timestamps. Designed for status pages and customer monitoring.

Status Values

Example Response

{
  "status": "ok",
  "timestamp": "2026-05-08T11:00:27-05:00",
  "api_version": "1.0",
  "model": {
    "version": "2.0",
    "trained_at": "2026-05-03 05:00:35",
    "training_games": 10616,
    "test_accuracy": 57,
    "test_brier": 0.2408
  },
  "checks": {
    "model_weights": { "status": "ok", "age_days": 5.25 },
    "grading":       { "status": "ok", "age_hours": 11.3 },
    "predictions":   { "status": "ok", "age_hours": 2.9 },
    "team_elo":      { "status": "ok", "age_hours": 0 },
    "database":      { "status": "ok" }
  },
  "response_time_ms": 5.1
}