Methodology

A technical overview of the prediction pipeline, bet sizing, and verification protocol behind TiltDataLabs.

The Model

Each sport is modeled by an ensemble of five algorithms: Logistic Regression, Linear Discriminant Analysis, XGBoost, a multi-layer perceptron (MLP), and Random Forest. Predictions are aggregated across models — a pick is surfaced only when a consensus threshold of models agree on the direction.

Feature Engineering

  • Rolling team stats — win rate, points, offensive/defensive efficiency, pace, and rest days over configurable windows (last 5, 10, 20 games).
  • Pitcher / player stats — season and rolling game-log metrics for starting pitchers (MLB) and key players.
  • Opponent-adjusted metrics — strength of schedule corrections so a win streak against weak teams doesn't receive the same weight as one against contenders.
  • Betting-market features — historical ATS records, over/under trends, and closing-line movement.

All features are strictly backward-looking. Training uses chronological splits — the model never sees future data during training or validation. This leak-free discipline is the single most important design constraint in the pipeline.

Finding the Edge

For each game and market (moneyline, spread, over/under), the ensemble produces a model probability. This is compared against the implied probability derived from the best available decimal odds:

edge = model_prob − implied_prob

A positive edge means the model believes the true probability is higher than what the market is pricing. Only picks above a per-sport, per-market threshold are surfaced — typically 3-5% minimum edge depending on the sport.

Bet Sizing: Kelly Criterion

The Kelly criterion determines the fraction of bankroll to wager on each edge, maximizing the expected logarithmic growth of wealth:

f* = (p · b − q) / b

where p = model probability, q = 1 − p, and b = decimal odds − 1.

Safeguards

  • Quarter-Kelly — we stake 25% of the theoretical optimum to reduce variance and protect against model mis-calibration.
  • Min/max constraints — no bet exceeds a per-sport cap (e.g., 5% of bankroll) and no bet falls below a minimum threshold.
  • Edge threshold gating — picks below the minimum edge are dropped entirely, regardless of Kelly sizing.

Blockchain Verification

Every daily pick file is a canonical JSON document containing the game date, picks, odds, and model probabilities. The pipeline computes the SHA-256 hash of this file and submits it to OpenTimestamps calendar servers.

Protocol

  1. Pipeline generates wagers/YYYY-MM-DD.json with all selected picks.
  2. SHA-256 hash is computed and submitted to three independent calendar servers for redundancy.
  3. Calendar servers return a receipt; the .ots proof file is saved alongside the JSON.
  4. Within 1-24 hours, the hash is included in a Merkle tree that gets anchored in a Bitcoin block header.
  5. The .ots proof is upgraded with the Bitcoin block attestation — a permanent, immutable record.

Verification is trustless: anyone with the JSON file and the .ots proof can independently confirm the timestamp using the open-source ots-cli tool or our in-app verify button. No trust in TiltDataLabs is required.

Markets

SportMoneylineSpreadOver/UnderFirst 5
NBA✓✓✓—
Soccer✓✓✓—
MLB✓✓✓✓
NFL✓✓✓—

First 5 (F5) variants cover moneyline, spread, and over/under for the first five innings of MLB games.

See it in action

Check the live track record or sign up for a free trial.

View PerformanceStart Free Trial