2026-06-18 · 9f68c4a

How our predictions work

The data, models, and methods behind every spread, total, and bet on this site — from a one-page overview to the full handicapping reference series.

Start here

How Our Model Works

One-page tour of the full prediction pipeline — data, models, tuning, scoring.

Handicapping Series — Overview

Why this reference exists, and the five numbers to memorize.

Reference series

1 · Bookmaking Fundamentals

How books set & move lines, the vig, sharp vs. square money, Closing Line Value.

2 · Predicting Spreads

Power ratings, Elo, key numbers (3 & 7), spread↔probability, QB & home field.

3 · Predicting Totals

Pace, points per drive, weather (wind is king), game script.

4 · Variable Catalog

Every predictive input ranked, with regression-to-the-mean caveats.

5 · Modeling Methods

Ridge ratings, walk-forward validation, calibration, market-aware blending.

6 · Betting Strategy & Bankroll

Edge identification, de-vigging, fractional Kelly, CLV.

7 · Engine Roadmap

How the theory maps to our codebase, and what's next.

8 · CLV & Odds Data

Closing Line Value, where to get open/close odds, and the capture/backtest tools.

These pages render the Markdown under docs/handicapping/. Educational / modeling reference — not betting advice.