A full breakdown of the model, from raw sportsbook odds to a +EV pick in your Discord.
Our Python worker polls The Odds API v4 for every active market across MLB, NHL, NBA, EPL, MMA, and Boxing. All major US sportsbooks are included: FanDuel, DraftKings, BetMGM, Caesars, and more.
We separate books into "sharp" (efficient market makers: Pinnacle, Bookmaker) and "soft" (retail books that lag). Sharp odds are aggregated using a weighted average to estimate the true probability of each outcome.
Every set of odds has a bookmaker's juice baked in. We use multiplicative devigging to remove that margin and convert sharp odds into true implied probabilities, the most accurate estimate of the real-world likelihood.
For each outcome, we compare the true probability against the price offered at every soft book. When a soft book's odds imply a lower probability than our model's estimate, there's positive expected value. EV = (true_prob × (odds − 1)) − (1 − true_prob).
Not every positive-EV line is worth playing. We apply Kelly criterion (minimum 0.25%), EV floor, and CLV probability filters to surface only high-quality opportunities. Lines with EV > 30% are excluded as likely data errors.
Each pick gets a 1–5 star rating based on CLV probability: the likelihood the pick beats the closing line. This is derived from a bagged logistic regression model trained on historical closing lines. 5-star picks are the sharpest.
New picks fire to your Discord tier channel the moment they're found. Basic members see 1★ picks. Premium sees 1–3★. VIP sees all picks, including every 5★ play. The dashboard shows your tier's picks in real time.
The math works. Every pick tracked and published since launch.
See the full record →EV (Expected Value)
The average profit per unit bet over many wagers. +5% EV means for every $100 bet, you expect +$5 back on average.
Devig
Removing the sportsbook's margin (juice/vig) from their odds to reveal the true implied probability.
CLV (Closing Line Value)
How much your odds beat the final pre-game price. Consistently beating the close is the strongest indicator of long-term edge.
Kelly Criterion
A formula for optimal bet sizing based on edge and odds. We use half-Kelly to reduce variance, capped at 3% of bankroll.
Sharp books
Sportsbooks that accept sharp action and have efficient, well-calibrated lines. Used as our benchmark for true probability.
Soft books
Retail sportsbooks that move lines more slowly and tolerate recreational bettors, where our edge is exploited.
Ready to get started?
Join Discord free or subscribe for full dashboard access.