How it works.
Everything you need to understand sports betting edge — from the math behind the model to how to read a pick card.
1. What is Expected Value (EV) betting?
In any bet, the market sets a price — an implied probability that something will happen. If you believe the true probability is higher than what the market implies, you have positive expected value (+EV). Bet enough +EV situations and you profit long term, regardless of short term variance.
For example: if a book offers a team at -110 (implied probability 52.4%) but your model calculates the team's true win probability at 58%, you have a +5.6% edge. At those prices, you're getting paid more than the risk warrants.
EV betting is about finding these discrepancies systematically and acting on them at scale. It's not gambling on hunches — it's probability arbitrage backed by data.
2. How EdgeIQ finds edge
Our XGBoost model is trained on 14,000+ historical games across MLB, NFL, and NBA — from 2019 to present. For each game, we engineer features across five signal layers specific to each sport:
- Market pricing: vig-removed Pinnacle lines as the true market implied probability
- Team efficiency: sport-specific metrics — pitcher xERA/xwOBA for MLB, EPA per play for NFL, net rating for NBA
- Matchup context: home/away splits, divisional history, rest days, recent form
- Venue and environment: park factors, weather (temp, wind) for outdoor games, dome and turf flags
- Situational signals: umpire run tendencies for MLB, referee factors, public vs. sharp line movement
The model is validated out of fold (OOF) to prevent data leakage — the same rigor used in professional quant finance. We never test a pick on data the model was trained on.
When the model probability significantly exceeds the market implied probability, we publish the pick. When there's no edge, we publish nothing. The model runs daily and picks are live by 10 AM ET on game days.
See the full model report with feature importance and calibration data →
3. Reading a pick card
Every pick card shows the same information:
4. Kelly criterion and bet sizing
The Kelly criterion is a mathematical formula for optimal bet sizing. It tells you what fraction of your bankroll to bet based on your edge and the odds:
EdgeIQ displays the full Kelly fraction for each pick but recommends using fractional Kelly (25 to 50% of full Kelly) to manage variance. Even with a real edge, bad streaks happen — Kelly assumes unlimited capital, but you don't have it.
A practical starting point: 1 to 2% of bankroll per pick on High Confidence, 0.5 to 1% on Medium Confidence. The app's bet tracker lets you set a default stake percentage and track how your bankroll evolves over time.
5. Closing Line Value (CLV)
Closing Line Value is the single best metric for evaluating a betting system long term. It measures whether you consistently beat the final closing line — the price available just before game time.
The closing line is set by millions of dollars of sharp money. If you consistently get better prices than the closing line, you're finding edge that the sharpest bettors in the world couldn't eliminate before game time. This is the strongest evidence a model works.
EdgeIQ tracks CLV for every pick automatically. You can see average CLV on the track record page. A positive avg CLV over a large sample is the most reliable indicator of a sustainable edge — more reliable than raw win rate in the short term.
6. Responsible use
Sports betting carries real financial risk. EdgeIQ provides analytical tools and picks backed by math, but no model predicts the future with certainty. You will lose bets.
A positive expected value edge means you profit over a large sample — not necessarily over 10 picks, 50 picks, or even a single season. Variance is real. Only bet money you can afford to lose. Never chase losses.
If you or someone you know has a gambling problem, resources are available. See our responsible gambling page →