Betting analytics glossary

The simplest market: pick the team or player you think wins outright — margin does not matter. Odds are skewed by how likely each side is, so a heavy favorite pays little and an underdog pays more.

The book sets a margin (e.g. −6.5) the favorite must win by, or that the underdog can lose by and still cover. Betting the spread is a bet on the margin of victory, not just the winner.

The book posts a number for the combined final score (or a single stat). You wager whether the actual result lands over or under that line — independent of which side wins.

A market on a single player’s output — passing yards, points, strikeouts, shots — with an over/under line. Props are isolated from the game result, which is why historical hit rate against the exact line matters so much.

Convert any odds to the percentage the book is effectively pricing in. American −150 implies about 60%; +150 implies about 40%. Comparing implied probability to your own estimate is the foundation of finding value.

The margin sportsbooks add to both sides of a market so the implied probabilities total more than 100%. On a standard −110/−110 line the vig is roughly 4.5% — the cost of placing the bet, win or lose.

What a wager is worth on average if you could place it repeatedly. Positive EV means the price beats the true probability and the bet profits over time; negative EV means it loses over time — regardless of any single result.

The share of past games in which a result happened — for a prop, how often a player cleared this exact line over a recent window (e.g. 7 of the last 10). A descriptive record of the past, not a prediction of the next one.

A composite score built from weighted, visible factors — recent hit rate, line-vs-average gap, matchup, usage, and injury context — that summarizes how strongly the data favors one side of a line. Higher means a stronger statistical case, never a guarantee.

Combines multiple selections so the payout multiplies — but every leg must win or the whole ticket loses. The longer the parlay, the higher the potential payout and the lower the true probability of cashing.

A parlay whose legs all come from the same event. Because the legs are often correlated (a QB’s yards and his receiver’s yards move together), books price SGPs with extra margin and the legs are not independent.

Placing a wager on the opposite outcome of an open bet to guarantee a profit or cap a loss — common on the final leg of a live parlay. A hedge calculator finds the exact stake to balance both sides.

Betting every outcome of a market at different sportsbooks whose combined implied probability is under 100%, locking in a small guaranteed return regardless of result. Rare, short-lived, and limited by stakes and account restrictions.

A staking method that sizes each bet by your edge and the odds, maximizing long-term bankroll growth. Full Kelly is volatile, so many bettors use a half-Kelly fraction to cut variance.

The difference between the odds you bet and the line right before the event starts. Consistently getting a better number than the close is one of the most reliable signals that you are betting with an edge.

A standardized stake — usually 1–2% of your bankroll — used to track results independent of dollar amounts. Measuring profit in units rather than dollars keeps records comparable as a bankroll grows or shrinks.

Recalculating a market’s probabilities so they total exactly 100%, stripping out the vig to reveal the “fair” line. The no-vig number is the cleanest baseline for comparing books and judging whether a price offers value.

When the result lands exactly on the line — a 3-point favorite winning by 3, or a player finishing on the prop number. The wager is graded a tie and the stake is returned, with no win or loss.

An alternative to the main number — a higher or lower spread, total, or prop line — priced with different odds. Buying a safer line costs payout; taking a tougher line pays more.

How many games in a row a player has gone over (or under) a market. A streak is context, not destiny — it can reflect a real role change or just variance, which is why it is one input among many, never the whole case.