Expected Value in NBA Prop Bets: When the Maths Says Bet — and When It Doesn’t

Expected Value Is the Single Number That Separates Strategy from Guesswork

Early in my betting career, I kept a spreadsheet of every prop bet I placed. I won 54% of them over the first three months and felt brilliant — until I ran the numbers and discovered I was losing money. My average winner returned less than my average loser cost, because I had been chasing short prices on heavy favourites without checking whether the odds offered genuine value. That was the week I learned about expected value, and it changed everything about how I approach the player prop market.

Expected value — EV — is the mathematical answer to a simple question: if I placed this exact bet 1,000 times, would I make money or lose money? A positive EV bet is one where the implied probability in the odds is lower than the true probability of the outcome. A negative EV bet is the opposite. The bookmaker builds margin into every price, which means that most bets are negative EV by default. The job of the prop bettor is to find the exceptions — the lines where your estimated probability exceeds the bookmaker’s implied probability by enough to overcome the vig.

Player props carry a margin of 5-8% on standard markets and 15-25% on same-game parlays. Those margins are the headwind you are betting into, and EV is the tool that tells you whether your analysis is strong enough to overcome it.

The EV Formula Applied to a Player Prop

The formula is not complicated. For a player prop over/under, EV equals the probability of winning multiplied by the net profit if you win, minus the probability of losing multiplied by the stake lost if you lose. In decimal odds, the calculation looks like this: take the decimal odds, subtract 1 to get the net profit per unit, multiply by your estimated probability of winning, and then subtract the probability of losing.

Suppose you are evaluating a rebounds over at decimal odds of 1.90, and your analysis puts the probability of the over hitting at 56%. The net profit per unit if you win is 0.90 (since 1.90 minus 1 equals 0.90). The EV per unit staked is (0.56 multiplied by 0.90) minus (0.44 multiplied by 1.00), which equals 0.504 minus 0.440, or +0.064. That is a positive EV of 6.4% — meaning for every pound you stake, your expected long-run profit is 6.4 pence.

Now compare that with a points over at 1.87 where your estimated probability is 53%. The EV is (0.53 multiplied by 0.87) minus (0.47 multiplied by 1.00), which equals 0.461 minus 0.470, or -0.009. That is negative EV of 0.9% — a losing proposition in the long run, even though you believe the over will hit more often than not. The margin ate your edge.

The backtested data from the 2025-26 season is instructive here. High-variance categories like blocks (69.9% win rate on the under) and three-pointers (63.2%) generate more frequent positive EV opportunities than points (55.7%), because the bookmaker’s model is less accurate on those categories and the gap between your probability estimate and the bookmaker’s is wider. The margin structure across different prop markets determines how large your edge needs to be before the bet is mathematically justified.

Estimating True Probability Without a Simulation Model

The EV formula is only as good as the probability estimate you feed into it, and this is where most punters get stuck. Building a full simulation model that runs 10,000 iterations per prop is beyond the reach of most casual bettors. But you do not need a simulation model to estimate probability with enough accuracy to find positive EV bets.

The simplest approach is historical hit rate analysis. Take the player’s game log for the last 30-40 games and count how many times the prop would have hit at the current line. If a player’s rebounds over 7.5 hit in 24 of his last 35 games, the raw hit rate is 68.6%. That is your starting estimate. Now adjust for context: is tonight’s matchup better or worse than the average of those 35 games? Is the player healthy? Is the pace projection above or below normal? Each adjustment nudges the probability up or down by a few percentage points.

The danger of this method is overfitting. A 35-game sample is small enough that random variance can distort the hit rate. If 5 of those 35 games were against the league’s worst rebounding defence, and tonight’s opponent is the league’s best, the raw rate overstates the probability. The correction is to weight recent games more heavily than distant ones, and to exclude games where the context was dramatically different from tonight.

A more robust approach is to work backwards from the bookmaker’s price. If the over is priced at 1.85, the bookmaker’s implied probability — stripping out the vig — is approximately 51-53%. Your job is to determine whether your own estimate is meaningfully higher. If your analysis puts the probability at 58%, the gap is 5-7 percentage points, and the bet is clearly positive EV. If your estimate is 54%, the gap is only 1-3 percentage points, and the bet is marginal — probably not worth the variance.

How Much EV Is Enough to Justify a Bet

Not every positive EV opportunity deserves a stake. The size of the edge matters, and so does the variance of the outcome. A bet with +2% EV and high variance — like a three-point prop — will need hundreds of repetitions before the edge materialises. A bet with +2% EV and low variance — like a points prop on a consistent scorer — will converge to profitability faster.

My personal threshold is +3% EV for standard prop bets and +5% EV for higher-variance categories. Below those thresholds, the expected profit is too thin to compensate for the emotional and financial volatility of short-term results. Above them, the mathematical case for the bet is strong enough to withstand a bad week or a losing streak without questioning the underlying process.

The margin on NBA player props — 5-8% on standard markets — sets a natural floor for how much edge you need. If the bookmaker is charging 6% vig, you need an estimated probability that exceeds the implied probability by at least 3-4 percentage points to generate meaningful positive EV after the margin. That sounds like a narrow window, and it is — which is why most prop bets are not worth placing and the discipline to pass is as important as the ability to find edge.