Understanding the Math Behind Prop Bets
Why prop bets matter for the modern bettor
Look: a prop bet isn’t just a side‑show; it’s the arithmetic engine that can turbo‑charge a bankroll. You see the line, you place the wager, you hope the odds line up. The flaw? Most punters skim the surface and miss the calculus that separates a lucky guess from a repeatable profit.
The raw numbers behind a “player points” proposition
Here is the deal: a player’s scoring distribution is a binomial process—each shot either lands or misses, each possession is a trial. If a guard attempts 30 shots at a 45% success rate, the expected points sit at 13.5. Slip in free‑throw attempts, three‑point attempts, and you have a Poisson‑ish tail that dictates variance.
Probability, not superstition
By the way, the odds posted by bookmakers are simply the inverse of the implied probability plus the vigorish. A -110 line translates to a 52.4% implied win chance, not 55% as the sportsbook pretends. Subtract the vig, and you get the raw probability you need to compare against your model.
Expected value (EV) is your compass
EV = (Probability of win × Payout) – (Probability of loss × Stake). If the true probability of a player hitting over 25 points is 58%, and the line pays +120, the EV is positive: (0.58 × 1.20) – (0.42 × 1) = 0.696 – 0.42 = +0.276 per unit. That’s green on paper.
Bankroll management meets Kelly
And here is why you can’t ignore variance. The Kelly criterion tells you how much of your bankroll to risk: f* = (bp – q) / b, where b is the net odds, p is your win probability, q = 1 – p. Plug the numbers above: b = 1.20, p = 0.58, q = 0.42 → f* ≈ 0.133. Roughly 13% of your bankroll on that single bet. Too aggressive? Half‑Kelly to smooth the swings.
A real‑world example
Suppose the Knicks’ rookie is projected for 19.5 points, but your regression model says 21.2. The sportsbook line is at -115 for the over. Convert: implied probability 53.5%, your model 58%. EV: (0.58 × 1.115) – (0.42 × 1) = 0.647 – 0.42 = +0.227. Kelly suggests a 10% stake on a $1,000 bankroll—$100.
Odds conversion tricks you can use today
Read the line, flip it. American odds → decimal = (odds/100) + 1 for positives, = 100/|odds| + 1 for negatives. Then strip the margin: true odds = decimal / (decimal – 1). That’s the number you compare to your calculated probability. Any gap, and you have an edge.
Spotting the edge on nbapropbets.com
Most sites inflate the line to protect themselves. Your job: pull the raw data, run a Monte Carlo simulation, let the distribution settle. When the simulated win chance exceeds the implied probability by even a single percent, you’ve unearthed a bet worth taking.
Actionable move right now
Grab the next “total rebounds” market, plug the player’s per‑game rebound rate into a negative‑binomial model, compare the resulting win chance to the posted odds, and stake the Kelly‑fraction of your bankroll. No fluff, just math‑driven profit.
