Poisson Distribution
A probability distribution that models the number of times a discrete event occurs in a fixed interval. It's ideal for rare, independent events like three-pointers made or turnovers per game.
Like counting how many buses arrive at a stop per hour — it's always a whole number, usually small, and somewhat random.
Why it matters
Props on low-count stats (steals, blocks, threes) don't follow a bell curve — Poisson gives more accurate over/under probabilities for these lumpy outcomes.
How DMP uses this
DMP uses Poisson P(over) specifically for fg3m (three-pointers made) props, producing more accurate probabilities than a normal approximation for these discrete, low-count events.
Common mistake
Using normal distribution math on stats that average 1-3 per game — the probabilities at integer boundaries can be significantly off.