Binomial distribution
The number of wins in n independent trials that each win with the same chance p. It tells you how likely every possible count of winners is.
01The formula
- the number of wins in n independent trials
- the number of trials (trades)
- a particular number of wins
- the chance each trial wins
- n choose k: the number of ways to place k wins among n trials
02See it move
03Where it is used
- Quant researcherChecks whether a run of winning trades is more than a fair coin would produce.
- Risk managerBacktests value at risk by asking whether the number of breach days is plausible for the stated breach rate.
- Market makerEstimates how many of n quotes will be hit when each is filled with chance p.
04Why traders use it
It turns a win rate into a range of outcomes. A strategy with a 55% edge still has losing weeks, and the binomial says exactly how often.
05Worked example
With 20 trades that each win 55% of the time, the chance of exactly 14 winners is 7.46%, and the chance of at least 14 is 13.0%. On average you get 11.0 winners, give or take 2.22.
06How it is built
- Any one sequence with wins and losses has probability , because the trials are independent.
- There are such sequences.
- Multiply: .
- X is a sum of Bernoulli trials, so its mean is and its variance .
07The common mistake
Using it when trials are not independent. Trades from the same signal on the same day often win or lose together, so real counts are more spread out than the binomial says.
08In Python and LaTeX
from scipy.stats import binom n, p, k = 20, 0.55, 14 print(round(binom.pmf(k, n, p), 4), round(binom.sf(k - 1, n, p), 4)) # P(X=k), P(X>=k)
P(X = k) = \binom{n}{k} p^k (1 - p)^{n - k}, \qquad E[X] = np,\quad \operatorname{Var}(X) = np(1-p)09Where it fits
Rather not code the maths yourself? Describe your trading idea and let AI build the bot.
Build a bot with AgenKitThe worked example and the model run the same code, so every number matches. Models simplify markets on purpose. This is education, not financial advice.