Real World Maths Busters

The Maths Behind Everyday Life

The story, the formula and a worked example — see exactly where this maths shows up in the world around you.

Back to all stories
Games & Chance 4 min read

Detecting a Biased Coin

How many flips to prove a coin isn’t fair?

Using a binomial test, you need enough flips so that the expected number of heads with p=0.6 is statistically distinguishable from p=0.5.

Sample size estimation

A rule of thumb uses the formula for detecting a difference Δp from 0.5 with 95% confidence.

n≈z2⋅p(1−p)Δp2n \approx \frac{z^2 \cdot p(1-p)}{\Delta p^2}

Detect p=0.6 vs p=0.5

  1. 1

    Step

    Δp = 0.1, z = 1.96 for 95% confidence

  2. 2

    Step

    n ≈ (1.96² × 0.5 × 0.5) / 0.1²

  3. 3

    Step

    = 3.8416 × 0.25 / 0.01 = 96.04

Answer

About 96 flips needed

Why it takes so many

A 60% coin looks very similar to a fair coin over small samples. With 10 flips, getting 6 heads happens 20.5% of the time even with a fair coin.

P(6/10∣p=0.5)=(106)⋅0.510=0.2051P(6/10 \mid p=0.5) = \binom{10}{6} \cdot 0.5^{10} = 0.2051

The formula behind it

Binomial Distribution

P(k)=(nk)pk(1−p)n−kP(k) = \binom{n}{k} p^k (1-p)^{n-k}
See the full formula, variables and more examples

The takeaway

Detecting a 60-40 coin takes about 100 flips — small biases hide remarkably well.

Back to all stories