Modified Even Coins

2025-07-26 [Probability, Conditional Probability, Law of Total Probability]

Problem:

$n$ coins are laid out in front of you.

One of the coins is fair, while the other $n-1$ have probability $0 < \lambda < 1$ of showing heads.

If all $n$ coins are flipped, find the probability of an even amount of heads.


Let $\mathcal{F}$ represent the outcome of the fair coin and $\mathcal{H}$ the total heads from the $n-1$ unfair coins.

This means the total number of heads is given by $\mathcal{F} + \mathcal{H}$ and we are specifically interested in the probability that $\mathcal{F} + \mathcal{H}$ is even.

Now if we condition on the fair coin (assuming that $\mathcal{F} + \mathcal{H}$ is even):

$$P(\mathcal{F} + \mathcal{H}) = P(\mathcal{F} + \mathcal{H} \mid \mathcal{F} = 0)P(\mathcal{F} = 0) + P(\mathcal{F} + \mathcal{H} \mid \mathcal{F} = 1)P(\mathcal{F} = 1)$$

We know that the probability of the outcomes for the fair coin are:

$$P(\mathcal{F} = 0) = P(\mathcal{F} = 1) = 0.5$$

and we also know the outcomes of $\mathcal{F}$ is either 0 or 1, so we can simplify the above to:

$$P(\mathcal{F} + \mathcal{H}) = P(0 + \mathcal{H}) \cdot 0.5 + P(1 + \mathcal{H}) \cdot 0.5$$

In this case where $\mathcal{F} = 0$ we need $\mathcal{H}$ to be even and for $\mathcal{F} = 1$ we need $\mathcal{H}$ to be odd.

This then leaves us with:

$$P(\mathcal{F} + \mathcal{H}) = 0.5 \cdot P(\mathcal{H} \text{ is even}) + 0.5 \cdot P(\mathcal{H} \text{ is odd})$$

Because of the law of total probability both cases of $P(\mathcal{H} \text{ is even})$ and $P(\mathcal{H} \text{ is odd})$ must sum to 1. This means $P(\mathcal{H} \text{ is odd})$ is equivalent to $1 - P(\mathcal{H} \text{ is even})$.

If we put this back into the equation we have:

$$\begin{align*} P(\mathcal{F} + \mathcal{H} \text{ is even}) &= 0.5 \cdot P(\mathcal{H} \text{ is even}) + 0.5 \cdot (1 - P(\mathcal{H} \text{ is even})) \\ &= 0.5 \cdot P(\mathcal{H} \text{ is even}) + 0.5 \cdot 1 - 0.5 \cdot P(\mathcal{H} \text{ is even}) \\ &= 0.5 \cdot P(\mathcal{H} \text{ is even}) + 0.5 - 0.5 \cdot P(\mathcal{H} \text{ is even}) \\ &= [0.5 \cdot P(\mathcal{H} \text{ is even}) - 0.5 \cdot P(\mathcal{H} \text{ is even})] + 0.5 \\ &= 0 + 0.5 \\ &= 0.5 \end{align*}$$

Which means the probability of an even number of heads is 0.5.