Testing a coin

You will test a coin both ways: by simulation and by exact counting.

The idea. Under the null idea, the coin is fair. The p-value is the chance that a fair coin gives at least as many heads as the real one.

Your tasks.

1. Finish heads_in(n). Toss a fair coin n times with random.random() < 0.5 and return the number of heads.
2. Finish p_value_sim(observed, n, sets). Return the share of sets simulated sets with at least observed heads.
3. Finish p_value_exact(observed, n). Return $\dfrac{1}{2^n}$ times the sum of comb(n, h) for h from observed to n.