Msprime records $\sum_i (l_i - r_i)$ will _not_ sum to `n-1` for `n` samples (unless there is no recombination)

msprime_covering.py

import msprime
pop = __import__('random').randint(2, 20000)
sum = 0
with open('example.tsv', 'w') as f:
    msprime.simulate(pop, recombination_rate = 10).write_records(f)
with open('example.tsv', 'r') as f:
    f.readline()
    l = f.readlines()
    for line in l:
        vec = line.split('\t')
        sum += float(vec[1]) - float(vec[0])
assert int(sum) + 1 == pop, "population size: {}, 'coverage': {}".format(pop, int(sum))