probs = np.array([5, 10, 9, 14, 8, 4, 6, 8, 15, 4, 4, 7 ])/100
in_next_two_years = probs+((1-probs)*probs)

res = []
for _ in range(300000):
    sample = (np.random.random(in_next_two_years.shape) < in_next_two_years).sum() >= 2
    res.append(sample)
np.mean(res)

@NoaNabeshima replace 2 with 4

import numpy as np

probs_for_one_year = np.array([5, 10, 9, 14, 8, 4, 6, 8, 15, 4, 4, 7 ])/100
in_next_two_years = probs_for_one_year+((1-probs_for_one_year)*probs_for_one_year)

res = []
for _ in range(300000):
    sample = (np.random.random(in_next_two_years.shape) < in_next_two_years).sum() >= 4
    res.append(sample)
np.mean(res)

@NoaNabeshima returns 0.09

After some thought updated probs to
[4, 15, 8, 20, 12, 6, 20, 15, 15, 6, 6, 12]

@NoaNabeshima what is the original probs array storing?

@firstuserhere My eyeballed probabilities for the next year. Then I assume that the year after has the same probability

@firstuserhere https://docs.google.com/document/d/1vzgzbLqGEuiA5EougVrxB2AJe4RvIFXwdFluiik8nVs/edit