In answer 3 for solution 8, the function estimate_asymptotic_variance uses a biased variance estimator (1/C instead of 1/(C−1)):
It uses:
varE = 1/C * sum((Es - mean(Es))^2)
It's the biased population variance rather than the unbiased sample variance. The correct estimator should likely divide by C−1.
Hopefully this counts for the challenge question mark!
In answer 3 for solution 8, the function
estimate_asymptotic_varianceuses a biased variance estimator (1/C instead of 1/(C−1)):It uses:
varE = 1/C * sum((Es - mean(Es))^2)It's the biased population variance rather than the unbiased sample variance. The correct estimator should likely divide by C−1.
Hopefully this counts for the challenge question mark!