Potentially invalid approximation in power spectrum normalization

[isands17]

Describe the bug

I wanted to ask about an issue mentioned in footnote 14 (page 12) of https://arxiv.org/abs/2112.04515: there is an error in the public BOSS power spectrum normalization due to replacing the geometric factor A with a sum over random particles, which is apparently the same normalization used in the Nbodykit code. I have noticed some deviation in amplitude between estimated multipoles and our modeled multipoles and was wondering if this bug could be related, although I am using masks on lognormal catalogs to approximate surveys other than BOSS. Is there a fix I need to do here, either on the estimator side or the modeling?

Thank you!

Version
3.15

Edit: fixed link to paper and added page number