You're merely mentioning two variables with a obvious correlation. You can simply eliminate this restriction by setting review at 9999, so why is this even mentioned in your advice?
First we don't know if 35-40 is a good number, it's an arbitrary number op picked like everyone else.
Lets just assume there are only two variables, the amount of time OP wants to spend on Anki, and number of cards reviewed per minute. Assuming all cards are already made and there are no other factors number of mature and learning cards already in the collection.
OP needs to find his delta, change in review number per day over change in number of new cards added per day. Assuming the correlation is linear, using simple linear regression, may we can find the optimal number of new cards per day.
I think this would be the basic anki modelling to optimisation, hopefully some genius can come up with a proper econometric model with hard empirical evidence.
All other "advice" with number to back it up is just waste of time.