productivity modules for byproducts in #-# range seem to undestimate the output

[Quezler]

Problem Description

(read the reproduction steps first, this textarea is a picture dump)

module-less setup:

with modules:

the belt test:

Reproduction

according to the laws of aviation, there is no way less than 15 bees would not fit on a belt per second, the bees of course do not care.

i have a hunch the recipes with a 0-x aren't accurate sometimes, as you can see on the pictures of the belt test it outputs just under 15/s, so the belts can handle the 2 byproducts just fine.

however when i add some productivity modules into the mix the byproducts drop to 1.5 items per second, which will surely fit on a yellow belt:

except, this is far more than 1/15 of the belt's capacity:

sorry for the messy issue report, i typed it up as i rubberducked my setup to prove it happened.

one final pick with the machine numbers rounded, yet its still nowhere near 10/s:

(re)discovered it when the byproduct line for this factory couldn't be yellow or green: (i recall running into this ones before on a run long ago but couldn't remember which recipe it was back then)