tompoland Retention would normally mean stripping the grinder and weighing the grinds captured in the burrs, chute, auger etc. What you are doing here is really assessing the ‘dose consistency’. I think this is more practical aspect to measure, as I don’t think people strip & clean their grinders every 100 grinds 😃
I would always use 0.01g resolution scales. When I first got my Niche Zero I checked the dose consistency for 10 grinds with 0.1g scales, I always got the same answer - 0.0g difference. There were visibly grinds that exited the grinder, when I then used 0.01g scales I found the standard deviation over 10 grinds was 0.19g, if I add this to the average difference in beans in, to grinds out, I get 0.23g. Calculating to 95% confidence level 0.44g.
The 0.1g scales do not ‘see’ this difference.
I measured the dose of whole bean, then the difference between that and the grinds produced, but I wouldn’t say that was any more ‘correct’ than what you’re doing,
Grinder dose consistency