So, in shopping for a beverage / wine fridge to be used as a cellar, I've been faced with capacity limitations to contend with (who doesn't contend with those, though...). This made me realize the important formula for what my rate of consumption from the cooler would have to be: Rate of Consumption = Quantity of Beers / Average Age of Beer Some beers I'm looking to cellar for 5 years, others only 6 months. But if I average them all out, I'd guess 18 months would be the average. The cellar I'm looking at holds about 40 beers (mostly bombers). Therefore RoC = 40 / 18mo = 2.2 beers per month, or about 1 every 2 weeks. Who else is nerdy enough to calculate such things? And what do your numbers look like?

Not me. From experience I would say whatever size you get you will need something larger in about 6 months to a year.

The ideal scenario would be to figure out how often you want to drink beers from the cellar, and what you want the average age to be, and you can see how many beers you should have on hand - and then try to stick to that (and subsequently fail, undoubtedly)

I've been thinking about the same thing, but only looking at one type of beer and thinking about how much storage I will need (since I'm not space-constrained). So if I want to build this vertical of gueuze with this many bottles per year and I'm drinking this many per year, how much should I plan on having? Really it's a differential equation: dN/dt = R - cN, where N is the number of bottles and R is the rate in and c is the percentage of bottles you're drinking (I think that it makes sense to think of it this way, since you might drink 6 bottles of something in a year when you have 12 of them, but you probably won't drink 6 when you only have 6). This formulation is nice and general, set dN/dt to 0 and you end up with basically what you had, that the rate in (which is the same as the rate out) is equal to the number times your percentage constant, which is the inverse of your average time. But if you set R to zero (the beer is no longer released, etc) then you get the classic decay equation, N(t) = N_0*e^(-c*t), where your supply is an exponential decay based on initial supply and elapsed time. I don't really know how useful any of this is, but you can make guestimates of the variables and plan accordingly.