Science on grit size in a 12# barrel

I've done some science, testing 60/90sc vs 120/220sc in a 12# barrel, like what adam5 did in 3# barrels based on that grit article. The premise is that we typically measure out grit by volume, but different grit sizes have different density. So what if you measure out the grit by weight instead? You buy grit by weight, and at least where I buy grit, 60/90 and 120/220 cost the same amount.

TLDR: The two grit sizes performed almost identically. I will be transitioning over to 120/220 just so I can skip a stage.

Results:
Each run got 1100mL of water and 275g (about 12-13 Tb) grit in a QT12. I used a mix of rocks similar to my normal loads, most of which have already been through stage 1 and need more work.

60/90
4674g initial weight of rock
4251g final weight
9.05% reduction

120/220
4674g initial weight of rock
4255g final weight
8.96% reduction

I was happy to see that the 60/90 had about 9% loss, close to what adam5 saw for 80 grit. I was honestly hoping the 120/220 would come in around 11% like adam5, but there you go, it's also around 9%. Some parts of my method were different from adam5, so that could be a factor, and of course the larger barrel. I don't know the real reason I got different results from adam5, but it doesn't really matter: 120/220 is still so close that I think it's worth switching over. Of course I still have some 60/90 and I'll use it for the first 1 or 2 runs when I have a lot of fresh rocks, then switch to 120/220 so that when rocks get smooth enough, they're ready to go on to 500.

Further research:

• We only have a couple data points on this, I encourage others to try running a test if you feel like it. I think it would be interesting just to get some reports on weight loss for different grits and barrels, you don't even have to do a side-by-side. It seems like there's new ground to be tested too, like amount of water, or length of tumble.
• When I laid out rocks for a pic, it really sank in how much wear happens in the first run. It was hard to tell the difference between 1 and 2 runs, and even some rocks that have been through many runs. I'm not sure of the value of carefully staging so many rocks. It might be fine to run tests with smoothish rocks plus an appropriate number of raw rocks. (Assuming that's your normal situation that you're trying to reproduce.)
• I might check the loss when going from 120 to 120, as mentioned below under Possible issues.
• I'm a bit curious to try a run with raw crushed rocks and no grit. Would the rough jagged bits grind away at each other and give you a decent first run, and you could save a little bit of grit? I know some grinding would happen, but how much?
• There will be no more lists because that was a pain.

Detailed method and some reasoning behind it:
I could have set up a more stringent test, but really, I just wanted to know if the two grits are even reasonably close. Even if 120/220 were a little less effective, I might switch over just to avoid the extra stage. The test was based on my normal process, because my main interest is whether this works for *me*. I hope it's useful to other people, but YMMV.

-Two runs back to back in a fairly new (< 6 mos) QT12, first 60/90 and second 120/220
-Clean barrel normally. The barrel always rinses out easily and only needs a light scrub on the rim and edge of lid with a toothbrush.
-Rock Shed grit, purchased within the last 6 mos.
-Tumbler between 2/3 and 3/4 full. I always measure with a caliper and aim for the middle of this range but of course there's some +/-. Each run was at a very similar level. Exact weight listed above.
-1100ml of water in each run. The normal amount I use, a little below the top of the rocks, but this time I measured it by volume and used the exact same amount in each run.
-I normally use a more-or-less level 3/4 cup (12 Tb) 60/90 grit in the QT12. I measured a proper level 3/4 c of 60/90, scraping a flat edge across the top. It weighed in at 275g, and I used the same weight of 120/220. (FWIW this amounts to roughly a heaping 3/4 cup of 120/220, which is how I will do it in the future.) Grit was weighed by itself, without the ~4600g of rock plus water plus barrel.
-Tumble for 7 days.
-Rinse well, but no burnish.
-Air-dry rock for 24 hours indoors before weighing, then weigh every 24 hours until weight remains unchanged (48 hours), and use final weight.
-All rock is hard: various jasper, quartz, agate, pet wood. All bought online, mostly crushed.
-Variety of sizes, from 3/4 to 2 or 3 inches, which is what I normally try to include. (I specifically try to buy a variety of sizes.)
-For me, most of the time in stage 1 is spent with rocks that are going back in after one or more runs and are somewhat smooth. So for each run I used a handful of raw crushed or cut rock, a handful that have been through stage 1 once, a handful that have been through stage 1 twice, and the rest are pretty smooth from multiple runs. Also a couple rocks that I have split with a chisel, so they're mostly smooth with a rough spot. This is a pretty normal mix for me in stage 1.
-I hand-picked the rocks to be as similar as possible between the 2 test runs, in type, size, and shape. When it came to the "pretty smooth" rock I hand-picked as many matching rocks as I could, but eventually just grabbed hand-fulls by size, so each test batch gets 1 handful of small, 2 handfuls of medium, and repeat. With such a large barrel and so much randomly-selected fill, I hope it averages out. I sorted out the rocks for both runs at the same time, and the final weights were initially within 27g. I added 1 rock to the lighter batch to make them equal. So I feel I did a pretty good job of balancing the 2 batches.

Here are the raw, 1, and 2 run rocks, large rocks, and a pile of smoother rocks for each run. This is a little over half the rocks put in each run, remaining rocks were similar to the piles of smoother rock.


Possible issues:
-I found a small 5g rock near where I dried the rocks after tumbling. It may have been from one of the test runs, it may have been from a different run. Adding 5g to either test run doesn't significantly change the results. (about 0.1%)
-I lube the tumbler's bushings once a month, and did the test runs at the end of the lube cycle. I figure if anything this puts 120 at a slight disadvantage.
-I used a kitchen scale. I was concerned about precision, especially with the weight being near its max capacity. But I weighed the 4674g of rocks all together, and then again in smaller batches and added up the smaller weights. The results were within 1g.
-I ran many rocks for multiple runs in 60/90 and set them aside partially finished in order to set up for this test. I don't know if it affects the results that some of these rocks are going back into 60/90 and others are going into 120/220. I might do one more test where rocks go from 120 to 120 and just confirm that the loss is still around 9%.