Talk:Composter

Could we also rank them by % chance, not just Alphabetically?
I.e., like the Snapshot description Page:

10% chance: all saplings, all leaves, all types of seeds (except nether wart), (sea) grass, sweet berries, and (dried) kelp. 20% chance: dried kelp blocks, tall grass, sugar cane, cactus, vines, and melon slices. 50% chance: wheat, lily pads, melons, sea pickles, pumpkins, apples, potatoes, beetroots, carrots, cocoa beans, all mushrooms, mushroom stem blocks (not mushroom blocks), all ferns, and all (tall) flowers. 80% chance: hay blocks, bread, cookies, and baked potatoes. 100% chance: cake and pumpkin pie.

It also helps show those 5 only seemingly specific % amounts. Easier to keep track of same-#'s. Yilante 1 /16 /19  5:02 p.m. 76.209.248.192 01:02, 17 January 2019 (UTC)


 * Th table can be sorted by the chance to compost, by clicking the column header. Having a compact per-chance list in addition to the table might be a good idea, though. 「 ディノ 奴 千？！ 」☎ Dinoguy1000 01:32, 17 January 2019 (UTC)


 * I've added so hopefully it might sort correctly now. – Nixinova Nixinova sig image 1.png Nixinova sig image 2.png 01:35, 17 January 2019 (UTC)


 * And it looks very niice. But as pointed-out in the MC Forum, "their % 's are weird."


 * So we should probably also analyze the best-ranges of given plants /crops /Crafted-Food, yields. So for example, do each by basically-efficiency of input-type to amount of output for it.


 * To specifically quote, "As things stand you're always better off composting with raw materials." ("Shadow" icon).  Yilante 1 /17 /19  2:33 a.m. 76.209.248.192 10:33, 17 January 2019 (UTC)

Percentages
I really don't think the percentage for wheat seeds is 10%. After extensive testing, I'm convinced it's 12.5%, or one chance in 8 rather than 10. I can do a full-blown chi-square test if anyone wants, but you can see for yourself - run a full chest of wheat seeds through a composter, and you'll get ~31 bone meal.

I haven't yet tested the other items in the "10%", but it seems likely that they also have the higher percentage. TheShadowLC (talk) 13:41, 18 January 2019 (UTC)

I just realized that I made an error in my calculations. A hypothesis test for the true percentage being greater than 10% on my current dataset gives a p-value of 0.137 - a tad high, but nowhere near high enough to reject the null hypothesis. Solution: Collect more data! TheShadowLC (talk) 14:30, 18 January 2019 (UTC)

Sigh, a tad *low*, I meant to type. TheShadowLC (talk) 14:40, 18 January 2019 (UTC)

Okay. Having run 3 double chests of wheat seeds (10368 total) through a composter, I got 196 bone meal, corresponding to 1372 composter levels. I calculate the following:

A one-sided hypothesis of whether the true mean proportion is greater than 0.1 has a p-value of 1.3e-22. I therefore reject the null hypothesis that the true mean proportion is 0.1.

I calculate a 99% confidence interval of [0.12375, 0.14090] for the true mean proportion. That's a range of 1 in 7.0971 to 1 in 8.0803.

The percentage could conceivably be higher than 12.5%. Doing a hypothesis test on that gives a p-value of 0.0138 - much higher than before, but still rather low. TheShadowLC (talk) 18:25, 18 January 2019 (UTC)