Digiconomist kWh data accuracy about ethereum-nft-activity HOT 6 CLOSED

hi sterling :) energy use and emissions for bitcoin is certainly better studied than for ethereum.

when i reviewed this topic, i found one recurring sentiment: it is very important to tie crypto energy use to a date. koomey stresses this in his work. because cryptoart is a very recent phenomena, estimates from 2017-2019 cannot be used.

as far as i've been able to find so far, digiconomist is the only source that attempts to provide a continuous estimate across multiple dates. so it's the only source that meets a basic requirement that is agreed upon by all researchers.

i would absolutely love to see a bottom up estimate for ethereum emissions. i've seen some for bitcoin, but for ethereum it would be a bit more intensive since we are talking about GPU releases and not infrequent ASIC releases.

i would also add that regarding point 1, the mining revenue, this is easy to calculate based on the price of ethereum. the other points are indeed back-of-the-envelope assumptions.

personally, what i would like even more than a bottom-up time-continuous estimate is something that provides upper and lower bounds and confidence intervals. if you find any of these things, please drop them here. for now i'll keep working with what i've got!

from ethereum-nft-activity.

I assume you have seen this already:

https://cbeci.org/

The methodology they present for lower bound makes sense to me. Choose the best mining device on the market, treat it like every miner is using it, and then apply that to the current network hash rate to understand GW. This can give you the lower bound, assuming all miners are using the best equipment.

Their method for calculating upper and best-guess estimates also seems reasonable, i.e. getting the average efficiency from the list of GPUs that are still profitable. Probably should be possible to apply to Eth, the hardest part is sourcing a list of mining GPUs I would think?

There is also this 2020 paper which was pointed to me by Sterling, it seems to have a fairly similar methodology, but includes a list of GPUs for mining Ethereum and their efficiencies.

https://www.cell.com/joule/pdfExtended/S2542-4351(20)30331-7

from ethereum-nft-activity.

@mattdesl the CBECI approach is exactly what i want, but for ethereum! i have looked into the 2020 TUM paper and i think i might just get in contact with all the authors. hopefully carbon.fyi and offsetra can adopt this approach, and we can move towards a better model.

from ethereum-nft-activity.

fwiw i'm rebuilding a model that is inspired by CBECI, but for ethereum. https://github.com/kylemcdonald/ethereum-energy

from ethereum-nft-activity.

Thanks Kyle – just reading it, looks fantastic. 👏

from ethereum-nft-activity.

Closing this issue as any Digiconomist references have been completely removed with my bottom-up emissions estimates, as of b674979

from ethereum-nft-activity.