unsedd, do you know of any GPU accelerated fuzzing?

Fuzzing doesn't seem like the kind of thing that'd benefit from that.

Hey everybody and @sgp_

Today I've been thinking about this notion of all-coinbase transactions, so that you know funds came from *_a_* miner, but you don't know *_which_* miner

I can't recall, has the decoy selection algorithm for these all-coinbase transactions been discussed?

Could just use timing from the standard algo, but I'm wondering if anybody has intuition about what the coinbase spend time distribution might look like.

s/know/statistically suspect/

Isthmus: medium.com/@JEhrenhofer/lets-stop-using-coinbase-outputs-da672ca75d43

I continue to advocate for a separate selection algo for coinbase output spends

Idk if you'd gain much from a new distribution, since most coin base are probably pool payouts. Those are easily identifiable regardless of distribution

UkoeHB_: did you read the post? I adocate for normal transactions never selecting coinbase outputs as decoys

Users never spend coinbase outputs

Isthmus is wondering about the ring member selection if coinbase were isolated

Best way would be to pick clusters from other pools (or your own pool), but that's pretty hard/annoying to implement and maintain

Yeah, for this thought experiment, let's say that we're not changing the protocol but pools are voluntarily adopting best practices for mixing their coinbases. In that case, what decoy selection algorithm should they use

Aren't the public pools indifferent?

Thought experiment

They publish all their payouts and blocks mined anyway

Yea, I wanna get away from that model though.

It matters for the change outputs though. Those can be improved by selecting decoys sent to miners in the same transaction that the change output was generated in

I don't want to publish blocks and transactions though. Instead of publish all of the blocks and transactions, will just publish an exchange rate between hashes & xmr payout.

Instead of every individual miner auditing the whole pool, each just audits their own contribution and payout

And if they feel that the ratio is unfair, they go elsewhere

is it possible to audit ex-ante when everything is published?

i.e. would changing the audit procedure not be neutral?

Not sure if I get what you're asking.

Any pool can choose whatever audit procedure they want, and users can choose whichever pool uses the one they like.

I dont fully understand pools. Does publishing everything mean a prospective miner can audit payouts before mining anything, or do they have to earn payouts first? If the latter, then changing the audit method to be exchange-rate based would be neutral for miners.

I think that's the principle? Looking at supportxmr.com now

In the bottom right you can click to see 2 lists

1) all of the blocks they mined

2) all of they miner payouts

My thought is pools are unlikely to adopt anything that loses them a competitive edge against other miners, such as possible better auditing

s/other miners/other pools

What does an honest pool have to lose from better auditing?

Note better doesn't mean able to catch more. Better means privacy preserving

I mean weaker auditing

idk what you gain from publishing this stuff, since it could all be faked..

or partially faked

Are you talking about publishing all the blocks and transactions?

yah

Yea, even putting aside reliability of self-reported blocks and payouts... I think that the publish-everything-so-everybody-can-verify approach makes a lot of sense for Bitcoin and the transparency=auditability ethos of many cryptocurrencies. But the paradigm isn't great for privacy coins (especially with decoy based privacy)

Imagine that you go to the gas station and want to know whether your interaction with the pump is fair. Do you ask the owner how many gallons of gas they sold in the last 30 days and how much total profit they brought in?

Nah, in that context, my notion of fair is whether the amount I am charged matches the price on the sign multiplied by the volume of gas I received.

Your point makes sense if you're OK doing the transaction before checking. Keep in mind that you don't have a "cancel" option once you're chared.

charged.

This is one of those situations where I think having data from transparent chains could be useful

e.g. data on spend ages for coinbase outputs only, to get a sense of differences from overall spend patterns