<c​haser:monero.social> that'd make the tx more expensive too, although sublinearly. btw, people have wondered before about constraining outputs to 2 (monero-project/research-lab #114), although I did so for transaction uniformity. I didn't consider its effects on a black marble attacker. it'd make it ~3.5x more expensive to create 16 outputs (one 1/16 tx vs 8x 1/2 tx's).

<c​haser:monero.social> sorry, I didn't account for tx weight. that way it's ~1.8x more expensive.

<c​haser:monero.social> they could create the 15 non-change outputs by "burning" a piconero for each (practically free), then there's nothing to consolidate, and no clustering either.

<c​haser:monero.social> *burning it to addresses external to the wallet (since AFAIK sub-dust enotes are spendable if you have another enote that can pay the fee to have both as inputs to a tx)

<c​haser:monero.social> thank you Rucknium, this is essential research, and a fascinating read even in its draft form. you know things are getting serious when research on blockchain privacy cites a paper on geomagnetic storms.

<a​rticmine:monero.social> The output distribution was very helpful since it provides evidence of an actual attack. Especially the ratio between 1/2 and 2/2 transactions

<a​rticmine:monero.social> Thank you Rucknium

<a​rticmine:monero.social> 👍