Doesn't look like a modified output selection algo to me

"Max output age per ring" is actually "Output age per ring" (i.e. plotted all rings' decoy outputs ages)

Detailed output age plot:

Both represent last ~24 days

Sorry, there is a mistake in both "Output age" plots, wrong sorting

The correct ones:

i wonder what the banding is about.. rounding maybe?

^ they are actually slopes raising consistently with the block number

4.74 rings per block are having all decoys less than 7 days old on average

If the attacker controls 50% of the recent outputs, they have 1/1024 chance to deanonymize of these rings after a week

*one of these rings

so after a week of attack 4.74 rings per block * 720 blocks per day = 3412.8 rings (<7d outputs) per day

~3 deanonymized 1-in txes per day

and growing if the attack continues

4.74 rings per block? is that how you are describing the attack spam level?

that's ~24d stats

4.74 rings per block are having all decoys less than 7d old

ah I see sorry

aight so where do things stand

Which things?

the things with the transactions

Ah, my understanding is that it was a misunderstanding with how decoy selection changes with increased transaction count over time, and that it’s expected behavior.

But someone else please correct me if I’m wrong.

The transaction count jumping 12k overnight was odd but explainable in theory, and not WAY past the norm lately.

gingeropolous: it is a bit unclear

we should just increase the ringsize to a BIG number

i mean, agreed. But decoys still probably need a better selection mechanism. And making it enforcable would be nice.... so i don't think its as easy as cranking the knob to 1024

oh look ^^^ :)

?

I echoed your refrain which resulted in you saying something more measured

this concludes my noise in this channel for the reminder of the year

<3

:) ah its late. there's no machinery on, all the math has been put back in the shelves for the night

sgp_> ArticMine: what's keeping us from increasing the base fee? <--- I am considering setting it close to the normal fee level. That is defined by being able to mine a tx ~3000 bytes. So about 4-5x the current value

The full set of txes has a much less pronounced bias than the original sample that was checked.

This is not really relevant though, it's still possible to do it regardless.

Prove that you really have the most critical thinking community. Come to t.me/get_monero and let's discuss things like adults.

how unexpected

I removed the message from matrix. I hope nobody is falling for that

Not visible on matrix ErCiccione

should show as removed

that's it. Shows as deleted

yeah that was my intent. They still see it on freenode