[polarized]

The section 3 vulnerability traces (aka "de-anonymizes") precisely no transactions at all. It indicates that the probabilities across potential outputs sources are biased, but does not offer any method at all to identify any actual source.

The estimate of the bias given in the paper for the current default and typical usage is that the most recent potential source has a probability of 45% instead of the ideal 20%. In fact this likely applies to 100% of current transactions, not 80%.

This is a known issue, and not ideal, and the quantitative results in the paper are helpful, but the paper does not show what you claim it shows.

[snovv_crash]

Yes, that's section 3. What about section 4?

[polarized]

Sorry my mistake. My comment was about Section 4, not Section 3.

RingCT is immune to the methods in Section 3 as stated in the last paragraph of Section 3.

Section 4 does not trace any transactions. It identifies a probability bias which make the ring sigs less efficient, but still functional.