Remember, we had already proved amplification, if by amplification you mean that the attacker can gain knowledge of additional payment linkages beyond those revealed directly by knowledge of his own outputs (our research bulletin called this a chain reaction). http://lab.monero.cc/pubs/MRL-0001.pdf

In that sense I would say you have proven amplification, since your method identifies all of the cases identified in our paper.

It seems likely at this stage that your method will also demonstrate additional amplification, but two things are needed:

1. To show the additional amplification indeed happens. This seems relatively easy.

2. To show that it is a viable method (resource requirements to achieve #1 are reasonably feasible). Not sure about this one, but I would say >50%

Almost there.

BTW, I don't think the BBR feature is directly effective against Sybil attackers. It prevents disclosure to the public (for example by someone spending without a mix) but a Sybil attacker can easily create these min-mix outputs, and then you will choose them as your "safe" mixin. But the damage is done as soon as you use the attackers outputs in a mix (he knows his are not the real ones so yours is identified by elimination); whether or how he spends does not matter.

The BBR feature may slow down the rate of chain reaction sunder some conditions, I'm not sure.

In any case we move on to...


Agreed, except I would say the skepticism is greater on the math, less on the code.