search

nopara73 / WasabiVsSamourai

10 stars 1 forks source link

Exclude coin join outputs that are also inputs #1

Open ratpoison4 opened 4 years ago

ratpoison4 commented 4 years ago

Many coins are mixed many rounds to increase anonymity set. I think it is more fair to count such coins once. What we are really interested in is the flow of the system. Can you add this new metric to the output, please?

nopara73 commented 4 years ago

Can you show an example? I'm not sure I follow.

ratpoison4 commented 4 years ago

I mean remixes. If I mix the same coin 3 times, it is not the same as 3 different people mix 3 different coins 1 times. I think it is interesting to see both the "flow" of unique sats going through CJ as well as total volume of mixed coins. In the example above the flow would be 0.1 and 0.3 accordingly and the volume would be 0.3 in both cases.

nopara73 commented 4 years ago

For the record I've created this statistics (and nice graphs) for Wasabi: https://github.com/nopara73/WasabiCjEfficiency

ratpoison4 commented 4 years ago

Interesting.

Which coins are considered fresh? Will the coin be considered fresh in these cases?

nopara73 commented 4 years ago

I see your point and I was thinking about doing that, but Core is not very performant when you want to walk the chain.

CJ - send to new address - CJ again

Yes. Partly because with Bitcoin Core it's expensive to go back on the chain, partly because we cannot say otherwise, as it can be a Wasabi to Wasabi transfer too between two users. It would be interesting to compare the results regardless, maybe we can make a conclusion that it doesn't change the results in any significant way.

CJ - pay 0.001 to somebody - CJ the change

Yes, it's considered to be fresh, too. I wasn't thinking of this scenario. Same commentas above, it would be interesting to write stats for that as I suspect it doesn't significantly change anything.