Topic: [ANN] Joinmarket - Coinjoin that people will actually use - page 10. (Read 84864 times)

There goes the immutable Law of Supply and Demand: https://en.wikipedia.org/wiki/Supply_and_demand

...as strong as the Theory of relativity.

Talking about Joinmarket (and also privacy), is it fully compatible with TOR?

No, I don't think so.

If there is more supply than demand, the price drops. That's normal, healthy, and to be expected.

You seem to be suggesting manipulating the market to force the prices to be higher than the market supports on its own. If we're currently short of "takers" then low prices is what we need to attract more of them, wouldn't you agree?

It's not clear volunteers do much harm. They're giving up an opportunity cost to earn money. Their hardware, electricity and internet costs them too. And eventually they run out of UTXOs and can't coinjoin anymore until they get confirmed.

There's also this idea to quantize the market prices to protect against front running. https://github.com/chris-belcher/joinmarket/issues/14#issuecomment-143509788 It's similar to what happens in stock and futures exchanges to stop someone placing an order one cent below everyone else so they can extract information from the market.

Is it sensible to require JM makers to do some proof-of-work to disincentivize the destruction of the market by the volunteers?

The address which the burned coins were sent to is mentioned in this deed bundle:

    http://deeds.bitcoin-assets.com/bundle-371897.txt

The first deed says:


and the second says:


Both deeds are signed with PGP key ID 7CDA03F9. When I verify the signature, I see:

    gpg: Signature made Fri 28 Aug 2015 05:50:16 AM PDT using RSA key ID 7CDA03F9
    gpg: Good signature from "Adlai Chandrasekhar <[email protected]>"

This was shortly after adlai had a falling-out with MP:

    http://log.bitcoin-assets.com/?date=25-08-2015#1252200, in which adlai dares agree that assbot is too spammy

    http://log.bitcoin-assets.com/?date=26-08-2015#1252289, in which MP asks him why he thinks he belongs in the channel

    http://log.bitcoin-assets.com/?date=27-08-2015#1253642, in which MP unpersons him

So it looks like the 'odd' coinjoin was adlai's parting message to #bitcoin-assets?


In the past 7 days my bot participated in 77 transactions for a total of 48.66 BTC.

No, I doubt that you'd get takers if a few makers destroy the market with their "charity" and a price race to the bottom early on.

What you're saying is you'd be happy to see fewer market makers and higher costs, so that you can feel good about it, as opposed to actual users being able to get proper service at a lower price.

I would also suggest that you look out for your own self interest.

Having said that, taker's use a significant amount of randomness in choosing offers to coinjoin with, so somebody offering zero fee in no way means they'll coinjoin with everyone. Of course there's also an upper-limit to how many transaction you can do per time, because of the requirement to wait for a confirm (or even many confirms when that feature is implemented)

If there are people that do this stuff for free, you're disincentivizing makers.

I wouldn't want to be a taker in a market that consists of a handful known enthusiasts instead of a huge amount of anonymous people fighting to find a fair price.

I truncate the csv and logs of anything older than 72 hours.  I do it manually each day, though, as I want to make sure to be able to save anything needed for troubleshooting.  The data I gave was from the current csv file.

Honestly, the small returns don't mean as much to me as seeing this project become commonly used, so I don't charge a maker fee to encourage more taker participation.  I buy coins pretty often, if I have extra cash, and I have daily mining income, so those are the coins being used by my bot, for the most part.  Of course, I don't keep my whole stash on the network, even though I feel like it would be relatively safe if I did.  I don't really have a need to tumble them, but I believe strongly in privacy.  Essentially though, yes, it's a slow and low cost way for my coins to be mixed, while also providing additional liquidity to the market.

My hope is that I won't be replacing another maker when my bot is used, but will just be one additional maker available for free to increase the maker count for each transaction.  That, of course, depends on the mindset of the takers however.  If they want to only mix with me as a sole maker due to having no coinjoin cost, I can't prevent that obviously.  Anecdotally, it seems most people are using above the default maker count when my bot is included.  I think most people realize that mixing with only one maker is not very private.

I wonder when and if we will start to see maker bots that pay takers.  I could imagine at entity (think DNM vendors, admins, etc.) with a "dirty" large stash of coins that constantly require tumbling would be willing to pay a fee to takers to encourage people to use their unclean coins in a join.  It's not my game, but I think it's only a matter of time before someone integrates joinmarket into their wallet structure for this purpose.  While I am sure that some people don't desire this (some joinmarket users, gubermints, etc.) , the beauty of an open-source, decentralized project like this is the freedom to do so.  This possibilities of this project are very intriguing and I am quite thankful for the people putting their time into it.

When I get cash, I often wonder how much cocaine and stripper ass the bill has seen, but I am glad I can use it still regardless, and virtually nobody that I pay it to will care about it's past history.  Bitcoin should also be like this, imho.

Don't you keep the yigen-statement.csv file?

I assume your zero-fee market maker is a slower cheaper way of mixing coins than running tumbler.py

I usually only keep my logs for 72 hours, and rarely do I even look at them unless there was an issue, or I am bored.  That said, over the past 3 days, my bot has done approximately 25 joins for a total of about 13 BTC transacted.  My bot has earned exactly zero satoshi during that time, which may explain it's higher transaction volume, relative to belcher's stats.   

There's a project for eventually doing that https://github.com/adlai/cjhunt
Alternatively me and others who run market makers could say how many CoinJoins they participated in last week.

Edit: I'll start, in the past 7 days my bot participated in 23 transactions with 1.94btc in total passing through my coinjoin outputs.

How about publishing some (weekly/monthly) "traded" volume statistics at Joinmarket in the (near) future?

That would be good to spread the word and to get more attention (and liquidity) to this project too.

Awesome job, this is a great update!

This is an important update, great job!

I've merged the code which implements a timeout for takers. If a maker doesn't respond for some time (default 30 seconds) the taker will connect to a new maker and try to finish the coinjoin. This means a tumbler.py run will not be stopped by one nonresponding maker. In fact in testing me and others ran very long tumbler.py runs and fixed several other bugs.

JoinMarket can now very robustly improve your privacy.

That's not really what BIPs are for. JoinMarket doesn't need any new consensus.
For advice and ideas I'd already talked to many people before I started coding.

This is a nice project.

So interesting that IMO It should even be suggested as a Bitcoin Improvement Proposal - BIP.

i wonder if this concept can be combined into the Multigateway project?