Topic: ChromaWallet (colored coins): issue and trade private currencies/stocks/bonds/.. - page 25. (Read 97112 times)

That is the open and operative question.  GLBSE had bots, bidding and asking.  It seems reasonable to assume that any P2P solution would quickly see the same, right out of the gate.

It's possible to make a p2p solution for laundering where parties would mix their coins via swap transaction. Only your peer will known where your output went.

If considerable number of participants are honest, i,.e. they'll never reveal details of swaps they do, after a large number of rounds coins will be well mixed so it won't be possible to trace inputs to outputs.

E.g. let's say we have 100 honest launderers and 900 FBI agents. After, say, 1000 rounds of coin mixing won't be able to trace particular output to a particular launderer, only to a group of 100 of them.

If a lot of people use this laundry (for example, it might be a feature of their client software: it will perform coin swap transactions even if laundry is not requested, just to protect privacy of others) it would be impractical to use taints, as a huge amount of coins will get "tainted". Thus tainting will neither help you to catch bad guys, nor you can deny receiving tainted coins as you' won't be able to accept coins of honest customers.

So distributed laundry is possible, it can be 100% secure in sense that you cannot lose your money, and it provides reasonable mixing.

The only problem is blockchain bloat, but mixing transactions are prunable, and also it is possible to do aggregation.

So, yes, I think that tainting won't work if enough people will help with laundry...

But, again, this has absolutely nothing to do with colored coins. Can you please move that discussion into some other thread? Thanks.

The colored coin algorithms do not violate any part of the bitcoin protocol. Instead, it is exactly how bitcoin works: you can trace every satoshi back to its generation to make sure there is no counterfeit. The untraceability of bitcoin is just your fantasy. If you believe this is the end of bitcoin, you don't understand bitcoin at all and you should sell.

Back to genius melting pool. You cannot broadcast your melting transactions as usual. You have to mine your own block to include them. These "hidden transaction" could be spotted out by monitoring the network. If there are lots of such transactions, that would looks suspicious. The difference between a legitimate miner and a laundering miner is: those hidden transactions will never go to a legitimate miner.

Also, to launder merely 1BTC (which is basically nothing in the world of money laundering), you need 200 transactions. With only 144 blocks each day, how many TH/s do you have?

You mention TOR. Yes, a much easier way is to launder through SR.

Where are people getting all these bots?

I am curious because despite seeing sometimes some quite large and long lasting arbitrage opportunities no bots seem to be closing the opportunities. Also it would actually be useful for things like when someone offers to sell asset X for asset Y and someone offers to sell asset Y for asset X, because although some fortuitous choices of scale and price can sometimes make obvious matches across the two markets often it is rather beyond what generic order-matching could reasonably be expected to handle.

For example if I offer on the hundred-scale market to sell fifty lots of a hundred asset X at 3 asset Y per hundred and meanwhile someone offers on the single unit scale to sell three individual units of asset Y at 33 asset X each, there is a profit of 1 unit of asset X to be made by cross-market matching between the X/Y scale 100 market and the Y/X scale 1 market...

-MarkM-

There were bots on GLBSE (and there are currently bots on currency exchanges).

One basic thing bot can do is market making. E.g. if existing bid is 6 and ask is 7, bot can place a bid at 6.1 and ask at 6.9.

If volatility isn't high, but there is some trading volume, bot can earn some money without a lot of risk.

Essentially risk is limited to bid/ask he puts, but he can buy and sell some amount many times through day.

Also bot can do arbitrage... For example, currently on litecoin global (https://www.litecoinglobal.com/security) there are two bond issues (ESECURITY-SA, ESECURITY-SA2) which have identical condition. Bot could exploit temporary price imbalances, e.g. if ESECURITY-SA is traded below ESECURITY-SA2 bot can sell ESECURITY-SA2 and buy ESECURITY-SA: his net position will be the same, but he earns some money.

Why? What would such a bot do? I know HFT bots exist on regular stock and currency exchange markets but they are exploiting very specific issues. Debt issued on a P2P bond market still has to be issued for some kind of purpose and under some kind of identity so you can have confidence you'll get it back.

If I think about the kind of economies we see today in bitcoin-land, I'd expect a P2P bond market to be used at the same kind of volumes as GLBSE was. I don't recall reading about bots operating on GLBSE. What algorithms would they use that require HFT?

Interesting...

Lol. Have you ever heard about TOR?

You have no idea about the system I proposed. I said that you, as a meltmining pool, tweak the bitcoin software in order to prevent the meltmining transaction to be broadcasted until it gets added in the block you found.

You understand nothing.

A meltmining laundry can melt a small number of coins every block it finds.

¿You say you can adapt your taint software to find the unusual fee transactions?

Ok, I say that I will send 2 to 5 tainted BTC fees in the next block and wait to be mined by p2pool or Deepbit.

¿What do you think will the miners do when they start getting their accounts freezed in MtGox because your taint lists and software? They can't dodge the tainted melting coins fake transaction I made and they can't prove they aren't the melting launderers.

So the tainted coins list/software adoption is the beginning of the end of bitcoin.

The standard fee at this moment is only 0.005BTC. To launder 1BTC without paying suspiciously high fee will take 200 transactions. You can't put these into the same block or it will become too obvious, and you need to generate more than 100 blocks for laundering only 1BTC. Even worse, the FBI may just monitor all unconfirmed transactions on the network. If they find many "hidden transactions" that were not broadcast to the network before they appeared in a block, they will further investigate that particular miner.

FTFY

There are some tricks you can play though if you have the mining power to create a block in a reasonable time. Then you can deliberately include transactions which increase the noise level and make it hard for any tracing software to infer the origin of coins. If done right, the tracing of the coins you want to launder becomes unpractical.

He absolutely have no idea what he's talking about. He thinks coins are magically laundered by using as transaction fee. Transaction fee is nothing more than a special type of output. If someone tries to melt some dirty coins by sending unreasonably high fee, the block reward will simply become dirty. Not to mention that setting up a coin melting pool cannot stop people using colored coins, unless he is able to spend coins that are not belonging to him.

However, I still hope he will set up a pool like that, so I may be able to grab some of those easy coins.

I think this is the main point of colored bitcoins, enable creation of new assets, and using bitcoin infrastructure as a digital asset grid, to enable people (and organizations) to hold value digitally, in the cloud, without any need for regulation or counterparty trust.
Today the biggest problem in finance is counterparty risk, and the reputation of a financial institution represent its counterparty trust and is the most valuable asset a financial institution has, and the biggest barrier to entry for new companies therefore for new technology. By decreasing requirements for regulation and counterparty trust the progress of technology will accelerate dramatically.

It seems likely that people will hook up bots to distributed bonds as soon as they pop into existence.  Some sort of mitigation strategy will be needed, at a minimum.  Otherwise the P2P network will be flooded with sales offers.

Bro, again, this has absolutely NOTHING to do with tainting. You have no idea what you're talking about.

In the last 2023 blocks, only 26 were orphaned. That's about 1,3% odds of loosing all the meltmined coins (to a lucky miner). I prefer a 1% chance of loosing the melted coins than your tainted coin system. Tainted coins will affect all the bitcoin economy badly but meltmining pools only have a 1% chance to loose the block. I insist, tainting coins is a very bad idea because it will hurt all the bitcoin economy and the solution is not so risky for a melting transaction.

I suspect HFT should be a non-goal for v1. It's hard enough to make things work just for regular trading by long-term investors.

If you want to also explore designs that could do HFT between a set of parties, Satoshi had the following to say on the topic:


I took this description and recast it into "micropayment channels" with WiFi hotspots as an exemplar, but it's actually a general technique.

That's really clever, Jeff. I might describe that as P2P with voluntary, spontaneous federalization.

pybond follows the distributed bond design laid out in that post -- including the P2P and DHT bits.

I am thinking there will be a third component as well:  P2P network will permit traders to advertise "coordination points", a single address where traders may connect for high-speed trading between themselves.  These coordination points may be created or destroyed at any time, and are self-organizing among participants.

There is a plenty of reasons.


I'd like to see it too, but that doesn't mean that there is no room for other solutions.

Particularly, I want to make one based on simple information exchange protocol on top of HTTP. (I.e. pieces of transactions and meta-information is serialized and stored/retrieved via HTTP.) It can easily scale from fully centralized to somewhat-decentralized to full p2p.

There is no reason to start with full p2p: it's far easier to start with something centralized and then scale it. (This isn't the only argument, though.)

(I should note that I'm not a stranger to p2p stuff: I once implemented an ed2k client, and I read articles on DHTs and whatnot... It's just that HTTP provides cleaner solution by decoupling of transport layer from layers which implement application semantics.)


As I already noted, p2p or not p2p is just a transport layer. It, broadly speaking, does not affect transferability of assets, trust requirements etc.