Quote from: kjj on May 11, 2013, 12:44:45 PM

I don't think that this will be as stable as you think. I'm seeing a few places where you can expect positive feedback which will amplify minor swings into huge rushes.

Criticisms are always valuable. Would you be willing to articulate them? There are also some corrective mechanisms not yet described here which target instabilities I'm presently aware of, like the ability under certain circumstances to short BTC via the token protocol at next to no risk.

Quote from: kjj on May 11, 2013, 12:44:45 PM

Your whole security model depends on backers paying more than the market rate, in hopes of being paid an even bigger premium later. Since they only profit when the value of the backing instrument rises, they would be better off just holding that instrument directly, rather than using it to go upside down in a new venture.

Backers don't create cvTokens in order to hold them. They are using the cvToken system as a type of secured margin trading, so they only create new cvTokens when they can be sold immediately for BTC. If BTC rises, they profit more than if they had merely held BTC. A drop in BTC will put them upside-down, it's true. But that's the speculative risk they took in hopes of higher profits. In that case a cash-out acts like a margin call.

Creating more cvTokens also increases BTC value through network effects. People who wouldn't have bought BTC will buy BTC in order to buy cvTokens (even if they don't realise they're doing it because an exchange does it for them). And although people want to be able to cash out cvTokens, they don't really want to cash them out unless the cvToken is losing popularity or the collateral is suspect. Also, backers have the right of first refusal on above-market value bids in the distributed auction. This type of opportunity seems hard to match for someone who believes Bitcoin will rise.

There is another mode backers can operate in, where their risks and profits are reduced to exactly those of holding the instrument, but while simultaneously increasing the available quantity of cvTokens. It's achieved by treating deposited BTC as traditional collateral, and holding the full value of the issued cvTokens in the non-digital basis as soon as they are sold at market. This is a role that businesses have an incentive to play in order to facilitate a cheap, stable, fraud-resistant payment method for their customers--especially if they are holding BTC already for some other reason. It's called "gift-card mode", and the two modes can happily co-exist.

cvTokens aren't perfect. But they do have three key points in their favour:

1. They have all the characteristics of bitcoins, except that they separate the speculator and market participant roles somewhat. This is where they compete, rather than against fiat/bonds/other traditional instruments. They let people use bitcoins without holding bitcoins. Any improvement in stability over bitcoins is considered to be a useful property for cryptocommerce.

2. If the reference point for "stable" in a cvToken is an inflationary currency, bitcoins will tend to appreciate against them over time (all other things being equal). This gives the backers a slight edge over straight speculation at a price that consumers seem willing to pay--an incentive for stability.

3. Successful cvToken colours tie up BTC, reducing the supply and increasing the price--a positive feedback loop for the token's collateralisation if they are eventually used widely. Although this additional effect would disappear if a cvToken was rapidly cashed out, it allows for different cvTokens to indirectly back each other; especially if some of a cvToken's fleeing users flee to another token.

Quote from: kjj on May 11, 2013, 12:44:45 PM

At least that's what I'm seeing from reading this thread. I'm not entirely sure that I understand everything as well as I should, and I'm automatically skeptical of value fixing schemes since none has ever worked in all of history, at least so far.

To set expectations appropriately, I would expect cvTokens to be less successful at value fixing than first-world fiat, and more successful than Bitcoin or gold. True value-fixing is an extremely hard problem.

I think it's also valuable to identify the shortcomings of the system, which I haven't yet had a chance to detail. The largest one is that the ratio of interested speculators to interested users would likely limit availability of a cvToken, especially during periods of high BTC valuation. The nature of a cvToken bunches demand into a very tight band, which means users won't just pay more to get them (unless retailers offer a "cash" discount for them. Even then the effect is small). This could have a stifling effect on their utility when they can't expand in volume quickly enough to match consumer interest.

Another problem is that they are hard to launch. You have to tie up a lot of capital for a long time, and then attract a significant number of merchants before anyone considers your token worth a second look. This is mitigated slightly by the ability of larger businesses to launch them as "gift-card currencies" and then convince other locations to accept them. But they lack the incentivisation for adoption that Bitcoin has, with its limited supply that can turn any holder of BTC into an evangelist. Again, "cash" discounts would help, but cvTokens are still less viral than most altcoins.

Finally, implementation is a chicken-and-egg problem. The protocol is inefficient unless additional transaction types are created to accommodate it. But introducing additional transaction types into Bitcoin is ill-advised unless there is a clear, non-esoteric use case; and it's hard to know how widely cvTokens would be adopted until they are up and running. Initial implementation could be on an altcoin, but market depth in the underlying instrument would be thin, so it may not be indicative--especially since few actual businesses will be interested in accepting altcoin-based cvTokens. And altcoins can be....scammy.

cvTokens are not magic. But I do believe they offer strictly less trust than Ripple and strictly more stability than Bitcoin, which seems like a valuable thing for commercial use in particular. Perhaps a simple programmatic model would be easier to probe for vulnerabilities. I'll see if I can put something together soon. I hope this post was coherent--it's past my bedtime.