Currently, there are NOT millions of Bitcoin users out there, if you take a look at the simultaneous online nodes statistics and/or client download statistics. Also nice "if everyone in ... would ..." ideas don't work - if they would even KNOW of Bitcoin, the price would soar much higher and your "stable" ~3/BTC would be like my offer of 21 EUR in comparison and a return to a rate of 3/BTC would be seen as "crash" + "failure".
Therefore we can assume the floor set initially will be much lower than the actual traded price, thereby setting a bottom boundary - exactly what the bot purports to do.
I don't set up such a company, because Tradehill is currently doing the same - and being under the jurisdiction of Chile I figure it might be easier for them than me under EU law.
Therefore Tradehill is willing to take on the risk. It could pay off for them, or the early adopters could put them underwater at a moment's notice. Hopefully it'll work out - this business is excellent for the ecosystem.
Also:
How/why does it hurt a currency if it is being exchanged for another currency at various prices or even if it is a meta-currency (like SDRs
http://en.wikipedia.org/wiki/Special_Drawing_Rights)?If you are not holding that money and the system is voluntary, you have the exact same thing as you already have on MtGox: People have a pile of USD in their account and when they think the price is right, they buy BTC. All you would do is to say: "As there are 21 million USD on MtGox right now that I know of, it does not make sense to set offers below 1 USD!". I still don't see how this would stabilize anything, if you cannot even remotely approach the current prices.
To approach current prices would require a far greater amount of money than this bot needs to set a floor. The more money is used by instances of the bot, the higher the value of the floor. As you said, it won't get to $10.00 per BitCoin the moment it's released - and that's a good thing. A sudden spike in demand only serves to artificially inflate the value to an unrealistically high level - where BitCoin's price currently sits compared to its purchasing power or backing (actual value). Hence, the floor would be unlikely to sit near the current price, as the current price has little to support it beyond speculation interest.
Most central banks DO have reserve requirements (
http://en.wikipedia.org/wiki/Reserve_requirement) but it again only works because they are central. You're free to try it out with an alternate chain. If you let people come and go at will, your "backing" is also completely unstable so there's no trust in it.
Most central banks have nowhere near the amount of wealth backing their currencies as they claim the total amount of currency in circulation is worth. This is an inherent weakness of fiat currencies as a whole. They are backed by primarily words. Anything less than complete backing is merely a psychological measure to engender trust to the asset. Similarly, this bot is meant to set a minimum price, to help BitCoin users know a minimum real value of their BitCoins - just like the real total value of the gold in Fort Knox is equivalent to the actual backing of the USD, but not to the actual value of every USD printed.
In the scenario you described, you have successfully devaluated BTC into nothingness, even though your bot was correctly working someone could still play pump + dump. To me this sounds like a complete failure of this bot.
My scenario also requires that everyone in the world simultaneously loses all faith in BitCoin except for one person running the bot - a highly unlikely situation. This situation doesn't describe a failure of the bot - it describes a total meltdown of the market. I chose it as it is the most extreme situation in which operating the bot could ever result in real asset risk to the user. Did the bot fail in the instance of total market collapse? Yes. Was the bot the source of the failure? No. Did every other thing in the BitCoin market also fail before the bot did? Yes. Did the bot fail to set a stable floor? Yes.
I concede the point than in the instance of total market failure, when someone sells every BitCoin into the floor, the bot will probably not work to force BitCoin to hold a price as people leave the fund and sell their BitCoins into the floor. In the case of 100% market collapse, the bot only slows the end game by a few minutes, giving the people in the fund a chance to get out and minimize their losses.
For a backing to work EVERYONE who trades fiat for BTC needs to put the same amount of fiat into the fund as well (I buy 10 BTC @ 10 USD --> I pay 100 USD to the seller + 100 USD to the fund to keep the value up).
For full backing to work, yes. For fractional backing, like we see in other currencies, no. Otherwise it wouldn't work in other currencies. Of course, really, all backing is full backing. We just pretend that the amount backed means more, and call it fractional backing. Because the chance of actually needing to back the entire currency is slim to none.
Then over time the fund will become large enough to really back Bitcoin.
With even $0.01 in it, the fund backs BitCoin, just not very strongly. Over time as(summing) the fund accumulates users, the backing becomes more powerful. Then the price of BitCoin stabilizes compared to an actual value it takes on. The bot is simply meant to speed the process along and make market collapse less likely in the mean time.
If you only pay in a lower amount of fiat into the fund than you trade, you have a fractional reserve and risk that amount you overpay (If you trade 10 BTC@10 USD and pay 80USD to the fund, over time the fund will only grow to a maximum of 80% of the then current trade price).
If only my goal were to lock in the current price, your math would matter. I'm not trying to set a permanent price - I'm trying to guide the price range to a stable one.
All in all your bot's a nice idea and all, but it has 3 shortcomings:
1) It won't work for ceilings (no BTC will be traded above x USD) as the amount of USD is unknown (depending of the money amount you take into account) or so high, that you will need far over 20 million BTC in that fund to bring it down to current sub-100 USD levels.
The math for setting the ceiling is tougher, certainly. I've been looking at a few theories of how to aim the ceiling, I should have something usable, or at least discussable, in a day or so. While there is a finite (though ever-changing and unpredictable) number of USD in circulation, there is no way to know how much of that is in the BitCoin market.
2) It requires people to publicly tell a lot of people how much buying power in fiat they have and pledge to use to back Bitcoin.
False. It requires individuals to tell a server how much they intend to put towards backing the BitCoin, and then to listen to the server's recommendation on a price. Because the bot is open source, different people will likely change the math to alter the level of the backing in their particular pool, or to a level they feel is more correct than the server, be it lower or higher. This organic process helps to set the actual floor.
3) It is voluntary, so panics + dumps can still occur (floors will most likely be magnitudes below current values and your system only works if noone needs their fiat money for anything else but backing Bitcoin. Ever.). You even gave an example in your last post.
The bot is a band-aid, not a panacea. Until the market stabilizes (and even afterwards) some strategies like pump-and-dump can still affect the market. I stand by the concept that the best thing for BitCoin stability is real purchasing power set in the form of merchant prices. The bot is voluntary explicitly by design for the reason of democratization of the floor price - which makes the floor a far more useful indicator of minimum real value. This means the floor can easily slip away - should people lose faith in BitCoin en masse. This is to protect the assets of the people that compose the floor. Why would I require them to stay in a tanking market? Furthermore, why would I do it if someone required it of me? People will stay voluntarily because they want to support the market, or leave because they don't want to. Just like people who aren't using the bot.
The panic/dump example I gave was a worst-case scenario: the complete simultaneous evacuation of the entire BitCoin economy. Nothing can protect BitCoin from sudden absolute devaluation except a
huge backing fund, first of all. Secondly, the chances of total simultaneous abandonment of BitCoin actually happening are slim, and if it did, the chances of any one particular individual being the last one running the bot are even slimmer. Thirdly, this final running instance of the bot could have as little as $0.01 running on it - as little or as much as the owner of that contribution to the pool chooses to let the bot use.
So, that is to say that in the case of total BitCoin implosion, one person running the bot
may be left with a few cents of net loss as a result of their decision to run the bot. But until the time that the entire BitCoin economy up and dies, the bot (or variants of it) sets a minimum value, as determined by the free market with voluntarily ceded shared buying power that is only pseudo-contributed to a backing fund (never leaving the hands of the person running the bot until the entire BitCoin economy collapses suddenly on itself while they're sleeping), to help stabilize the value for the long term, thereby stepping away from total BitCoin disintegration.
Anyways, I still haven't seen a single line of code yet! If it is such a great idea, just do it, don't listen to me and argue with me - prove me wrong!
I'm not a php coder, but if no one else does it, I may have to learn php eventually just to get this project off the ground. While the coding part is boring to me compared to the theory, I still see the bot as a necessary step towards stability, so how boring learning php is may not matter.