Exactly. QED.  bytemaster, please send the bounty to: 1CKeoT8vBDQDEpMHz5VAswV39pZJ2GTGYd.  Remember, I'm saving you 20000 bucks.

Edit: But if I have to sign me up for the game.  I'll take every currency anyone creates, instantly mint 100000000000000 units (or whatever the maximum currency units are) for 1 bitshare (or whatever the minimum is), therefore reducing the currency's value to essentially 0.  These currencies will never track USD.   And as a bonus, if I'm the only one doing the above, I'll get all the dividends.

Bytemaster: RE: Simulation :

1. "minting" currency is not defined in your game rules.

2. In fact again the actual rules of the BitShare system is not defined.  Your "rules" just define how we issue buy sell orders.

3. Choosing Mt. Gox exchange rate is just going to delay and obfuscate the situation.  Its theoretically (and practically when you look at crypto currencies) possible for the exchange rate to vary by 90% in weeks.  But I don't want to play this game for an entire year just to prove a point.  And just to remind you, the point was NOT to make money, but to prove that crypto-USD does not track USD relative to BitShares.

Please respond to my prior posting.  You have offered the 10BTC bounty.  Now you are changing the rules AGAIN, with this endless game.

 

Ok, is Q "fungible" or not?  That is, if I have 1 Q and you have 1 Q and we trade them did end up with the same value?  When you say "...backing his Q."  You are telling me that Q is not fungible.  If it's not fungible, its not a currency... Its a whole bunch of different currencies called the exact same name.  How many Q does a Mango cost? You can't price it because my Q is different then yours.

Ok so if we assume Q is fungible, then the sum of all the backing BitShares must apply evenly to all Q notes.


So if Q is fungible this is what happens:

You mint 100 Q and back by 100 BitShares.  Then I mint 1000000 Q and back them by .01 BitShares.  Now I don't need to sell.  I just sit on my 1000000Q and pull in the dividends coming from the 100.01 backing BitShares.  Your original 100 Q gets you (your ownership/total Q issued) 1/10001 th of the total dividends.  You get almost no dividends.  My 1000000Q gets me 1000000/1000100 (or almost all) of the dividends.

So Q is not a viable currency if the next guy can mint for less then what the previous guy paid.

However, if BitShares rises relative to USD, then to get the price of Q down relative to BitShares (on par with USD) I must mint for less backing.

Therefore Q is either not a viable currency or cannot track USD when BitShares rise relative to USD.

This is not about a wager.  This is about me putting a h*ll of a lot of time and effort into understanding your system which, frankly nobody is understanding and everybody seems to be giving up on because it is either not clear or not workable.


Yes, a fact that you have completely failed to address until I brought it up.  If you are going to design-by-question-and-answer please start admitting it.

You have not previously stated anything about price information.  That is my point.  Fiat currency price information is not entering the system.


There IS no trade from the point of view of the system.  There is only "I give you 100 Q".  The fact that you gave me 5 twenties from your pocket for them cannot be encoded honestly into the system.


My example starts with minting at Q==1 USD.  At that point someone wanted Q==1 USD.  But not now.  Now, the holder of Q wants it to diverge from USD and become infinitely valuable so he can buy a private island.   So I missed nothing.  And all subsequent minters want to mint Q for NO backing BitShares at all.  That way the second minter gets the dividends from the backing BitShares put up by the first! 

But if the first minter has any say whether to allow minting, he'll vote NO for any backing less than his original.



Great.  I'm looking to sell USD and buy Q.  So I'll decide the price and buy 1000000Q for USD .01 in BitShares.  Now there are 1000100 Q backed by 100.01 BitShares.  I get 99.999% of the dividends of 100.01 BitShares and paid .09% of the cost. 

This is not trade between 2 people, this is minting new Q!


Ok what you are saying now contradicts or at least limits the above.  So your rule sounds to me like I can't mint unless I offer a little bit more backing then the previous guy (and nobody sells me Q for that).  That is fine, and it means that Q will rise slightly in value relative to BitShares.  Since in my example BitShares has increases 10x relative to USD, Q will now be worth MORE then 10x USD.  So minting cannot bring the price of Q into line with that of USD.

You need to pick one rule or the other.  If you let me mint for less backing, I will print Q to worthlessness.  If you do not let me mint for fewer backing BitShares, Q will never descent to USD parity if BitShares rises relative to USD.  You cannot let me mint for backing USD because USD is not in the system.  You cannot let me mint for a backing of BitShares-equivalent-to-N-USD because BitShare-per-USD pricing information cannot be honestly inputted into the system.

If my description confused you, just note that you said nothing about USD in the above rule.  As I said before, USD pricing information is not entering the system so it CANNOT be part of whatever equation governs Q.  So Q tracks BitShares not USD.



No, I'll bid low and then choose to mint Q at a low price and sit on it for the dividends paid by your minting it at a high price (see example above). 

Now you'll argue that every minting has unique BitShares that back it.  So now we actually have N different currencies that are called the same name.  In other words Q is not "fungible"; it breaks a key property of a currency.  My low-backed Q is not as valuable as your high-backed Q.  Anyway, I'll still choose to mint it at a low price and sell it to some sucker who doesn't know the difference.  This will happen until Q is valueless which is why "fungibility" is an important property.



No it doesn't because the exact price paid in paper-USD cannot be honestly recorded in the transaction.




Start by actually posting how you do the accounting so we can put this to bed once and for all.


Sure...

Ok, here's my attempt to collect the 10 BTC bounty.  I am going to flesh out out a large issue in your proposal that I mentioned in my first post and hopefully make you see the problem.  Even if not you, it will make everyone else see it.  Or, if we engage in "design-by-answering-questions" I ask you to consider paying me some small portion of your development bounty.  Because good design is a the most important part of development.  I will publicly acknowledge any receipt of funds so others know you are serious with your bounties. 1CKeoT8vBDQDEpMHz5VAswV39pZJ2GTGYd

The issue is fundamentally "There is no way for USD pricing information to honestly enter the system."  Also, there are several smaller-but-still-critical-issues which I'll designate with [N].  So we'll call the above issue [1].  And in the end you'll see that you end up with a currency that tracks some multiple of BitShares, not USD.

Since you seem to think in examples, here is the problem illustrated by example:

Initial conditions.  1 BitShare = 1 USD,  1000 BitShares mined

Person A creates currency Q, mints 100 Q backs it with (say) 100 BitShares.  So 1 Q is worth 1 USD at this point. Now, the dividends he would be getting for the 100 BitShares are now diverted into paying the owners of the Q.  Right now, that is Person A so the system is in equilibrium.  Money is neither gained or lost by minting a new currency because if Q had not been created, Person A would be getting dividends from BitShares.  Other schemes result in runaway currency creation or a draining of the backing (the value) behind Q [2]. 

Now, BitShares like BitCoin are awesome so in 3 months the price goes to 1 BitShare = 10 USD.  Now those 100 Q, backed by 100 BitShares is nominally worth 10 USD each.  So when somebody bids to buy 100 Q for 100 USD (the original dollar-parity price) nobody is willing to sell for that price.  So the currency is created by someone who offers BitShare backing. 

But how many BitShares is the new currency backed with?  For parity with USD, it must be backed with 10 BitShares.[1]  But who decides that value 10? [4]  No voting scheme works because there is no way the holders of currency Q will agree to that backing level, because they lose money.  Regardless, any voting scheme causes people to vote in their self-interest, NOT to vote in a way that causes Q to track USD. 

Within the system, a trade of BitShares for USD simply looks like a BitShare transfer (the USD portion happens outside the network of course -- it could be physical cash).  Even if you add a field to the transaction to "report" the USD transfer amount, there is no proof that the USD actually changed hands [4].  Or even that BitShares changed hands actually.  It would be easy to transfer BitShares to (another account owned by) yourself with a fake USD amount to drive the reported exchange rate where ever you wanted it to go.


[1] Fiat currency pricing information never enters the system.
[2] runaway currency creation
[3] Proven the short term Q's price follows the backing currency.  Point [3] is that there is nothing else for it to follow...
[4] Bitcoin is based on distrust of all participating parties.  Your system requires some trust or centralization at a minimum to inject fiat pricing into the system, but trust/centralization could also solve [2].


A quote and rebuttal:


Within the example given above (where the backing BitShares are now worth 10x), why would buyers demand near parity?  Just because the name is crypto-USD?  In fact, if they did there would be NO SELLERS (at that price).  Crypto-USD owners would just sit there receiving the same dividend payment (denominated in BitShares), but those same BitShares are actually worth 10x USD.  How can the dividend payment be reduced to force parity with USD?  This requires that USD->BitShare price information be available within the system (and its not).

I'm happy to help; personally I think multi-currency is a key feature in the next generation blockchain (and sub-chains, described in several postings recently, but much earlier here: https://bitcointalk.org/index.php?topic=92398.msg1019392#msg1019392) .  PM me a google doc link or something if you want to co-author.

bytemaster, I have read you reply but do not have time right now for a point by point reply.  But let me ask a few questions and offer one piece of writing advice that I hope will clarify things.

Advice:  you tend to write anecdotal examples and stories and "prime" your and other people's mind to a sub-conscious assumption by using terms like crypto-USD.  In other words, you assume crypto-USD and try to argue backwards why it works.

Instead what you want to do is something like a true white paper:

Proposal:  A crypto currency X with these properties:
1
2
3

Effects:  Combining these properties creates the following interesting effects:
1
2
3:  [Ultimately I think you want to say] Payment of dividends creates a negative feedback loop that drives the price of X in THC (the "native" currency) towards parity with the price of the USD in THC.
This is clear because [now prove it].

Extreme situations:  [Prove the effects work during major market upset]

Result:  crypto currency X will track the value of USD


Questions:

1. How does information about the price of USD in THC enter the system?

2. When I was talking about anyone creating their own currency... you said at one point that "we" will start with just the major currencies.  Who is "we" here?

3. How does crypto-USD enter the system?  I get that someone buys it for USD, but I mean technically how is it actually created?

4. How does your dividend system cause crypto-USD to approach USD value?


And then an observation:

By not letting other people "mint" their own instruments, you tremendously limit the crypto-infrastructure.  Doing so would let:
A company issue its stock or bonds
An individual finance his own mortgage
(and a million variations)

Today, all current crypto-currencies (bitcoin, litecoin, etc) are really the same; they may have minor variations but hold the same fundamental role.  The above allows crypto-currencies to hold roles that interact directly with real life entities, rather then through the indirection of a floating exchange.

Whether an individual instrument is legally issued and enforced is not a problem to be solved by the crypto-infrastructure, just like the same is not solved by the paper these instruments are written on today.  In fact, a crypto-infrastructure really is like digital legal paper, allowing verifiable signatures and near-instant transfer.  But ultimately courts will decide about the legality of issuance and enforcement of a particular instrument.

 

timescale too short  you can find anything at that granularity

where do you guys find this stuff!?!?  seriously are there gif meme websites that u guys troll all day 4 pics??

I do not think that you structured this very clearly.  And this is causing errors in your own beliefs.

First of all, the best way to characterize this is in fact a true open source distributed "ripple" based on the bitcoin codebase.  Another definition would be a "native" colored coin blockchain.  Since colored coins is frankly awkward, I have been considering a blockchain that accepts multiple currencies deeply.  What is blocking me is frankly that I'd want bitcoins to somehow be a "native" part of the solution, not an external currency that requires a backer.

Point 1: You somehow think collateral makes your USD not need a gateway.  Uh, No.  There's really no nice way to say it but your thinking here is flawed.  Essentially each individual is a mini-gateway, securing the genesis of your "crypto-USD".  And BTW, nobody who really needs a loan can put up that kind of collateral... people put up cars and houses because they can simultaneously USE them.  A loan is truly underwritten by FUTURE the earnings potential of the person who received the loan.  Irrespective of your blue-sky scenarios, when black monday comes around, the people who received the $1 USD and essentially "created" the 1 crypto-USD have to make good on that and convert the crypto-USD back to USD.  Or be sued to get it.

Fundamentally, there IS no crypto-USD unless its the US government issuing it.  There is only someone's individual promise to pay.  And as we learned in the 2008 mortgage crisis it can be REALLY BAD to "bundle" these promises together under the assumption that they are equivalent!

Ok, now that we've got that out of the way, lets structure the system.

Base: 
  cleaned up bitcoin codebase with a single "native" currency, let's call it "thecoin" (THC)

  new transactions:
  1. Coin-type genesis.  A coin type is represented by a public key.  URL and hash of a coin contract are included.  The private key controls that coin type.  Could require a nontrivial txn fee of THC to miners, not to just one miner but distributed across the next 1000 mined blocks or something to reduce coin type spam.  However, just like there can be essentially an infinite # of bitcoin accounts, so with coin types.  This is good because as I was saying above, you can't combine fiat promises together.
  2. Coin genesis.  Signed by the private key, this allows new currency to be minted.  (THC txn fee)
  3. Coin destruction.  A transfer to the coin type's public key is essentially coin destruction because it gives control of the coin back to the issuer.

  Transaction changes:
  4.  Of course, all txns can have multiple inputs and outputs that take different coin types (coin type public key is of course included in the TXO).  This allows atomic exchange. 
  5.  A new transaction type called "market-based exchange" that is essentially the exact same as a normal txn but just marked as an exchange between anonymous parties (see below).

Exchange:

  For some reason you feel that bids/asks should be encoded in the blockchain.  There are huge issues with that, including blockchain bloat and matching speed.  But there is no need.  Bids and asks (actually they are ALL bids, just offering different currencies if you see what I mean) do not need to be remembered forever. 

These could be a special half-signed txn specifying what is offered and what is asked in return.  These are propagated throughout the network.  Someone trades by supplying and what is requested and signing, making a full txn.  The issue here is the half-signed txn cannot easily be withdrawn.  The only sure way to withdraw is to "double-spend" the underlying TXO so the $ the half-signed txn represent does not exist.  So perhaps bids have an "expiration block number".

Another choice is an unsigned half-txn.  The only issue here is your client needs to remain on-line to sign the txn if someone matches it.

Matched txns are committed to the blockchain, with a fee in THC paid by either/both parties.  Multiple matches against the same half-txn are resolved when the next block is found -- whichever full txn makes it in that block wins.  Note a higher THC fee would encourage more miners to use your trade (txn) as opposed to someone else's, which might make things very interesting...

Another issue would be bid flooding to attempt a DOS.  This could be handled to some degree by miner's fees -- essentially clients would refuse to forward the bid if the TXin does not offer enough THC as a miner's fee.  Clients can't tell the size of the bid because they don't know the value in USD of the coin-types being traded.  So we don't know if a txn is "dust".  So the miner's fee essentially needs be used to eliminate dust txns.  Clients could also refuse to forward half-txns that are far outside the trading range of currency-types as known by looking at txns committed to the blockchain and marked as being an "market-based exchange".

Another possibility is proof-of-stake.  You sign the bid half-txn with an account holding at least M THC, clients only hold N bids from a particular signature at one time.


Done!  Easy!  :-)

if you could exploit that then you could use the same knowledge to reliably make yourself money in the market w/o the bot.

But if the bot buys enough to move the market more then the trading fee, you could make yourself money by buying, triggering the bot, selling.  Given a flat market, this would on average tend to erode the bot's trading funds.  But during a bull market it might not.

its a lot more convenient to install virtualbox and hack around with a Linux virtual machine..

lol... on the same day Bitcoin more stable then both gold and stocks...

Here's to hoping I have to do another update tomorrow!  Gold price almost warrants it!

  I resisted the ignore button till now but this post pushed me over.  Wasting my time to go to bitcoincharts looking for even tiny evidence of something...

Gold is gasping... the last telegram was sent Feb 2006.  It takes a while for obsolescence to end.

the silverbox update (comparison from the beginning of this thread, March 13th, 2012, gold=1690, nasdaq=3055, Bitcoin=5.4):
Bitcoin is 118.60.  Gold is 1367.50.  Nasdaq is 3482.74
Bitcoin: 2096.30%
Gold:    -19.08%
Nasdaq:  14.00%
Gold Diff:  2614% advantage Bitcoin
Nasdaq Diff:  1827% advantage Bitcoin[/b][/size]

I'm glad I added the Nasdaq numbers though.  I think NDQ is probably a good proxy for inflation+economic growth, so the 1827% may be a pretty good indicator of bitcoin's organic growth.

I doubt it.  If there is ever a significant zero-conf market (especially a brick & mortar), they can pay mining pools to NOT allow replacement for TXNs originating from them.

This would reduce the likelihood of a double-spend dramatically.

As the credit-card business aptly shows, problems not theoretically solvable are often 99.93% (7 cents out of 100 bucks) solvable in practice http://en.wikipedia.org/wiki/Credit_card_fraud

Sure accepting zero-conf has a few problems.  But its a LOT safer then using a string of letters and numbers that any other individual can also use (credit card).  

Large merchants could accept cheap zero conf if they have a relationship with a significant amount of the mining resources:
Don't replace TXes coming from me.  Forward but flag as illegal double spend attempts (and notify me).  This makes it likely you can refuse service instantly or if it becomes a problem arrest the guy right there.

Given mining pools this relationship would not be hard.  Or the merchant makes his own pool.

If the double-spender was in collusion with a significant fraction (say X%) of the hashing power, and those miners were configured to NOT forward his TX then the double-spender would have X% likelihood of a successful double spend.  

That's a lot of work for a 1% (say) chance for a free $5 coffee and BTW never show your face at that store again :-).


So in practice we are fine for small brick and mortar payments.  But it may be a problem for $100+ brick and mortar txns.

China is coming... the quote below is what its like to really invest.  Especially a year ago when (like China now) service was thinner. 

tl;dr: It takes a month+ to do more then just stick your toe in.

Thanks!  I suppose the tl;dr; takeaway from my post is that trading this situation is just gambling.  So instead use the little dips or rises to re-position yourself for moves in either direction.



Maybe some gamer hooked his frag count to his trading bot?