All,
Here is the list of block-chain rules under active development. I will payout a 0.5 BTC bounty each time someone finds an 'attack' on this rule set that results in me changing the rules. Note, I may change the rules at any time in response to my own efforts at finding bugs. I will then kindly request that everyone who finds bugs re-invest their 0.5 BTC bounty into the project and receive 500 BitShares instead.
1) Anyone may sell short any asset on the exchange provided:
a) there exists a buyer who is willing to take the other side of the trade *at built-in market price*
b) after the short sale, they have 3x the value of the short sale as collateral. (actual collateral requirements subject to change)
2) Any short position may be redeemed by the market when the value of the collateral falls to 1.5x the value of the short.
a) half of the position is sold and the proceeds are used as collateral for the other half of the position (unless collateral would still be insufficient, or the balance would be 'dust')
b) there is a 5% fee paid by all shorts which force the network to cover their position. The goal is to
encourage any short to keep their margin sufficient or close out their position early. This fee also
motivates miners to give closing out of short positions priority over most (all?) other transactions.
4) All 'short' positions must be closed in full before any of the collateral may be spent.
- as the price of an asset falls, the effective interest rate paid to longs will go up as the ratio between short and collateral grows.
- this will cause increasing opportunity costs for the short position which will motivate them to cover the entire position
and re-open their position at a new base.
3) Dividends paid on BitShares held as collateral are redirected to individuals who went Long (taking the other side of the trade).
a) As a result, crypto-USD pays 1.5x to 3x to dividend rate rate as BitShares.
4) Users place their bids / asks into the blockchain as 'outputs' that can be canceled by spending them, or accepted by
spending them as part of a transaction that satisfies the bid and market requirements.
5) No block may execute a trade below the highest bid or above the lowest ask in the block chain.
- The order in which trades are executed is based upon 'price' first, 'fee' second, and otherwise up to the miner
- If the highest bid is greater than the lowest ask, then the transaction occurs at the *bid* price.
6) No transactions that contain multiple currency units are allowed outside of the bid/ask system.
- This requirement may be lifted after a careful audit for potential attacks by circumventing the 'market'.
In the event that the price of an asset changes so rapidly as to blow through all 'margin', the Longs will eat the losses.
- this is the justification for the higher dividend rates paid to the longs and the opportunity cost incurred by the shorts.
- No system can gurantee 0 losses and BitShares is no different.
9) All trades on the built-in exchange incur a 0.05% transaction fee that contributes to mining fees / dividends. This fee is
designed to minimize the profitability of 'rapid trading' and generate profits for the BitShare holders.
- minimum transaction size limits will also be imposed (like Bitcoin) to prevent dust spam.
- minimum transaction fee just like bitcoin also applies.
10) All dividends paid to 'transaction outputs' in the last 120 blocks are recaptured as mining fees, spending these unconfirmed
dividends would result in chain-splits invalidating the tranaction.
- as a result, those who spend money rapidly will receive no dividends, while those who save will receive the dividends.
11) Users may transact in any currency just like they do Bitcoin (provided all non-market transactions only deal with a single currency).
- this includes trading of their short position.
12) No block may clear out more than 5% of the value of all open bids/asks for a particular asset.
- this prevents certain classes of attack in 'thin' markets.
13) A maximum reduction in exchange rate of 5% per block. The goal is to give market participants time to add collateral or buy the
dip. It would also prevent certain types of attacks based upon 'rapid manipulation' of the price.
14) No trade may occur unless there are at least N? bids/asks capable of 'reversing' the position.
- this aims to prevent attacks on new issuance and insures that there exists a deep enough market to justifiy creating
a new asset class. It also 'halts' trading when the market gets thin.
- the definition of 'capable of reversing' is still TBD
15) You must wait 10 blocks before spending the output of a trade.
- if we allow people to immediately spend with the proceeds of a trade then, chain forks could be exploited to
reverse trades, manipulate prices, and cause losses.
All of the rules above ultimately mean that trading can only occur at 'human speed' and all high-frequency trading will be
forced off-chain. Trading is not 'free', but cheaper than any current exchange.
In particular I am looking for ways that the market can be manipulated that do not also apply to traditional markets. Some avenues of
attack that must be considered:
1) What would happen if someone had 51% of the hashing power?
- they could control what bids made it into (or out of) the blockchain.
* prevent people from canceling bids.
- they could control who got want bids.
* play favorites
- they could do anything they could do with Bitcoin.
2) What would happen if someone had 1% of the hashing power?
- they may gain some advantage in picking/choosing bids.
- would this motivate professional traders to invest heavily in mining?
- would the competition ultimately be good for the network?
3) What weaknesses would be exposed by having all short positions and margin available as public information?
* Somone with significant capital could 'trigger' a short-squeeze by bidding up the underlying asset.
- is this mitigated by not allowing uncollateralized shorting?
* The short-squeeze would then enable new shorts to sell at higher prices (offsetting their attempt to push it up)
* In theory someone could take advantage of such moves... but only if they could move fast enough between
short and long positions to 'head-fake' the network. Because all positions require 6 confirmations before they can be adjusted does it
make it difficult or impossible to benefit from this kind of manipulation?
4) In theory all 'shorts' are naked, but backed, and are ultimately settled in BitShares. What are the implications?
- don't trade in illiquid, rare, or non-fungibile/divisible items. It would be up to the Longs to assess this risk.
- the total 'short' position for any asset class is public and therefore can be audited. If the total short position
is too-large the market will respond by discounting the 'long' position from face value.
- how does 'naked' shorting enable manipulation? In theory, someone with a large amount of capital (BitShares) someone
could keep selling into a market. This would result in pushing the price down but would also drive the dividends paid
to longs up.
- Because longs are not buying with leverage, short-selling to push the price down CANNOT trigger margin calls and further selling.
- another way this can be viewed is that the 'shorts' are 'borrowing' the USD from the longs and are posting collateral and
paying interest to do so.
- any naked-short is ultimately has to cover and thus 'unwinds' his position. He can only profit if supply and demand
actually creates a fall in prices independent of the action of the short-seller.
5) Why would anyone go 'long' against someone known to be naked-short? Perhaps you can think of the short-long market as
a betters market where the winner takes home BitShares. All market participants are attempting to manipulate the price and
predict which way it will move. 50% think it will go up, 50% think it will go down and the result is a tug-of-war. What
are people really betting on? They are betting on what *other market participants* will do! How do you know what the
other participants will do? You have to assume they are all expecting the price to follow real market prices. Anyone who
is out-of-sync with the emergent consensus opinion about what a price should track will ultimately end up making losses
in this market.
