"But reputation is useless as a hedge against the real nightmare of a setup like Ebay: the long con. It doesn’t cost much, nor does it take much work, to build up sleeper identities on Ebay, fake storefronts that sell unremarkable goods at reasonable prices, earning A+++ GREAT SELLER tickmarks, even for years, until one day, that account lists a bunch of high-value items on the service, pockets the buyers’ funds, and walks off."
That was the business plan of at least half the Bitcoin exchanges.
Of course, in both cases, that is a bug of the particular reputation system, not of reputation systems in general:
1) In the Ebay case, reputation can be conditioned on the type of items and the flow of money being processed by the merchant. Someone who is highly reputable for $50 transactions, isn't necessarily so for $5,000 ones. You could also always have "anomaly detection" built into the system, which kicks off whenever the seller moves too quickly from $50 items to $5,000 and either warns users or pro-rates their reputation (if nothing else, radical behavior change could mean the seller's account got compromised...).
2) In the Bitcoin exchange case, there is currently no real reputation score for Bitcoin exchanges as far as I know. But you could have a proof of stake system that requires exchanges to hold X% of their funds available, similar to the requirements for banks. You could also operate the exchange in a reduced risk way if say, to convert $1000 USD to BTC you sequentially post 1000 $1 USD transactions and don't send them the next dollar until the corresponding BTC amount clears on the network and is safely stored in your hardware or multi-server wallet.
Not saying that there aren't fundamental issues with reputation economies, but lets give the dystopia a fair trial and assume that the non-essential problems will be solved.
In response to (1), plenty of exit scams involve exactly the same product that the company previously sold with 5-star ratings. Either an expensive product from the beginning, or long fulfillment times that allow a large backlog of orders before anyone takes notice, or both.
(No, that $100 ounce never came, and I lost my money just like fifty other fools on Agora.)
EDIT: oh yeah, this is also how I got scammed out of $200 trying to purchase a fake ID. Kids, friends make friends use an escrow service.
> But you could have a proof of stake system that requires exchanges to hold X% of their funds available, similar to the requirements for banks
In Bitcoin's case you can even do that provably and without trusted third parties: https://github.com/olalonde/proof-of-liabilities. I haven't worked on this in a while so there might be more recent proposals.
#2 is pointless. People want to keep bitcoins on an exchange because every bitcoin > bitcoin transaction has overhead.
So, now the exchange has account X with Y money in it. Once Y get's to 100,000,000$ (or whatever) they defect.
PS: The bitcoin network only handles ~7 transactions a second, 1,000 1$ transactions would take over 16 minutes if your willing to pay more money than anyone else to do those transactions. However, if you wait for a transaction to clear that's ~10 minutes per transaction. So, 1000 1$ transactions would take 1 week.
The problem with 2) is that the USD transactions would be too hard to do with the current banking system. While in many cases you can already do instant and free wire transfers, virtually no banks offer APIs good enough to program such scenario.
Also, I'd be afraid of triggering some kind of banking system warnings if I sent a hundred transactions within one day.
With crypto to crypto exchanges a better solution would be to do it as a DAO/smart contract.
Why should an exchange have any less than 100% of the funds available? The exchange should be holding all customer funds in segregation, not mingling and floating them.
It's strange to think about, but realistically the exchange does not need 100% of the funds available, since they will never all be withdrawn at once, even in the event of a massive crash.
The actual proportion an exchange needs to keep to be able to pay everyone (who asks for it) at all times is about as predictable as the bitcoin market itself.
Would we trust "100% funds available" exchanges more? Possibly, but the "cost" would be very high so I would actually be a little wary of an exchange which does this, as to me it would be "trying too hard" to look nice (in reality there's always some trust involved, and so some risk that the owners run off with your coins/money).
> It's strange to think about, but realistically the exchange does not need 100% of the funds available, since they will never all be withdrawn at once.
This was never true, but in the aftermath of 2008, makes a really poor argument. Two points worth mentioning without going into details:
1. Bitcoin doesn't have a central bank to easy the bank-run.
2. When financial businesses go burst, it's extremely difficult to quantify the loss a priori for the accountants
Any sane financial business with half-brain at this time and age would keep at least some percentage of the funds intact.
> Possibly, but the "cost" would be very high so I would actually be a little wary of an exchange which does this
You lost me there. Why would that cost be high? All you need is a wallet (or two, hot/cold) which holds your customer deposits and which you are not allowed to spend from.
With fractional reserve banking, the institution has assets greater than liabilities. With insolvency, the institution has a plan to someday have assets greater than liabilities.
Because then you get all the "features" of the current banking system that bitcoin wanted to avoid.
Bitcoin is by design deflationary currency. Fractional reserve banking is by design inflationary. No matter where you sit on how much inflation and what monetary policy is best - it is obvious that it is against the intentions of the original adopters and creators of the BTC network.
Also partly how pirateat40's big Ponzi scheme worked. He built up a strong reputation on the Bitcoin trading community's web of trust-based feedback system and enlisted a bunch of other highly-rated people from there to resell his "investment" program. The people who accurately pointed out that it had all the markings a ponzi naturally had much lower reputations on there, especially after they'd done so.
What's really funny is EVE online has proven this exact phenomenon dozens of times and is essentially an accurate model of what 'anonymous digital currency' leads to.
That doesn't sound right to me. Scammy exchanges run away with the balances people carry at them, right? That seems very different (and much more lucrative and viable) than a retail storefront changing overnight from reputable practices to theft. Obviously the former is a real problem and a big threat, although it has little to do with reputation systems. The latter doesn't sound like that big of a threat. It seems like the cost to establishing a retail storefront with a solid reputation is not worth the quick payout from theft. In fact, if you have the initial capital and business acumen to run a stable retail storefront, it's probably more profitable and less risky to just keep doing that.
Or a ponzi scheme. The initial investors get paid out which builds up the schemer's reputation, then when the pot is big enough (or new investments slow down) the schemer walks away with the pot.
They only protect you if the seller has money in his account. My Friend bought a video card that was never sent. Ebay said that the seller did not have money in his account so there was nothing they could do. He was out $500 with no recourse.
Exactly this. Had to go through this situation. PayPal said they'll flag the account, but these accounts are already abandoned once the money is withdrawn.
In addition to my other comment. You want to pay by credit card since it has fraud protection. Never use PayPal credits or bank transfers or debit cards. Debit do have some protection, but usually have shorter time frame on when you need to report.
Well, that's kinda fine by me. If I'm out $1000 from a fraudulent lens purchase on eBay, and they/Paypal won't refund, then I'm fine not doing business with them anymore and recovering my $1k from credit charge back.
Perhaps if we had some formal description of all the store fronts we could write a program to calculate the "effective rating" for the item being sold and that could help prevent attacks like these.
That was the business plan of at least half the Bitcoin exchanges.