I believe MtGox users have unfortunately learned the definition of "counterparty risk". This is one of the major overlooked risks that almost took down the global financial system in 2008. People would hedge their worthless bonds using AIG as a counterparty, and once things blew up, these people realized their insurance policies were completely worthless because AIG was about to go under itself, causing them to almost go bankrupt as a result.

Unfortunately, in a world where none of the bitcoin exchanges have any sort of regulation, there's no way that BTC use can expand via BTC financial institutions until people have full confidence they can get their money back in the case of another collapse like mtgox.

The collapses we've seen with traditional currencies stem from the inherent risks of fractional-reserve lending.

Is this a risk with Bitcoin? According to my limited understanding, it doesn't seem to be. That should in theory reduce the need for things like FDIC and complicated risk calculations, like counter-party risk.

MtGox doesn't seem like a "bank failure" or "bank run" to me. It seems like a bank that forgot to lock the doors at night.

At least if your money is held in an FDIC insured bank and the bank goes under, you can rely on the government to insure your investment. With bitcoin you have no such thing.

That sounds very smart, however if your money is held in an FDIC insured bank, then it's likely to be in a national currency that is systematically devalued by the issuing institution through progressive year-on-year inflation. (Note that I use the term institution instead of government as two major issuers, the Federal Reserve and the Bank of England, are private organizations).

PS. I'm not characterizing inflation as bad, nor am I a goldbug. I am merely pointing out that the entire population of FDIC-insured banks are paying constantly for the mystical "it might happen some day" promise of a bailout, due to an inflationary system that reserves priveleged access to a central clique - ie. it's not like the conservative viewpoint of "faith in <the national central bank>!" comes without strings. (I am most aware of this perspective as someone who used to live with someone on the interest rate committee for the Bank of England.)

> That sounds very smart, however if your money is held in an FDIC insured bank, then it's likely to be in a national currency that is systematically devalued by the issuing institution through progressive year-on-year inflation.

Which may be a good reason to limit currency holdings to what you need to use as currency (that is, short-term medium of exchange for things of value), and invest the rest in productive assets, whether those are stocks, bonds, real estate, or digital collectibles like internet funbucks. But if you aren't directly holding those productive assets yourself but are trusting them to account held by someone else who promises to transfer them to you on demand or dispose of them at your direction, you probably want to do some consideration of their ability to keep them securely and execute the trust you are placing in them.

Inflation isn't bad and the only people that think it is are goldbugs with no understanding of economics. Asset backed currency have had MANY more crashes than fiat currencies. Bitcoin is a wonderful experiment in how long it will take people to learn the lessons the world learned over the previous 200~ years.

I'll bet on the long term stability of the US government over that of MtGox or Coinbase anytime.

Just to add to some of the other great responses here:

The British empire was built on the back of the Pound -- the first truly international, modern fiat currency. It reigned until it was muscled out by the fiat Dollar.

An immense amount of wealth has been generated for humankind on the backs of fiat currency.

And one more thing...

Bitcoin is also a fiat currency in the sense that it does not contain intrinsic value. It is only as valuable as its network effect. If the network falls in love with a different coin, bitcoin dies.

Fortunately there's ways of proving available deposits when using bitcoin. In theory at least, this can address the issue of counterparty risk. Hopefully all the exchanges will start adopting this.

Not really. If the exchange says "We have 500,000 BTC" and they publish their wallet, but their actual customer liabilities are 2M BTC, everyone can still lose a lot of money. What they really need is to be externally audited and undergo consistent regulation.

I'm talking about a more sophisticated solution that addresses this- See https://github.com/olalonde/blind-solvency-proof

Disclaimer: I only skimmed this solution.

However, it looks like you don't/can't address all of the counterpart's liabilities. So salaries, tax dues, rents, loans, and other obligations denominated in fiat currency don't appear in the calculations. But any of these can destroy a counterparty as surely as insufficient reserves. Crucially, depositors don't really care where the risk is, they care that there is some external entity (like the SIPC or FDIC) that can insure that risk.

You are right, those other factors cannot be accounted for with this approach, it only proves that the business is not running with fractional reserves.

You can only lose what you give them, which should be no more than is necessary to complete a set of near-term transactions.

I've been wondering if the systemic portion of this risk mightn't be helped by allowing judges the power, during bankruptcy proceedings, to magic a small percentage of the balance owed into existence for the lender if the lender also forgives some (larger portion) of the balance provided that the loan is determined to have been a real loan made in good faith. This could help inject liquidity precisely when and where it's needed, while the requirement of someone undergoing bankruptcy, forgiveness of some of the debt, and judicial discretion would help check abuses.

The main feature of BTC is that it's unregulated, so indeed, if you don't provide the security usually enforced by governments, there isn't any.

Usually, when using online services such as MtGox, you're secured by the fact that their reputation is worth more than what they can scam out of you. It's vital for eBay, AirBnB, the Silk Road etc. that you trust them, so they won't let an horror story unfold publicly just to steal you a couple of grants.

If a service / person is in a position to pull off a scam worth more than the reputation / survival of the service, it's dangerous to trust their good will. MtGox was predictably dangerous. Given the poorly aligned interests of MtGox, and the target they represented for black-hat hackers, all this is a (very expensive) lecture in basic economics for libertarian masses.

> The main feature of BTC is that it's unregulated

The main feature of BTC is that, independently of regulation by any other entity that may be applied to it, it is "regulated" by the design of the protocol, which, among other things, provides decentralized issuance and transaction verification, and predefined supply constraints.

A popular myth about BTC is that it is unregulated (and, in some forms of the myth, impossible to regulate) by government.

Which is as close to "unregulated" as useful definitions go: once unleashed and bootstrapped, no human nor political organization can directly drive it. It's only self-regulated the way a free market is self-regulated, which is almost the exact opposite of what one has in mind when they talk about "regulations".

(Actually governments can influence its value significantly, by stating on its legality, and the legality/taxation of its conversion into state moneys / commodities).

What boggles the mind is Mt Gox started as a site for trading Magic The Gathering cards. I suppose in the code btc was just a special type MTG card? Incredible it got this far.

There was never a MTG trading card site on that domain. This is a myth you should stop spreading.

Indeed, it seems relatively apparent that the name was originally related to MTG, and the original person had plans to create a MTG website, but I have never seen any evidence to suggest that a MTG trading website was actually created, let alone then re-purposed to be a bitcoin exchange.

There are plenty of things to find stunning or hilarious about mtgox without making things up.

Well if that's so untrue then you should take your mythbusting campaign to the Mt Gox page on WikiPedia where it clearly states this was the original intent of their business. Mt Gox stands for "Magic The Gathering Online eXchange", does it not?

Stripe.com once hosted a gallery of Furby pictures. Does that mean we should blame any bugs in their payment platform on misbehaving toys? No, previous hypothetical uses of the domain (by a different owner no less; Jed McCaleb who registered the domain and Mark Karpeles who owns it today are different people) are irrelevant, and mtgox.com does not stand for that.

It was almost certainly the intent on purchasing the domain, but that doesn't establish that one was actually running. My understanding of what evidence is available leads me to think it never was (particularly a lack of anyone posting anything about the domain in conjunction with Magic anywhere except comments like the above).

>(1) Split a $100 transaction into 10 $10* transactions executed incrementally (send $10 of BTC to Mt. Gox, exchange it, transfer it out, and only then send the next $10 of BTC over). This keeps exposure at any given time at a manageable level. The downside is it increases exposure to BTC/USD volatility.

This is an old anti-scamming solution and is very useful in some settings but this isn't really one of them. The main problem here isn't really losing your money immediately after making the transaction. Just making the transaction in one swoop and immediately withdrawing your money after is much more efficent (in terms of both time spent and fees).

Most people weren't screwed because they made one large transaction just before things went wrong - they were screwed because they kept money for long periods of time on gox. And yes, you can argue that some people just couldn't withdraw for months - but if they wanted to use gox and use your scheme it would've taken a year to execute all of those transactions.

Assume, a deposit-trade-withdraw transaction takes 1 hour, and the chance of an exchange failing in a given hour is 0.01%.

If you perform a single $100,000 transaction, there is a 0.01% chance you will lose $100,000. So on average your loss is $10 but in the worst case your loss is $100,000.

If you perform 100 transactions of $1,000 (taking 100 hours) there's (roughly) a 1% chance you'll lose $1,000 so your average loss is still $10 but in the worst case your loss is only $1,000.

I agree that performing 100 transactions over 4+ days sounds like a right hassle, and prices might change by the time your trade finishes - whether this precaution is worthwhile depends on your taste for risk and whether you're trading $100 or your entire life savings in bitcoins :)

Personally if I had a million dollars of bitcoins, the last person I'd trust with it would be a bitcoin exchange :)

You are forgetting the extra fees.

(Also the 'a deposit-trade-withdraw transaction takes 1 hour' assumption is nowhere near accurate, especially for Mtgox.)

It's a pity your post isn't at the top of the thread. It is sound advice.

I find it fascinating that there are (apparently) people in this thread who lost $500k at MtGox and despite the loss would still have no problem paying $250k in taxes on it if they were required to do so.

I would hope if you're making that kind of money you'd either have a sound sense of financial risk management, or the wisdom to hire somebody who does to manage it on your behalf.

Of course, it's easy to criticise with the benefit of hindsight.

My comment below highlights the clawback risk for those that were successful at pulling money out. If this turns out to be a Ponzi scheme there is a real chance that people who profited and were able to get their USD out may actually be forced to forfeit all of their profits depending on the jurisdiction.

I'm not sure I've heard any claims that would point towards MtGox being a Ponzi scheme. It was a straight forward exchange, not an investment with returns of any sort, let alone those derived from later investors.

The allegations are generally much more straight forward, either some sort of incompetence in the form of transaction malleability, or some sort of malfeasance (pilfering the bank for operating costs, outright theft of the funds). Ponzi schemes these are not.

I find it fascinating that people take at face value the words of a throwaway account who is quite likely making the whole thing up, given that you'd want wealth of 100m to just shrug off 250k like that. (Based on the fact that someone worth 100k can just about shrug off 250 like its nothing).

They claim to be a miner. If I made (say) 1m, then lost 250k of it (due to risks which I was aware of when I started playing the game), I'd still be pretty happy.

I think they are just chalking it up to a cost of doing business.

If they lost 250k which they'd made from a different source, they'd be more upset. As no doubt may people are right now.

I'm not sure if it's rational. But ... that's the way people often respond.

Given that most miners have been operating at a loss (after power, space, hardware, etc) for quite some time, I'd also take that statement with a pinch of salt.

Dude, I'm poor has hell and I'm not crying about the €30k I lost there. And those 30k are several multiples of what I got left with. Not everybody commits suicide for losing money, you know...

Mt. Gox and many other exchanges have put into people's minds that you HAVE to forfeit your rights to your Bitcoin when using online services. It doesn't have to be that way, though. You could just use a system like we use at Bitalo [1], where fractional reserves or funds misuse are impossible because of use of multi-signature Bitcoin addresses, which means funds are specifically tied to user wallets and exchange operators cannot use them, because the user has to sign all transactions by himself.

[1] https://www.bitalo.com/why_bitalo

This isn't realistic for an exchange. They have an enormous transaction volume, which is represented as internal/off-chain transaction in an internal accounting system.

Using multi-signature transactions requires every fund transfer to be represented as a Bitcoin transaction, which the Bitcoin neAtwork simply can't handle.

Since we are a P2P exchange, we don't foresee to have more than few dozens transactions per minute when we grow big. I'm sure the Bitcoin network can handle such volume.

The protocol can only handle 7 transactions per second.

The system is currently configured with a limit of 7 TPS. But the limit can be lifted, according to https://en.bitcoin.it/wiki/Scalability

Why would bitalo ever require users ID themselves or provide bank details? multisignature scheme already adds a latency on transaction. just not clear how this exchange works. Is it P2P where users transact between each others bank accounts. If so the AMA/KYC seems unnecessary.

The reason is two-fold. One is legal/regulation stuff, as jccooper said (for seller account especially). Second is seller protection - if a buyer wants to scam sellers, we can blacklist him based on his ID, so he cannot scam anyone again.

Legal reasons, almost certainly. I don't know about Finland or Germany or EU specifically, but it seems likely they have "money transmitter" requirements to combat money laundering. I doubt the regulators and/or law know or care much about multi-sig transactions. If they touch money changing hands, they almost certainly have to comply.

You can't move money without the government's permission, you know. Very naughty.

Hello! fellow diæresis user:)

The only problem with doing multiple transactions is the amount of time required to both withdraw and send. If I waited, the price could fall (or rise...)

What's needed is a non-governmental/industry-driven method of showing solvency and competence. Blockchain-transfers and signed messages to show bitcoins solvency maybe? I'm not sure how to show solvency in a national currency. I'm also not sure how to show full compliance with the protocol without an auditing organization.

Also, users need to pay attention. As soon as mtgox started making bad transactions people should have started questioning them more. When they didn't properly fix those transactions, people should have started moving away.

>The only problem with doing multiple transactions is the amount of time required to both withdraw and send. If I waited, the price could fall

There are two risks: the risk of the (1) price moving against you (delta risk) and (2) exchange defaulting (CP risk). The longer your funds spend waiting to transact, the greater your delta risk. The longer your funds spend at the exchange, the higher your CP risk.

Bitcoin is a volatile asset, making delta risk expensive. But the exchanges are opaque and operate in a legal grey area, amplifying CP risk. The risks are also correlated - if BTC/USD falls, the odds of the exchange collapsing increase. This makes the problem complicated. What I can say is this: the solution is far from either end. You should never have a significant fraction of your liquid net worth in any of these exchanges.

The most innocent excuse for taking weeks to process withdrawals is operational inefficiency. Even this increases your CP risk. Do not react to long wait times by increasing your transaction size. Instead, aggressively seek out alternatives.

> The longer your funds spend at the exchange, the higher your CP risk.

But we're talking about doing 10-$10 transactions versus 1 $100 transaction. The difference in time that money is at risk for the exchange defaulting is negligible in this scenario.

> You should never have a significant fraction of your liquid net worth in any of these exchanges.

Agreed. The other side of that coin, I would say, is that a "significant fraction of your liquid net worth" shouldn't be used in a single transaction either.

> Do not react to long wait times by increasing your transaction size.

I agree, you shouldn't make a transaction larger than it should be or force a transaction to be sooner than you'd like because of the wait time, but specifically breaking up one transaction seems a bit much.

> Instead, aggressively seek out alternatives.

Agreed. I did a long time ago. Their invalid transactions should have been a huge red flag.

How would you transfer out $10 from an exchange? My bank charges more than that in fees for any transfer at all... Also, it always took weeks to get anything out, so you'd be waiting for months to transfer $100.

You wouldn't transfer $10; that's just an example. You would split your transaction in such a way as to reflect the maximum permissible exposure you are comfortable with (the largest amount you're willing to risk losing - not forgetting the other transactions you may have on the table elsewhere). If this value is $100 you probably have no business dealing with Bitcoin anyway.

If, on the other hand you're the guy trying to cash out $500k worth of BTC, a $25k loss is probably acceptable risk so you might exchange your $500k over 20 independent transactions of $25k each. In reality, you're not going to be able to sell $500k of BTC in a day anyway, so it's difficult to see this process as onerous -- it's just part of the game.

> In the case of Mt. Gox, or any other Bitcoin exchange, that credit is secured by nothing and insured by noöne.

I believe that's not quite true. IIRC there is a european bitcoin exchange, where at least the euro balance is insured up to a certain sum[1].

[1] http://spectrum.ieee.org/tech-talk/telecom/internet/bitcoinc...

It's kind of sad to me that this even needs to be written out, it should be obvious: If you can't trust somebody, you have to break any financial transaction with that person into little pieces.

To be fair, I'm guessing most people burned by MtGox in a big way were day trading, where this advice won't work.

This is part of the reason I never day traded bitcoin- The counterparty risk is (and clearly was) too great.

Hm. Are there any wallet-exchange metaclients that would do that? Salt your $100 evenly between a few exchanges of your choice?

True only for fixed-exchange-rate currencies.