Bitcoin August 14 Flash Crash | Hacker News Reader

Bitcoin August 14 Flash Crash

coinprices.io

106 pointsmurtali63 comments

I am the author of this post. Let me know if you have any questions.

Essentially what happened is that Bitcoin Exchange Bitfinex allows you to trade with leverage. There were approximately $28M USD worth of leveraged orders and they were margin called as the price fell. This created a feedback loop that resulted in the flash crash.

So just to make sure I understand, when BitFinex offers leverage trading, are they betting that the price of Bitcoin won't go up as high as the folks asking for leverage think it'll go?

If not, why do people offer leveraged trading? Just for the interest? A fee?

> why do people offer leveraged trading

Bitfenix doesn't provide the leverage, they use other trades to provide it.

Those traders get hourly interest for participating.

If you have a balance and you don't want to use it, loaning it out can make some money you wouldn't otherwise get.

It is similar to loans but on a shorter window.

A broker offering leverage is earning interest nearly risk free. They retain control of the collateral and can (as observed) force close the position. Only in really extreme situations is their money at risk.

Correct, but it is actually other users offering the leverage. Bitfinex created a P2P market for it. If you offer your USD as leverage then you receive interest.

You said there was $28M in leverage before the crash. Do you know what the leverage is currently?

It fell from $29M to $24M: http://www.bfxdata.com/swaphistory/totals.php

But that drop doesn't happen at the same time as the big drop on Bitfinex, as far as I can see. I assume all times are UTC, unless otherwise stated.

The drop in total sum of USD swaps fell from 2014-08-14T07:30 to 2014-08-14T08:30, but the drop on Bitfinex happened around 2014-08-14T10:00

The thing is that Bitfinex only reports their leveraged volume every hour so you can't really pinpoint the margin calls.

Especially since bfxdata is a third party site.

If you go to BFXdata.com you can see historical graphs for it. Bitfinex reports their leveraged volume every hour.

Here's a screenshot of what you are looking for: https://i.imgur.com/SvSFYMK.png

edit1: I also added that graph to the post.

Just a heads up, I was trying to figure out what graphing/charting library you're using and noticed that your <html> tag closes immediately after opening it, rendering the rest of your mark-up invalid.

All that aside, Good article, interesting read.

A stop hunt?

Reminds me of the good old days of Bitcoinica: http://i.imgur.com/TTMjGvB.png

50% swings in minutes.

I know there were suspicions that Bitcoinica did that intentionally, and were making money by forcibly liquidating people's positions through fake price swings.

Was there any evidence?

Anecdotally, in the Bitcoin world, if someone lost money and there's a chance to allege fraud, someone will do so.

That's a good thing. Bitcoin fraud is the easiest large-scale fraud you could hope to deploy in the modern day.

What do other US citizens use to trade BTC with stop orders? I've been wanting setup an account to take advantage of short crashes like this but I'm not sure what exchanges are open to US citizens. And don't say BTCe, it must have transparency.

Be careful, stop orders only help if there is a reasonable amount of liquidity when you reach your stop price. If it's not there (and often isn't during a flash crash) you may get terrible execution.

If you are in the US the easiest way to get money into an actual exchange seems to be by buying BTC via coinbase (they support ACH transactions for US citezens) and then transferring it to an exchange such as Coinsetter, Bitfinex, or Bitstamp to trade. Coinsetter is based in the US and will rebate you your Coinbase transaction fees.

That wouldn't be a stop, it would be a buy/limit order with a price far below (20-30%) the current market price.

Sorry, you are correct. I misspoke. Buy limits are what I want for purchases but really I want all standard order types.

Bitfinex does shorts, which is similar I guess.

Margin calls were one aspect that magnified the magnitude of the 1929 stock market crash.[1] After that, margin regulations (Regulation T) made the minimum margin rate 50% for stocks, though it appears some brokers may have been increasing margin requirements to near this level before the 1929 crash.[2]

1. http://stocks.fundamentalfinance.com/stock-market-crash-of-1...

2. https://en.wikipedia.org/wiki/Margin_call

It's not just margin calls but leverage in general that turns downcycles into crashes, if you believe Kindleberger (and you should). [1]

1. http://www.amazon.com/Manias-Panics-Crashes-Financial-Invest...

If the US dollar fluctuated like this, it would cause huge problems to world economy. But if it happens to bitcoin, it's the most innovative thing to have happened to the world economy ever!

If people used Bitcoin the same way they used the USD, it would be much more stable as a currency.

Right now market liquidity on Bitstamp (I think that's the biggest exchange?) is 3 million per day. If you want to trade and buy $3m of Bitcoins, you're going to have a huge impact on the price. This is not an unreasonable transaction for larger parties and investment firms.

I don't know what the liquidity on the USD is, but consume an entire days worth of trade volume between USD and say, the Euro, you pretty much have to be a small country.

A lot of these fluctuations will smooth out greatly as Bitcoin's actual use and liquidy goes up. If investment firms and hedge funds start giving serious attention to Bitcoin, you'll see it approach a stability closer to that of gold. Still not the USD standard of stability, but a lot better than what it is today.

You realize Bitcoin and the USD are quite different things, right? One has its price regulated artificially by creating more of it, while the other depends solely on the whims of the free market. And besides, Bitcoin is much smaller, and therefore it fluctuates the same way startups' stock fluctuates.

So one could say this: If a startup's stock flucuates like this, it's the most normal thing in the world. But if it happens to Bitcoin, it's something we should criticize!

There are extremists that believe Bitcoin is poised to be the new world's reserve currency (which to me means a highly sought after commodity to compare against), as gold once was in the past 200 years to some degree. This party of thought see it as digital gold that is fungible through the internet. It doesn't have enough critical mass of actual economic value hedged on it yet to be that, if ever, so it's going to be a wild ride when any big players decide to put some monies in the slot machine of the market. Small level purchasers aren't moving the market, and if they're in the US they're probably using Coinbase or some other "Bitcoin bank" that does off-chain transactions. The begrudgoned trusted entities that the current world requires to exist for people to get involved aren't moving the price for the most part, and their prices reflect what the other exchanges are saying. Regardless of the potential for transaction exclusion collusion, any large amounts of Bitcoin bought off the chain can cause wild swings on the markets that are averaged for the prices, and for a while if Bitcoin continues to excel, it will remain just as exploitable as any other financial system such stocks w/ press.

I believe that Bitcoin (or perhaps another cryptocurrency based on similar concepts) absolutely has the technical capability of and some desirable traits for being an important reserve currency. I don't predict with any confidence that this will happen, but I see no reason why it couldn't happen. Not sure if that makes you label me an extremist.

Have I fundamentally misunderstood bitcoins? I thought it was a crypto-currency, not a crypto-stock. If it's supposed to be a currency, then judging it in relation to how other currencies behave isn't unusual at all. If it's supposed to be a stock, why does everybody call it a currency?

If you don't see the difference between a currency that can regulate its price artificially by printing more of it and Bitcoin, then you have definitely misunderstood it.

If you don't feel comfortable comparing it with stock, then compare it with gold. Even after thousands of years, and a market capitalization of 8 trillions of dollars, it still has price fluctuations.

Bitcoin has only ~7 billions of dollars of market capitalization. It's comparable to what we call a "penny stock". Anyone with a few millions can move the price. And even with hundreds of thousands of dollars, you could move the price in one particular exchange.

If you still don't understand it, watch this video:

https://www.youtube.com/watch?v=qHUPPYzzZrI

It's more of a crypto transaction system, which has obvious potential uses as a payment system or a stock. There is no objective "supposed to be" for the technology, but obviously the primary focus of the Bitcoin community is as a payment system.

Bitcoin is brand new and it has to have a growth process because of its nature, unlike state-sponsored currencies like the Euro that can just be "rolled out." So you can't fairly compare it to mature currencies, even though it is a currency.

> One has its price regulated artificially by creating more of it

Not on these scales. Artificial adjustments to the dollar happen on month increments, possibly days.

> while the other depends solely on the whims of the free market

Free is relative, the Bitcoin market is still too small to be isolated from manipulation.

> same way startups' stock fluctuates

Care to quote a startup stock that is traded decently (Bitcoin isn't a penny stock) and has a 10% dip in minutes that isn't newsworthy for anyone interested in that stock?

Because I can bet anyone following it would find it newsworthy.

> > One has its price regulated artificially by creating more of it

> Not on these scales. Artificial adjustments to the dollar happen on month increments, possibly days.

To give you a sense of perspective, the FED was printing $55bn/month this year [0]. So the FED prints the whole Bitcoin economy ($6.7bn) every 3.6 days, ... day after day, month after month ... for many many years now [1].

[0] http://www.forbes.com/sites/samanthasharf/2014/03/19/fed-cut...

[1] http://static4.businessinsider.com/image/52e4250869bedd404cb...

Milton Friedman is famous for saying the inflation of the dollar is exclusively controlled by the FED, but that isn't 100% true today. There are other branches of government that capture back large amounts of the printed paper from operations that are illegal. So in a sense, the "war on drugs" could be keeping inflation lower than it would be without it. For inflation to be felt the money has to be in circulation. Also, when the US broke away from the gold standard and printed a lot of money for police actions, it took about 10 years for the inflation to become rampant. It is important to realize this was before the crack wave.

Any references on the volume of paper that is confiscated and how it compares to $55bn/month? In any case, isn't drug money re-distributed into the economy? (I don't know, so I'm genuinely curious)

paper confiscation isn't what he's talking about when he says capturing back dollars.

> Care to quote a startup stock that is traded decently (Bitcoin isn't a penny stock) and has a 10% dip in minutes that isn't newsworthy for anyone interested in that stock?

"traded decently", gotta love those vague expressions.

Any company that at some point was just a small startup, sold stock privately, and today is big, has had these swings. Eg: Facebook, Amazon, etc.

If you don't know this, it probably means you have never bought private stock.

This swing is newsworthy, but EduardoBautista said it would cause huge problems to the world economy, which is not the same thing.

If the value of the USD fell by 10 percent in fifteen minutes, I can see that causing some problems in the world economy; do you disagree?

This was on a single exchange. Bitcoin didn't really lose 10% of its value momentarily. If the same thing somehow happened at a currency exchange with such crummy controls as this one, a lot of people would lose a lot of money, but not 10% of the value of all USD.

"regulated artificially"

"danielweber"

"This is actually good for Bitcoin because..."

Flash crashes happen on regular stock markets too. People still use them. The flash crashes don't cause huge problems. It's almost like people can deal with brief anomalies somehow...

If bitcoin was used as widely as the USD this would never have happened. It happened because of weak depth in the bitcoin markets (i.e. not that many people are trading it).

Bitcoin is obviously useless and will surely never be of any use to anybody ever.

If grandma had balls, she'd be grandpa. So what?

I'd think you'd be insane to do anything involving debt or margin with bitcoins.

Can't agree with that. Betting say $1 on a highly improbable event can't really be considered "insane" by any metric I can think of - if you lose, so what, you're only out one dollar.

The risk of the event itself doesn't matter; what matters for your personal risk is how much of your personal fortune you put into it.

If losing a bet makes you homeless, it's hardly a good bet no matter how good the odds are.

Conversely, there's barely an "insane" bet in the world as long as the probability for a payoff is greater than zero (excluding e.g. Nigerian scam emails), and as long as the amounts involved are small relative to your disposable capital and the payoff expectation.

Think of it as "how many times do I need to place this bet before I win", vs "if I win, what's the payoff", vs "if I'm wrong, what's the most I could lose".

If you get those numbers right, I can't see that bitcoin margin bets are intrinsically insane.

I'd think you'd be insane to miss an investment with such a disproportionate potential returns / risk.

Lots of 'insane' people in the world. Makes it interesting.

Has there been fluctuations like this due to algorithmic trading of BTC yet?

This is called a liquidity trap and it's well-studied in finance circles. Basically, in order to have a well-functioning market, you need a roughly balanced number of sellers and buyers at any point.

If an external event triggers enough parties' simultaneous need to sell, it can suck all the buy orders out of the market, causing the price to fall lower and lower as the sellers have to submit ever-lower prices to find willing buyers. The event creates a positive feedback loop as sellers go lower and lower to find buyers, leading to sharp, discontinuous movements in price.

The ultimate answer to this is tons of market depth/liquidity, but absent that, exchanges have "circuit-breaker" policies in place that cause trading to halt if prices move too much, too quickly.

Ultimately, ensuring an orderly market is a massive challenge that shouldn't be taken lightly.

EDIT: Child is correct, I think liquidity traps are from macro, but same idea - not enough buyers, too many sellers.

This is not called a liquidity trap. That's a concept in macro econ, it has nothing to do with trading or (this kind of) finance.

It's usually called a liquidity spiral. Brunnermeier and Pedersen's 2009 paper [0] is the standard reference. For a quick overview, see their slides [1] or Pedersen's summary for Vox EU [2].

As Pedersen emphasizes in the Vox article, the amount and price of funding available to participants in financial markets is intimately related to macro factors. So market liquidity, through the funding linkage, can dive along with the real economy.

But liquidity can also evaporate due to endogenous factors, particularly when markets are set up so that the funding available to traders depends on the price of the asset being traded, e.g., through margin rules. That's what seems to have happened on BitFinex today. Positive feedback loops make for ill behaved markets.

0. http://pages.stern.nyu.edu/~lpederse/papers/Mkt_Fun_Liquidit... [PDF]

1. https://www.newyorkfed.org/registration/research/risk/peders... [PDF]

2. http://www.voxeu.org/article/understanding-liquidity-risk-an...

Is the grandparent comment otherwise correct, besides the name?

Well, the original article makes a claim that this was caused by a margin call feedback loop (thanks pash for finding the right term and some analysis).

There could be other causes. It's hard to untangle whether BitFinex caused the other exchanges to drop, or if it just dropped faster due to margin but had no real effect on the other exchanges.

Yes, Bitcoin price fluctuates.

Why is this on the frontpage?

I upvoted because the article provided an interesting technical explanation of the fluctuation.

The technical explanations were (1) large sell order causes price depression and (2) ensuing margin calls.

Those aren't interesting. The first is expected behavior in any low-liquidity market. The second is expected behavior if foolish investors are involved in low-liquidity markets. In other words, this was expected behavior in the bitcoin markets.

Both factors are also seen regularly in traditional stock markets.

A 10% swing in minutes is out of character even for Bitcoin.

Not really. From this and your other comments it's now clear that not only you don't know anything about private stock, you don't anything about Bitcoin's history either.

Reply on news.ycombinator.com