5 ms·

On BitCoin this type of behaviour seems quite easy to spot but it happens elsewhere too - eg the S&P E-Mini futures and probably the Nasdaq futures. It takes r

by jarym 7y ago

On BitCoin this type of behaviour seems quite easy to spot but it happens elsewhere too - eg the S&P E-Mini futures and probably the Nasdaq futures. It takes remarkably few contracts to actually move the market - especially during the Globex session.

davidgerard 7y ago
oh yeah - this sort of thing is rife on thinly-traded commodity and forex markets, especially if you think the regulators aren't paying much attention - you'll see Barts all over the place. It's a symptom of just how thin trading is in crypto, even for the "biggest" one.
- antonvs 7y ago
  I'm not sure "thin" is the right word here. It took something like 20 minutes for someone to dump nearly $40 million worth of Bitcoin, and after the dust had settled, the price had dropped less than 12%. There are many other factors that lead to this sort of thing happening with Bitcoin, including the lack of regulatory oversight you mention.
  - village-idiot 7y ago
    The nominal market cap of BTC right now is about $129 billion. A $40m trade represents about 0.03% of the reported market cap. That a trade that small would cause the market to crash by double digit percentage points is absolutely a sign of a thin market. Also, the slow speed of price propagation isn’t a sign that the market is thick, it’s a sign that the market is inefficient.
    - logicchains 7y ago
      >That a trade that small would cause the market to crash by double digit percentage points is absolutely a sign of a thin market. It was only the market on a single exchange, Bitstamp, which is extremelly illiquid compared to the largest exchanges. The problem was that Bitmex, a much more liquid exchange, bases the settlement price of its popular perpetual Bitcoin swap on an index of just two exchanges, Bitstamp and one other, so manipulating the price on illiquid Bitstamp also screws with the price on liquid Bitmex.
      - davidgerard 7y ago
        Coinbase's isn't much deeper. The real problem is that the exchanges all have separate order books, and it's hard to arbitrage between them. This is a recipe for volatility.
      - village-idiot 7y ago
        I agree that Bitmex basing its settlement price on two less liquid exchanges is a bad thing. Personally given the number of margin traders being force liquidated, I suspect that BitMEX likes the lack of liquidity. However, if the issue was only a thin market on Bitstamp, the price fluctuations would have remained localized to Bitstamp and the futures market on Bitmex. But instead the price for bitcoin dropped by about $1k across the board, which to me signals thin markets everywhere.
    - antonvs 7y ago
      That's not a relevant percentage. You'd need to look at the amount of Bitcoin that trades regularly. Also, I wasn't saying anything about slow speeds. I was saying that the ability of the market, on a single smaller exchange, to absorb a $40 million trade in 20 minutes is not traditionally what you'd call a thin market.
      - jacques_chester 7y ago
        Suppose for the sake of argument you knew that this wouldn't require blockchain confirmation because it's an exchange trade. And suppose you knew that indexes follow exchanges. And suppose you knew that these exchanges are poorly scrutinised. To the point that, hmm, you could just trade with yourself to pump or dump the price. A market composed of yourself and yourself is, by any measure, pretty thin.
        davidgerard 7y ago
        Now, you might think wash trading is rampant on bitcoin exchanges, but that's because it is. https://medium.com/@bitfinexed/wash-trading-bitcoin-how-bitfinex-benefits-from-fraudulent-trading-8bd66be73215 https://medium.com/@bitfinexed/wash-trading-bitcoin-how-bitf...
  - bigiain 7y ago
    Maybe. Remember it takes "something like 20 minutes" for _any_ bitcoin transaction to get two confirmations. That's kinda "the speed of light limit" for any transaction in bitcoin, whether it's 1 satoshi, or $40mil... (I don't know if this $40Mil "dump" was one transaction or a significant portion of the ~4000 you can fit into one block, if it was in a series of transactions averaging less than $10k, if couldn't have been done in one block, and there's no way it could have been confirmed in 20 mins...)
    - bubblewrap 7y ago
      The author of the original article also cites that as some kind of reason for something. It's pretty irrelevant, though. If you have your Bitcoins stored on a market, your "transactions" don't go through the blockchain. It's just a field in a database then, indicating how many Bitcoins a user has stored on an exchange. The argument would be the same as as claiming "to transfer stocks, you have to deliver them by mail to the recipient, so it takes at least two days". Or indeed the same with money. If you wire somebody some money, no actual money is moved. It is just fields in a database of banks changing, that indicate how much money the banks owe you. We are so used to it by now that we think of it as actual money changing hands, but it is not. So I am sorry, I can only take the article half serious. The author seems to be driven too much by his disdain for Bitcoin than by actual facts.
    - antonvs 7y ago
      Someone else mentioned it, but just to reiterate - these aren't transactions on the blockchain, they're trades on an exchange. They happen essentially instantly, because all that's happening is the exchange is updating its internal customer account balances. The only time this hits the blockchain is when someone withdraws Bitcoin from the exchange.

by jarym 7y ago

davidgerard 7y ago

oh yeah - this sort of thing is rife on thinly-traded commodity and forex markets, especially if you think the regulators aren't paying much attention - you'll see Barts all over the place. It's a symptom of just how thin trading is in crypto, even for the "biggest" one.

antonvs 7y ago
I'm not sure "thin" is the right word here. It took something like 20 minutes for someone to dump nearly $40 million worth of Bitcoin, and after the dust had settled, the price had dropped less than 12%. There are many other factors that lead to this sort of thing happening with Bitcoin, including the lack of regulatory oversight you mention.
- village-idiot 7y ago
  The nominal market cap of BTC right now is about $129 billion. A $40m trade represents about 0.03% of the reported market cap. That a trade that small would cause the market to crash by double digit percentage points is absolutely a sign of a thin market. Also, the slow speed of price propagation isn’t a sign that the market is thick, it’s a sign that the market is inefficient.
  - logicchains 7y ago
    >That a trade that small would cause the market to crash by double digit percentage points is absolutely a sign of a thin market. It was only the market on a single exchange, Bitstamp, which is extremelly illiquid compared to the largest exchanges. The problem was that Bitmex, a much more liquid exchange, bases the settlement price of its popular perpetual Bitcoin swap on an index of just two exchanges, Bitstamp and one other, so manipulating the price on illiquid Bitstamp also screws with the price on liquid Bitmex.
    - davidgerard 7y ago
      Coinbase's isn't much deeper. The real problem is that the exchanges all have separate order books, and it's hard to arbitrage between them. This is a recipe for volatility.
    - village-idiot 7y ago
      I agree that Bitmex basing its settlement price on two less liquid exchanges is a bad thing. Personally given the number of margin traders being force liquidated, I suspect that BitMEX likes the lack of liquidity. However, if the issue was only a thin market on Bitstamp, the price fluctuations would have remained localized to Bitstamp and the futures market on Bitmex. But instead the price for bitcoin dropped by about $1k across the board, which to me signals thin markets everywhere.
  - antonvs 7y ago
    That's not a relevant percentage. You'd need to look at the amount of Bitcoin that trades regularly. Also, I wasn't saying anything about slow speeds. I was saying that the ability of the market, on a single smaller exchange, to absorb a $40 million trade in 20 minutes is not traditionally what you'd call a thin market.
    - jacques_chester 7y ago
      Suppose for the sake of argument you knew that this wouldn't require blockchain confirmation because it's an exchange trade. And suppose you knew that indexes follow exchanges. And suppose you knew that these exchanges are poorly scrutinised. To the point that, hmm, you could just trade with yourself to pump or dump the price. A market composed of yourself and yourself is, by any measure, pretty thin.
      - davidgerard 7y ago
        Now, you might think wash trading is rampant on bitcoin exchanges, but that's because it is. https://medium.com/@bitfinexed/wash-trading-bitcoin-how-bitfinex-benefits-from-fraudulent-trading-8bd66be73215 https://medium.com/@bitfinexed/wash-trading-bitcoin-how-bitf...
- bigiain 7y ago
  Maybe. Remember it takes "something like 20 minutes" for _any_ bitcoin transaction to get two confirmations. That's kinda "the speed of light limit" for any transaction in bitcoin, whether it's 1 satoshi, or $40mil... (I don't know if this $40Mil "dump" was one transaction or a significant portion of the ~4000 you can fit into one block, if it was in a series of transactions averaging less than $10k, if couldn't have been done in one block, and there's no way it could have been confirmed in 20 mins...)
  - bubblewrap 7y ago
    The author of the original article also cites that as some kind of reason for something. It's pretty irrelevant, though. If you have your Bitcoins stored on a market, your "transactions" don't go through the blockchain. It's just a field in a database then, indicating how many Bitcoins a user has stored on an exchange. The argument would be the same as as claiming "to transfer stocks, you have to deliver them by mail to the recipient, so it takes at least two days". Or indeed the same with money. If you wire somebody some money, no actual money is moved. It is just fields in a database of banks changing, that indicate how much money the banks owe you. We are so used to it by now that we think of it as actual money changing hands, but it is not. So I am sorry, I can only take the article half serious. The author seems to be driven too much by his disdain for Bitcoin than by actual facts.
  - antonvs 7y ago
    Someone else mentioned it, but just to reiterate - these aren't transactions on the blockchain, they're trades on an exchange. They happen essentially instantly, because all that's happening is the exchange is updating its internal customer account balances. The only time this hits the blockchain is when someone withdraws Bitcoin from the exchange.