4 ms·

That was a nice read. I like these pieces about the historical context of financial instruments. > Michael Lewis details one such practice in his book Flash Bo

by dsacco 9y ago

That was a nice read. I like these pieces about the historical context of financial instruments. > Michael Lewis details one such practice in his book Flash Boys, where market makers use superior networking technology to change the market price of a security after a customer sees it and places an order, forcing them to trade at an inferior price. This is a potentially misleading description of HFT algorithms. HFTs cannot see a customer’s order before it executes on an exchange. They cannot “jump in front of” an atomic order. HFTs only have insight like this at the inter-exchange level, i.e. in latency arbitrage. But by default a customer’s order won’t be routed to multiple exchanges unless there is insufficient liquidity to complete the order on the first exchange (which, if that’s routine, means you’re likely an institutional customer). Alternatively, if they submit several orders, HFTs will (logically) adjust their prices in response to the new order book activity, which would result in price slippage for the customer.

cryptonector 9y ago
HFTs == market makers. You need market makers in order to have a functioning, liquid market. Absence of market makers -> higher volatility. Market makers can lose a lot of money under certain market conditions. Their profit margins are rather small. Market makers are practically heros.
- vinceguidry 9y ago
  It's a bit naive to equate the two. Plenty of strategies can be found that leverages high transaction speed and volume that aren't market making. What defines market making is the fact that you buy and sell the same commodity at the same time, taking advantage of the spread to make money. Two other rapid transaction arbitrage strategies involve market manipulation, in which trades are placed that are designed to fool other trading algorithms into believing that the price is going to move a certain way, and statistical arbitrage, which searches for temporary price differentials across exchanges. Neither of these two strategies create liquidity on the market, that is, offers more opportunities for anyone who wants to buy or sell to do so. Instead they seek to create, and/or exploit opportunities.
  - cryptonector 9y ago
    ISTM that high frequency offers/bids that are not meant to complete will tend to become noise that the market filters. That's my sense, though admittedly I don't know enough and have no data to back that up. And when such offers/bids do complete, they really are market making. Either way, this is helping the market discover the price -- you call label this with "exploit", but it's not clear that that's anything but bias. Normal market making too could be seen as exploitation. Suppose the market is crashing. Without market makers the market ceases to function and uncertainty worsens. With market makers the market functions and crashes as much as it wants to. So your stock holdings lose value and here's a market maker glad to help that happen. How is that not "exploitation"? Politics is a war of words, I get that. I guess I should be happy that nowadays we don't punish market makers.
  - dsacco 9y ago
    > What defines market making is the fact that you buy and sell the same commodity at the same time, taking advantage of the spread to make money. You need not be a formal market maker to increase liquidity. Technically speaking even retail customers will receive small transaction fees for increasing liquidity in certain circumstances where it's particularly lacking. More generally speaking, any strategy that rapidly connects two parties is market making, even if the two parties are not immediately next two each other in the strategy's trading queue. > Two other rapid transaction arbitrage strategies involve market manipulation, in which trades are placed that are designed to fool other trading algorithms into believing that the price is going to move a certain way Market manipulation is not a good term to use for this, and it's a fairly serious accusation to levy against a trader. Market manipulation is an illegal activity that typically involves cornering or collusion. More abstractly, this is a net improvement in price discovery, because one party notices that another party can be fooled (which means their algorithm has a glaring pricing inefficiency). In repeatedly taking money from the other party, the other party becomes incentivized to stop executing that strategy. Once there is no counterparty to take that trade anymore, the inefficiency has been traded away, and the market is now (likely imperceptibly) more efficient. > Neither of these two strategies create liquidity on the market, that is, offers more opportunities for anyone who wants to buy or sell to do so. Instead they seek to create, and/or exploit opportunities. This creates liquidity. Any situation that increases trading activity increases liquidity by definition. You do not need purely symmetric market making to improve overall liquidity - if many low latency trades are being executed, and they are finding counterparties, the liquidity of that market is increasing as a result of the volume. HFTs do not buy and hold, which means they are rapidly connecting buyers and sellers, even if it individual securities are bought and sold asynchronously.
- crdoconnor 9y ago
  The sheer number of market making trades provides ample cover to hide some of the less legit trades if/when regulators come a-knocking. As you said, market making is super low profit. HFT certainly used to be a fairly rich source of profit. Not sure if what it's like these days. This is all probably just a coincidence.
  - cryptonector 9y ago
    Regulators should audit regardless. It's not like we should want lower volumes just to make it easier on auditors. We can't exactly go back to 1920s tickers so as to have lower volumes...
    - crdoconnor 9y ago
      Regardless of whether they audit or not, it's germane to recognize that something is off when a supposedly low profit activity is actually generating quite a lot of profit.
      - cryptonector 9y ago
        Low profit-margin x high volumes == potentially big profit. Nothing's wrong with that.
        crdoconnor 9y ago
        Which would work if there were one or two players trading an incredibly high volume, but that neglects to consider a little thing called competition which annihilated those margins. Now that doesn't matter so much if you're essentially using those high volume trades as an alibi. Zero profit is fine. All of this is an open secret. Hell, it's why dark pools exist. Those mean ungrateful dark pool customer fucks apparently decided that they didn't want all of that liquidity that HFTs were generously providing when they bought and sold. So rude.
crdoconnor 9y ago
>HFTs cannot see a customer’s order before it executes on an exchange. They cannot “jump in front of” an atomic order. I'm pretty sure this is what bid stuffing was all about.
- dsacco 9y ago
  No, quote stuffing can be used in latency arbitrage between different exchanges, but it is literally not possible to peek at an order before it arrives on at least the primary exchange of origin.
  - crdoconnor 9y ago
    You say potato, I say potato. I say "jump in front of" you say "latency arbitrage".
    - dsacco 9y ago
      Okay, you can use different terminology if you’d like, but per your original comment I responded to, it is still impossible to do.
      - crdoconnor 9y ago
        So you agree that jumping ahead of an order via quote stuffing is possible but jumping in front of an "atomic" order is not? I guess it's technically correct but it's not a particularly meaningful distinction given the context.

by dsacco 9y ago

cryptonector 9y ago

HFTs == market makers. You need market makers in order to have a functioning, liquid market. Absence of market makers -> higher volatility. Market makers can lose a lot of money under certain market conditions. Their profit margins are rather small. Market makers are practically heros.

vinceguidry 9y ago
It's a bit naive to equate the two. Plenty of strategies can be found that leverages high transaction speed and volume that aren't market making. What defines market making is the fact that you buy and sell the same commodity at the same time, taking advantage of the spread to make money. Two other rapid transaction arbitrage strategies involve market manipulation, in which trades are placed that are designed to fool other trading algorithms into believing that the price is going to move a certain way, and statistical arbitrage, which searches for temporary price differentials across exchanges. Neither of these two strategies create liquidity on the market, that is, offers more opportunities for anyone who wants to buy or sell to do so. Instead they seek to create, and/or exploit opportunities.
- cryptonector 9y ago
  ISTM that high frequency offers/bids that are not meant to complete will tend to become noise that the market filters. That's my sense, though admittedly I don't know enough and have no data to back that up. And when such offers/bids do complete, they really are market making. Either way, this is helping the market discover the price -- you call label this with "exploit", but it's not clear that that's anything but bias. Normal market making too could be seen as exploitation. Suppose the market is crashing. Without market makers the market ceases to function and uncertainty worsens. With market makers the market functions and crashes as much as it wants to. So your stock holdings lose value and here's a market maker glad to help that happen. How is that not "exploitation"? Politics is a war of words, I get that. I guess I should be happy that nowadays we don't punish market makers.
- dsacco 9y ago
  > What defines market making is the fact that you buy and sell the same commodity at the same time, taking advantage of the spread to make money. You need not be a formal market maker to increase liquidity. Technically speaking even retail customers will receive small transaction fees for increasing liquidity in certain circumstances where it's particularly lacking. More generally speaking, any strategy that rapidly connects two parties is market making, even if the two parties are not immediately next two each other in the strategy's trading queue. > Two other rapid transaction arbitrage strategies involve market manipulation, in which trades are placed that are designed to fool other trading algorithms into believing that the price is going to move a certain way Market manipulation is not a good term to use for this, and it's a fairly serious accusation to levy against a trader. Market manipulation is an illegal activity that typically involves cornering or collusion. More abstractly, this is a net improvement in price discovery, because one party notices that another party can be fooled (which means their algorithm has a glaring pricing inefficiency). In repeatedly taking money from the other party, the other party becomes incentivized to stop executing that strategy. Once there is no counterparty to take that trade anymore, the inefficiency has been traded away, and the market is now (likely imperceptibly) more efficient. > Neither of these two strategies create liquidity on the market, that is, offers more opportunities for anyone who wants to buy or sell to do so. Instead they seek to create, and/or exploit opportunities. This creates liquidity. Any situation that increases trading activity increases liquidity by definition. You do not need purely symmetric market making to improve overall liquidity - if many low latency trades are being executed, and they are finding counterparties, the liquidity of that market is increasing as a result of the volume. HFTs do not buy and hold, which means they are rapidly connecting buyers and sellers, even if it individual securities are bought and sold asynchronously.
crdoconnor 9y ago
The sheer number of market making trades provides ample cover to hide some of the less legit trades if/when regulators come a-knocking. As you said, market making is super low profit. HFT certainly used to be a fairly rich source of profit. Not sure if what it's like these days. This is all probably just a coincidence.
- cryptonector 9y ago
  Regulators should audit regardless. It's not like we should want lower volumes just to make it easier on auditors. We can't exactly go back to 1920s tickers so as to have lower volumes...
  - crdoconnor 9y ago
    Regardless of whether they audit or not, it's germane to recognize that something is off when a supposedly low profit activity is actually generating quite a lot of profit.
    - cryptonector 9y ago
      Low profit-margin x high volumes == potentially big profit. Nothing's wrong with that.
      - crdoconnor 9y ago
        Which would work if there were one or two players trading an incredibly high volume, but that neglects to consider a little thing called competition which annihilated those margins. Now that doesn't matter so much if you're essentially using those high volume trades as an alibi. Zero profit is fine. All of this is an open secret. Hell, it's why dark pools exist. Those mean ungrateful dark pool customer fucks apparently decided that they didn't want all of that liquidity that HFTs were generously providing when they bought and sold. So rude.

crdoconnor 9y ago

>HFTs cannot see a customer’s order before it executes on an exchange. They cannot “jump in front of” an atomic order. I'm pretty sure this is what bid stuffing was all about.

dsacco 9y ago
No, quote stuffing can be used in latency arbitrage between different exchanges, but it is literally not possible to peek at an order before it arrives on at least the primary exchange of origin.
- crdoconnor 9y ago
  You say potato, I say potato. I say "jump in front of" you say "latency arbitrage".
  - dsacco 9y ago
    Okay, you can use different terminology if you’d like, but per your original comment I responded to, it is still impossible to do.
    - crdoconnor 9y ago
      So you agree that jumping ahead of an order via quote stuffing is possible but jumping in front of an "atomic" order is not? I guess it's technically correct but it's not a particularly meaningful distinction given the context.