3 ms·

I am a professional trader, and by almost any definition I operate in the "high frequency" space. First, let's establish that Traders is an authority on the re

I am a professional trader, and by almost any definition I operate in the "high frequency" space. First, let's establish that Traders is an authority on the real world of financial markets in the same sense that PC World is an authority in the world of technology. So I've not read the linked article, nor am I going to. But I will say this: HFT does perform a viable, necessary economic function. A well-functioning capital market absolutely requires this kind of activity. HOWEVER, like most mainstream-media memes, what gets talked about / opined on is almost never relevant to what is actually important and/or controversial: in this case, the question of whether HFT creates a two-tiered playing field where individual (read: non-technically-sophisticated) investors suffer at the hands of the "pros". Most arguments against HFT basically say that algorithms are purely predatory and only serve to hurt the performance of large investors. This is naive at best and deceptive at worst; for every share I purchase "ahead of" a big order, a seller has been filled at the price he desired. Every transaction has two sides; you can't just pick one and say they got screwed. The other side has to have done as well as the other did poorly (assuming a fictional frictionless world). The reality is that HFT requires tons of knowledge and a technology budget of seven figures per annum at the barest minimum, and this provides a very real barrier to entry. What should be talked about, but never is: is that ok? Why or why not? What ramifications does it have?

richardw 17y ago
Each time you buy ahead of someone they lose the money you make. It's a zero-sum transaction. Instead of being between a buyer and seller, it's now the buyer, you and the seller. You make enough money over the year to justify the seven-figure technology budget, and all of your profit has to come from the other two parties. (Well, and the money you might get as rebates from the exchange.)
- brownegg 17y ago
  You make the assumption that they would trade. That assumption is only valid because of people like me. I don't need to justify my activity, nor do I want to go that route--my point is that the fact that participants can realistically expect a fill is not something that happens because of magic.
  - crux_ 17y ago
    > You make the assumption that they would trade. That assumption is only valid because of people like me. You make the same assumption yourself, every time you trade. The entire purpose of high frequency trading is to beat others to good deals -- not to do a better job of discovering good deals.
    - brownegg 17y ago
      Completely untrue. Pure speed games are "winner take most", and brutally difficult. It is very hard to win on speed.
      - andrewcooke 17y ago
        so how do you win? my naive impression is that speed must play an important part, or it wouldn't be called high frequency. (so speed is important to you, just as water is important to a fish; it may not be what you focus on, but that doesn't mean that it's not critical).
        brownegg 17y ago
        It's crucial, for sure. But there are trades for which it is the only determinant, and those are dangerous waters. There are market participants who make their money on pure latency arbitrage; it is a 100% speed game (because anyone can figure out that 1 - 1 = 0). The majority of HFT falls into the "you have to be fast enough to not get run over" bucket.
        andrewcooke 17y ago
        ok, so i think the original thread here is arguing that "being run over" is just the market doing what it would do anyway, without you playing in the traffic and endangering us all...
      - crux_ 17y ago
        If it's completely untrue, then I'd like to hear about how your computerized trades ever "discover" a good deal. I'm fairly sure this is the answer: They don't -- instead, they're just good at noticing, very quickly, the actions of the market actors who do. Which is just another way of saying that they're about beating others to the deal. (I'd also guess that they're probably also quite good at handing recreational day-trader's asses to them.)
rgarcia 17y ago
There are lots of markets with high barriers to entry. Does the fact that Kinkos invests millions in fancy printers make the photocopying business unfair?
coreyrecvlohe 17y ago
I'm not a professional trader, like yourself, but from what I've gathered over the last several years of HFT coming up in the news is that it is in fact a low-barrier to entry field. All you need is a decent quant, application designer, and some co-located machines as close as you can get them (and other easily acquired amenities.) So I could make a reasonable bet you could open a HFT firm with a couple hundred thousand, plus the talent. (And I've confirmed this with a few hedge fund guys, and they pretty much agreed.) And I do completely understand the argument for liquidity: more transactions = more accurate price discovery. But I think problem here is that this entire field is black box, meaning property trading algorithms and trading patterns can and are used within the system. This can allow a trading AI to go out into the market place, look for pattens, and to create and cancel millions of orders within the fraction of a second. I'm sure you know about the former Goldman Sachs programmer who was charged with theft by the FBI. It was totaled at around 32 megabytes of software code. Not very much. But Goldman insisted that if this code got out into the market it would be detrimental to their business and violate their trades secrets confidentiality clause. Not to mention Goldman's largest profit center in their business is their proprietary trading desk, which is heavily into HFT. My contention is that if we are really interested in a utility that uniformly benefits the market, then let's have an open source platform that provides that, so we can verify that any of these firms aren't "front-running" their trades. Thats my opinion. I think a lot of the debate over HFT is just filler, it doesn't matter. The guys who are hip to the scene are already making a killing on it now, and it may last a few more years before we begin to regulate.
- brownegg 17y ago
  Sure, you can get the point where you're placing trades for a couple hundred grand... as long as a lot of things fall into place for you. Writing your own apps? Unless you think you can get talent that is willing to roll the dice and possibly be out of a job in a couple months, you're looking at 12 months x the cost of those people... how much is that? $500k? This isn't framework-style assembly of components. If you want to be competitive, the SLOWEST things you can use are linux and C++. Colocation agreements usually require a minimum commitment of 12-36 months... probably $10k/mo if you have some rudimentary failover and the like. I'd say anyone who tries with less than 12 months and $500k to truly burn has 0 shot at success. And if you want to really do it right, you're looking at an order of magnitude more.
jfager 17y ago
for every share I purchase "ahead of" a big order, a seller has been filled at the price he desired. But they would have been filled at the price they desired if they sold directly to the big order, as well. The end effect of the hft systems is just to run up the big order to the max of what they're willing to pay, and take their profit as the difference between that max and the original offer (split up between however many of these things managed to make it over to the feeding frenzy before the real buyer got what they wanted).

richardw 17y ago

Each time you buy ahead of someone they lose the money you make. It's a zero-sum transaction. Instead of being between a buyer and seller, it's now the buyer, you and the seller. You make enough money over the year to justify the seven-figure technology budget, and all of your profit has to come from the other two parties. (Well, and the money you might get as rebates from the exchange.)

brownegg 17y ago
You make the assumption that they would trade. That assumption is only valid because of people like me. I don't need to justify my activity, nor do I want to go that route--my point is that the fact that participants can realistically expect a fill is not something that happens because of magic.
- crux_ 17y ago
  > You make the assumption that they would trade. That assumption is only valid because of people like me. You make the same assumption yourself, every time you trade. The entire purpose of high frequency trading is to beat others to good deals -- not to do a better job of discovering good deals.
  - brownegg 17y ago
    Completely untrue. Pure speed games are "winner take most", and brutally difficult. It is very hard to win on speed.
    - andrewcooke 17y ago
      so how do you win? my naive impression is that speed must play an important part, or it wouldn't be called high frequency. (so speed is important to you, just as water is important to a fish; it may not be what you focus on, but that doesn't mean that it's not critical).
      - brownegg 17y ago
        It's crucial, for sure. But there are trades for which it is the only determinant, and those are dangerous waters. There are market participants who make their money on pure latency arbitrage; it is a 100% speed game (because anyone can figure out that 1 - 1 = 0). The majority of HFT falls into the "you have to be fast enough to not get run over" bucket.
        andrewcooke 17y ago
        ok, so i think the original thread here is arguing that "being run over" is just the market doing what it would do anyway, without you playing in the traffic and endangering us all...
    - crux_ 17y ago
      If it's completely untrue, then I'd like to hear about how your computerized trades ever "discover" a good deal. I'm fairly sure this is the answer: They don't -- instead, they're just good at noticing, very quickly, the actions of the market actors who do. Which is just another way of saying that they're about beating others to the deal. (I'd also guess that they're probably also quite good at handing recreational day-trader's asses to them.)

rgarcia 17y ago

There are lots of markets with high barriers to entry. Does the fact that Kinkos invests millions in fancy printers make the photocopying business unfair?

coreyrecvlohe 17y ago

I'm not a professional trader, like yourself, but from what I've gathered over the last several years of HFT coming up in the news is that it is in fact a low-barrier to entry field. All you need is a decent quant, application designer, and some co-located machines as close as you can get them (and other easily acquired amenities.) So I could make a reasonable bet you could open a HFT firm with a couple hundred thousand, plus the talent. (And I've confirmed this with a few hedge fund guys, and they pretty much agreed.) And I do completely understand the argument for liquidity: more transactions = more accurate price discovery. But I think problem here is that this entire field is black box, meaning property trading algorithms and trading patterns can and are used within the system. This can allow a trading AI to go out into the market place, look for pattens, and to create and cancel millions of orders within the fraction of a second. I'm sure you know about the former Goldman Sachs programmer who was charged with theft by the FBI. It was totaled at around 32 megabytes of software code. Not very much. But Goldman insisted that if this code got out into the market it would be detrimental to their business and violate their trades secrets confidentiality clause. Not to mention Goldman's largest profit center in their business is their proprietary trading desk, which is heavily into HFT. My contention is that if we are really interested in a utility that uniformly benefits the market, then let's have an open source platform that provides that, so we can verify that any of these firms aren't "front-running" their trades. Thats my opinion. I think a lot of the debate over HFT is just filler, it doesn't matter. The guys who are hip to the scene are already making a killing on it now, and it may last a few more years before we begin to regulate.

brownegg 17y ago
Sure, you can get the point where you're placing trades for a couple hundred grand... as long as a lot of things fall into place for you. Writing your own apps? Unless you think you can get talent that is willing to roll the dice and possibly be out of a job in a couple months, you're looking at 12 months x the cost of those people... how much is that? $500k? This isn't framework-style assembly of components. If you want to be competitive, the SLOWEST things you can use are linux and C++. Colocation agreements usually require a minimum commitment of 12-36 months... probably $10k/mo if you have some rudimentary failover and the like. I'd say anyone who tries with less than 12 months and $500k to truly burn has 0 shot at success. And if you want to really do it right, you're looking at an order of magnitude more.

jfager 17y ago

for every share I purchase "ahead of" a big order, a seller has been filled at the price he desired. But they would have been filled at the price they desired if they sold directly to the big order, as well. The end effect of the hft systems is just to run up the big order to the max of what they're willing to pay, and take their profit as the difference between that max and the original offer (split up between however many of these things managed to make it over to the feeding frenzy before the real buyer got what they wanted).