Tradeworx' defense of their HFT strategies

Discussion in 'Automated Trading' started by MGB, May 6, 2010.

  1. … and a laughable attempt on pgs 7 - 8 to portray HFTs as the only market makers in the market, apparently serving us all by providing “orderly markets” (like yesterday’s!) ...

    And also this ...
    Poor, poor HFTs ...
    l
     
    #11     May 7, 2010
  2. I read that definition of high frequency trading too and thought it was just odd. I would imagine they wouldn't want to keep their positions open at the end of the day, but I would imagine something that is more rapid than daytrading would be going into the range of seconds and even shorter.

    I wondered about their argument that we've already passed the big hump in profits from HFT. Since everybody's obsessed with it I'd think there's more to be had, but that might just be that it's a fad now.

    If there was something to really frustrate me it's that they submitted a document in the form of Powerpoint slides. Does anybody write documents anymore? :(
     
    #12     May 7, 2010
  3. I think they should just delay the quotes at the data center to the point that everyone gets them at the same time. Use an average of the retail quote systems latency as a benchmark. Oh, and everyone agree on the minimum tick for all instruments and make it the same for all participants.
     
    #13     May 7, 2010
  4. MGB

    MGB

    That doesn't work.

    Latency from the data center to you and me will always be different. How will you control the latency so that you and I will get it at the same time? You can't.

    Latency from the you and me to the data center will be always be different. Again, how will you control the latency so that my price arrives first at the data center because I submitted it first even though my latency is slower than yours?
     
    #14     May 7, 2010
  5. rosy2

    rosy2

    thats the glamorized media version. sending orders and handling data feeds is the microsecond latency stuff you hear about but the trades themselves last longer than you think.
     
    #15     May 8, 2010
  6. I may be wrong but the way I see it is that HFT systems rip off those who place market orders and if they harm anyone, it is traders who have the same goal in mind as HFT- i.e. profiting from intraday volatility - but they do not have a better infrastructure.

    HFT systems can do no harm to limit orders. Just decide what price is fair for you and place your order. No HFT system can hurt you. At the end of the day IMO, this battle is between sophisticated HFT, individual trader morons who do not understand the game and stupid long term investors who place large market orders because it is OPM, not their money and they claim not to care about entry cost because they have a looooooooooong investment horizon (idiots). But when the stock falls they realize that entry cost impacts bottom line.

    I say let HFT punish the other morons until they learn. Why should any government protect morons? Since when, speculation, stupidity and lack of knowledge are protected qualities?

    If anyone does not agree, just explain to me how HFT can impact a limit order. Also explain why some morons who want to make a living out of rebates should cost government money by trying to protect them by holding special hearings and hiring consultants to explain to government workers how HFT works.
     
    #16     May 9, 2010
  7. Dude, sorry to put it bluntly, but either you have never been sub-pennied, or you are an idiot.

    There are many stocks I have traded, in the ~1.00$ range, that are subpennied. What does this mean? well, let's say that ABK is trading at 1.01 bid by 1.02 ask. And lets say you want to get a fill on the bid. If you put in a bid at 1.01, and you were the only one to have that bid price, any shares sold would get sold to you.

    This, however, is where the subpennying screws you... immediately after seeing that 1.01 order, a computer puts in an order at 1.0101.

    This means that your order will never be filled... Instead, you have to buy at the ASK in order to get filled. Which means you take an automatic 1% loss. And this is just one side of the transaction. When you try to sell, you cannot sell into the ask, because, lets say you want to sell at 1.02... immediately after placing your ask, you get subpennied and a computer steps in front of your order and puts an ask at 1.0199. This means you will never get filled on your ask, and there is nothing you can do about it because you, as the retail trader, cannot place subpenny orders... So you have to sell to the bid. BAM, another 1% loss.

    Over time, these add up.

    Another example which I have had happen to me MANY TIMES is where you can see a large bid which has been sitting there for several minutes. However, when you go to sell a limit order maybe 2-3% BELOW that price which is displayed, that limit order immediately disappears BEFORE filling your order, and your sell order does not take that bid out.

    I have had that happen with me with ETRM, and 100K share sell orders... The computer will have a large bid displayed, that sees your ask, which was placed BELOW that bid, then immediately that bid gets pulled before you can sell your shares to it.

    g
     
    #17     May 9, 2010
  8. HFT=bloodsucking leeches.
    Ought to be banned either through transaction tax, rules or other barriers.

    They say they "provide liquidity" by taking the other side of investors' trades; of course they do, that's the only way to skim money off the table!

    That's what these assholes do, make the market work the way of a slot machine.
     
    #18     May 9, 2010
  9. This comment by tradeworx is TOO FUNNY in light of the happenings on Thursday - especially this part :

    •
    How real is the risk of “runaway algorithms” in terms of likelihood?
    •
    Not very likely: empirically, this has been a non-issue thus far, and HFT is a mature business that has withstood many turbulent periods
    •
    Exchange-based risk checks, such as port-level limits on order size, along with post-trade risk checks by b/d's, eliminate most of the risk
    •
    Reg NMS prevents people from trading so quickly that their broker wouldn't notice excess risk on a post-trade basis
    •
    Orders to buy or sell shares at far-from-market prices are likely to get rejected because of Reg NMS
    •
    How real is the risk of “runaway algorithms” in terms of magnitude?
    •
    Huge pnl losses can not occur "instantaneously" -- they take minutes or hours to materialize even in extremely volatile markets
    •
    thus, HFTs or brokers using post-trade checks can cut their systems off with ample notice when losses mount
    •
    Exchanges already have policies to break trades that are at obviously wrong prices, and halt trading in the event of “extraordinary activity”



    Ha, ha, ha. ROFL ! I think, I have to contact Manoj Narang and have a chat with him what his firm's P&L was on Thursday...
     
    #19     May 9, 2010
  10. FYI I do not trade stocks priced below $20.

    Then, what you describe is typical MM activity and not necessarily HFT.
     
    #20     May 9, 2010