HFT Myths

Discussion in 'Automated Trading' started by hft, May 3, 2013.

  1. nitro

    nitro

    Have you actually tried anything of what you are quoting?

    People throw this stuff around without any idea how complicated any of it is. The hardware is the easy part, and even that requires a full time Network Administrator and a full time Systems administrator.

    One, day, one day, I am going to write a book.

     
    #271     Aug 26, 2013
  2. baglunch

    baglunch

    Yes, I do this kind of stuff every day, it's my job.

    Is there something that I said that is untrue or out of date? As far as I know it is relatively true. i am somewhat vague just because if we were to go into every aspect like numa or non blocking sockets or cache management we might as well write a book.
     
    #272     Aug 27, 2013
  3. zmax, who are you kidding 2 posts and already a discredited dweeb.

    trade at IBKR, they won't be kicking you anywhere.

    clowns
     
    #273     Aug 27, 2013
  4. zmax

    zmax


    I don't have time to lose on forums that's why 2 posts. If you don't believe what i say that's your prob, but i'm pretty sure others here had probably pretty similar experiences in the FX environment.
    IBKR is a shit broker for this, fills over 100ms are way too much, the topic of the discussion is hft...
     
    #274     Aug 28, 2013
  5. 100ms is too much you're right. What you need is a fill yesterday.

    get a life
     
    #275     Aug 30, 2013
  6. hft

    hft

    Generally speaking, liquidity providers gross negative and make on rebates. Liquidity removers gross positive.

    Sure they exist and HFT's use them, EDGA for instance

    I've never personally 'dinked around' with it, if that means trying to alter the opening/closing print. Not that I'm truly against it, but just something I've never pursued. I've never seen it personally done, but am sure others are doing it in some manner.
     
    #276     Sep 10, 2013
  7. jb514

    jb514

    Is SPY ES arb kind of a winner take all trade? Or is there an amount of profit sharing since the liquidity at different endpoints are often from algos watching the same signals? Also what do you think pnl looks like? I'm guessing it's probably similar to your previous example with majority of large trades scratch

    What do you think about arbing ETFs as in trading two different arca listed ETFs. Do you think it's possible to be competitive from one co-lo or do you have to be at multiple exchanges?
     
    #277     Sep 10, 2013
  8. hft

    hft

    I disagree on a few points. I don't think Reuters and EBS report specifics about counterparties to banks. I agree that CRNX and HSPX do. Keep in mind that 'banks' might be providing liquidity technically, but it's not necessarily (and even unlikely) that the bank is sourcing the liquidity. For instance, when I submit orders it shows up on the tape as my prime broker (say Goldman), but I'm behind the liquidity. So if I get Goldman to tell me what their counterparty makeup is, I still really don't know who was behind the orders that I'm making/losing money on. Yes, you can see some patterns, but it's pretty hard to break it down in great detail to exploit them. Again, this is EBS and Reuters specific only.

    I would think that if you trade through a big broker on the majors (EBS + Reuters) that you'd be able to get by unless you're REALLY raking it in, and you'd only get called out if you're taking out mispriced orders way off the book (topic of another discussion, yes it's unfair but it only happens every so often). Else, just use CME FX futures instead if you don't need decipip pricing. No way you get kicked there unless you blow messaging limits.

    Either you're talking about decipips, or you're dealing with the less liquid pairs, which I'm really not familiar with so maybe my thoughts are irrelevant. I've only traded the majors (G7 vs. USD).
     
    #278     Sep 10, 2013
  9. hft

    hft

    I've never always been the kingpin at ES-SPY (might have been at times :)), so I can assure you that there's some money to be split up amongst the rest of the group. It's impossible to narrow down the PNL from that one particular strategy since you're also often scalping interchangeably with spreading and also spreading to other products loosely in a portfolio.

    I think arbing any US equity requires colocation and all lit exchanges at the very least, and also good dark pool access to make consistent money.
     
    #279     Sep 10, 2013
  10. jb514

    jb514

    Where where you trading equities from when arbing ES v SPY? Were you at all equity endpoints or does the latency from Chicago negate that need?
     
    #280     Sep 10, 2013