Trying to understand slippage: 1.6% with Interactive Brokers.

Discussion in 'Order Execution' started by mickael28, Feb 8, 2019.

  1. dinn13


    wow, been nearly 15 years since I signed up, time sure does fly.. I run a 600 M equity book for a multi-manager fund, pretty high turn doing our own execution, collocated and what not and so have access to and use a ton of data
    #31     Feb 9, 2019
    traderob and murray t turtle like this.
  2. Thanks so much for having had a look at those details...

    Do you know any good retail broker with a good infrastructure for executions? I'm mainly trading US stocks, scaling in and out, so I was with IB mainly due to their low commissions and because I could borrow the US funds rather than having to convert foreign currency, but these bad fillings keep on happening so they're making the system trickier to operate as I'm usually taking more risk than planned due to them...

    Who are *they* in these sentences? ie, if my orders were already in the market not sure who decided that the first fill was just good for 100 shares and all the others will be executed a little bit later (when a larger slippage had happened).

    Sorry, what did 'sor' mean in the sentence of "before they evaluated their sor again"?
    #32     Feb 9, 2019
  3. Gotcha


    Impressive. But let me ask you this if I may, and it is a serious question. If you stopped working and had to be a retail trader from home, could you do it? Lets assume that you have the same constraints as most other guys (ie. not co-located, maybe you're not able to get the micro second data you showed here, and in some ways you would probably be at the mercy of having to use stops, which is something I assume funds don't do)

    I ask because here at ET, its often discussed how badly hedge funds do in a down turn, since it seems like many barely outperform the general market, and hence are only investing long, but its ok cause they live off of fees.

    And if you feel confident in your ability to trade as a retail trader vs. an institutional guy, what would you do different? Perhaps your answer will be that you won't even try because without the edge of being co-located and having algos for execution that won't be preyed on by HFT, it would be a hopeless proposition?
    #33     Feb 9, 2019
    Onra likes this.
  4. JSOP


    Ok so this is to prevent stop-hunting then. Even for that, a stop-limit order would've done the job. If the stop is triggered, the order won't be filled until the limit price is hit, at the higher price or not at all if it was truly a nasty stop-hunt if the price spikes right back up after the stop is triggered and at the same time it prevents slippages. You still won't need the OCO order.
    #34     Feb 9, 2019
  5. dinn13


    Don't know of a better retail broker, I'm not allowed to trade actively due to compliance reasons so haven't looked into it in probably over a decade.

    The they is Interactive Brokers, so 93ms after the stop was triggered there was the opportunity to take another 400 shares at 18.24 on top of the 100 shares you did get executed and thus reduce your slippage a bit.

    As far as I can tell IB was only sending IOC orders to hit orders already resting in the book. I don't see any limit orders in the book that would be obviously yours. So your orders were never in the market other than than when IB was trying to hit a specific quote. So it was IB that decided when to fill your first shares and then the rest. My guess is there was latency in their system that caused the 1 s delay for the last set of execs. If it was my system I'd be digging into it and fix it since 1 second is egregious.

    sor is a smart order router (and it's effectively the "who decided" to only fill the 100 shares and then wait a second to execute the rest), it's the piece of software that will determine where to route orders, anything as simple as looking at the top of book across exchanges and routing there, or fancier stuff like having some machine learning algo to forecast where some dark liquidity is and then ping those venues (if you ping too many venues then might add too much latency and increase slippage)
    #35     Feb 9, 2019
    Onra, mickael28, qlai and 1 other person like this.
  6. dinn13


    retail trader with how much capital? but I currently trade my personal account and have made 20% returns the past few years although largely due to luck as opposed to any quantifiable alpha. and when I was in college was making 10k a month trading on IB with some family money (100% returns taking a lot of risk but 15 years ago now), basically as a retail trader have to make trades that larger players are ignoring due to scalability issues. For example at work I have a signal that is 4+ sharpe and ~8% returns (unlevered) but it doesn't scale at all and doesn't net with other things, has to be run separately, so could maaaaybe run 5 M on it but not worth the operational headache so I don't run it. So could maybe find things like that to trade retail, then again at work my commissions for trading are effectively 0 and pretty low financing, so not sure if that strat could survive retail commission rates

    but yeah I couldn't really do what I currently do at work in any way shape or form. The data, connectivity, and compute is way too expensive, literally millions a year and very much dependent on all of it.

    there are so many hedge funds out there with different objectives that it's hard to draw any general conclusions from, so can only speak for myself but downturns have been great for me. 2008 was my best year, the flash crash best day ever, early 2018 with that vol absolutely printed money.. I do great in downturns and high vol, and often do ok or just blah when vol is low, which makes sense in that expected returns are going to be greater with higher vol and my role is effectively to transfer risk between longer term market participants and with higher vol there is a greater need for that.
    #36     Feb 9, 2019
    dakr, HobbyTrading, Onra and 3 others like this.
  7. qlai


    Awesome insight! I just want to mention that IB has many "smart" options to choose from and ET members like @d08 contributed greatly on the subject.
    #37     Feb 9, 2019
  8. qlai


    So let's say you have a OCO stop limit at 10.00 and stop market at 9.00. The 10.00 stop is hit and your order goes to NASDAQ as limit offer 5000@10.00. The stock is trading all the way down to 9.75, your market stop order not hit yet.
    It is very likely that the 5k resting order will get attention (from dinn13 perhaps) and someone will figure out a way to buy from you and flip at a profit. You are happy with zero slippage and they are happy for the option your order provided them. Everyone happy. Win-Win.
    #38     Feb 9, 2019
  9. Could this issue with the slippages I am experiencing be related to the following?

    I was assuming that when my Stop order was triggered IB would send it as a market order and that's it. But now I read about this:

    #39     Feb 9, 2019
  10. Robert Morse

    Robert Morse Sponsor

    I try not to be blunt, but in this case, I think it will help you. IMO, don't trade illiquid stocks with wide spreads. If you decide you have to, don't use stop orders and manually enter and exit your positions and assume to enter and exit, there will be a cost. As much as I would like to blame another broker, these are choices you make and depth and liquidity are not the responsibility of your executing broker. They are subject to market conditions at the time you enter and exit.
    #40     Feb 9, 2019
    tommcginnis, qlai and Overnight like this.