What % of daily volume in a trade do you start to run into significant slippage?

Discussion in 'Order Execution' started by yakyakgoose, Mar 19, 2016.

  1. I trade smaller stocks and the biggest elephant in the room for my system is at what point does slippage start to make my system invalid or become so significant it eats up my profits.

    I assume the logical way to think about this would be to think of it as a percentage of total daily volume. I also only do "at open" and at "close orders" so am trading at the highest volume times of the day.

    So...considering the following example....Stock XYZ is trading at $5 and trades 1,000,000 shares a day. Very roughly, how many shares traded at open or close starts to cause significant issues?
     
    Shadetree42 likes this.
  2. K-Pia

    K-Pia

    Slippage is a function of order size and liquidity.
    As you pointed out, liquidity is cyclical too.

    So How much ? is conditional on specific situation.
    It's too much when it becomes too expensive.
    And again ... Expensiveness is relative ...

    I'd consider it excessive if it takes 5 ticks to be filled.
    But it's my point of view only. Because 5 ticks becomes 50 ticks.
    Quick enough if everyone is caught by panic or any liquidity crisis.

    To be able to quantify how much liquidity lies at the bid & ask is insightful.
    I use a Deep Of Market to place my orders, so I can see the depth.
    Sometimes very big orders sit there, if you're lucky, it's for you.
     
  3. Ideally less than .5 % round trip. I have noticed that sometimes slippage is in my direction but I havent gathered enough data to see what averages out to.

    Has anyone here ever traded 2.5% of the daily trade volume in one trade? What happened?
     
  4. eganon69

    eganon69

    I have a simple rule that may sound a bit odd but I will explain.

    I trade (long/short) ABSOLUTELY no more shares than the 22 day average volume of the stock/780. ( MAX Shares <= (22 day average VOL/780)). There are 780 30 second periods in a trading day. I want to be sure that I can liquidate my entire position in 30 sec or less. Usually this means I need a volume of about 780,000 or more for whatever I am trading. Now I don't have the problem you have anymore. I used to sweat when I needed to liquidate my position trading a large volume of the daily volume. Follow this rule and sleep better.

    My .02
     
    hoodyap and Handle123 like this.
  5. I like your thought process but I may be limiting myself if I follow that exactly. My orders are only at open and close and the volume is significantly higher than the average during those times. I was trying to find a study I was reading earlier that I believe said 70-80% of volume is at the open and close and is trending higher. BUT...I should still be able to apply your logic to come up with something.

    If we assume 70% of the volume is in the first and last hour on a 1,000,000 volume stock

    1,000,000*.7=700,000

    Divide by 2 hours worth of 30 second intervals

    700,000/240=2, 916 shares
     
  6. Push it to the Limit;
    [​IMG]
    don't worry so much about becoming so big that you will slosh the market around ;)
    ...that's a good problem to have, assuming you even get to that point.
     
  7. Well its more that my backtested system assumes getting the open and close price within reason. Im just not sure how bad slippage can get and at what amount of volume.
     
  8. d08

    d08

    I try to be around 5-8% of the auction, it's quite high as others have said to never go above 1%.
    We can't really know what the slippage would be as your order can move the price which in turn triggers orders at other price levels.
     
    Shadetree42 likes this.
  9. I'd care more about knowing what the fair value of the stock with decent confidence than than what % of volume you are. If I was pretty confident of my model, I'd take a big % of volume at a good price.
     
  10. dratsum

    dratsum

    It's simple to measure the actual volume over the last 100 days and come up with an average or sort and rank the values ....or perform any kind of statistical analysis you're interested in. If you're not able to do this yourself then.... ask yourself why and make the appropriate changes.
     
    #10     Mar 22, 2016