Slippage of MOO/MOC

Discussion in 'Order Execution' started by albertly, Sep 4, 2011.

  1. albertly

    albertly

    It would be nice to have statistics about slippage in execution of MOO/MOC orders.
    I want to start new thread where everybody can publish real execution of MOO/MOC orders

    Broker| Date |Security | Order |Open Price |Filled Price
    IB | AUG 31|SPY | MOO BUY |122.46 |122.48


    P.S. It's real data.

    Plz, publish your data. It would be beneficial for everybody
     
  2. rmorse

    rmorse Sponsor

    MOO and MOC orders sent to the NYSE all get one price. So what type of slippage are you referring to?
     
  3. albertly

    albertly


    The open price was 122.46 - On Yahoo, On Bloomberg, IB's TWS also gives 122.46.

    Filled price was 122.48.

    I've got plenty of explanations that is not slippage and an official opening price of SPY on ARCA was indeed 122.48 and Yahoo and others publish price of first trade - blah, blah, blah.

    What the hell, the matter should be pretty clear for everybody who trades with automatic strategy.

    Call it - "slippage" or call it - "table" - it does not matter. What is really matter that my backtest uses 122.48 and not that official, hidden open price.
     
  4. rmorse

    rmorse Sponsor

    If you route your orders to the primary exchange, not a smart or managed route, there is one price that everyone gets. Smart routes and managed routes for open and close don't require best execution because no one know what the market is yet.
     
  5. albertly

    albertly

    I am not arguing with that.

    When I build strategy I've got one open price. All data feeds gives me so-called "first trade price" as open price. But real execution is different.

    There are two solutions for it.
    1. Get an official open price from data feed (I don't know where and how).
    2. Make a correction. To make a correction I need statistic.

    What I call "slippage of MOO" it is difference between execution price and my feed's open price.

    I need it to make my test as realistic as possible.
    I think many traders have the same problem as me.
     
  6. rmorse

    rmorse Sponsor

    It sounds to me like what you need is an accurate Open/Close price data source of the primary markets, NYSE, NYSE ARCA and NASDAQ, and the websites you're using don't offer that.
     
  7. albertly

    albertly

    The same as IB's data feed.

    Interesting what was the open price of SPY (AUG 31) on your data feed ?
     
  8. rmorse

    rmorse Sponsor

    I don't have my own feed. The NYSE charges a lot of money for historical data that will be accurate. Live Vol is a good source of historical data and you can pay for just want you want.

    http://www.nyxdata.com/Data-Products/NYSE-Arca

    http://www.livevol.com/historical_files.html

    I'll look around for a free source and let you know if I find it. What you need is the Open, Close for the primary excluding the ECNs. MOO and MOC orders sent to a Managed route or Smart routes go to equity market makers that pay for order flow. The open and close are always funky because it's hard to enforce best execution without the other exchanges to compare to at that time. I would avoid that.
     
  9. The best solution is to test your systems using intraday bar data. For example, what is realistic is using the close of the first 5min bar of the day (i.e. the 935AM bar close) and the second to last closing bar of the day, the 355PM close.

    Relying on executions based off the daily opening (and closing) is not a good way to go, unless, you're trading the opening cross. In that case, you're going to have read up on the exchange's rules and procedures for properly trading the opening.

    This might be of value:

    http://www.nasdaqtrader.com/trader.aspx?id=openclose
     
  10. evo34

    evo34

    In your example (ARCA-listed ETF), the probable reason you see a slightly worse price in your order blotter is that IB includes the comm as a part of the executed price.

    In other cases (NYSE-listed stocks), the published open (first trade on any ECN) can be massively different than the primary exch. open, which will occur whenever the specialist feels like it.

    In other news, the NYSE sucks.
     
    #10     Sep 11, 2011