TradeStation slippage parameter for NYSE stocks with >1M volume

Discussion in 'Strategy Development' started by garchbrooks, Mar 14, 2010.

  1. What's a realistic slippage parameter for backtesting NYSE stocks, assuming I don't trade right off of the open, if the stock trades 1Mil shares, 2Mil shares, and 2Mil or more shares per day?

    Also, would I be better off routing orders to ARCA or using that "Intelligent Routing" feature?
  2. I did a study back in 2006 and early 2007 for my automated trading for 23 stocks trading over million shares a day where all stocks were priced over $20 a share at the start of trading. This was a low volatility environment for the most part back then. Here is the table I built.

    Price Range Dollar Market Slippage per share

    General 0.074
    20 - 50 0.031
    50 - 80 0.072
    80 - and Up 0.152

    (Table did not display right. )

    These are average numbers. General is the composite of all ranges. If you are trading Google today $0.30 may be a better slippage? Some day I will go back trough my journal an build a new one
  3. Hey, this is really great information. Wondering if you split the data up by time of day on the fill? My opening fills kind of match your numbers, but trades after 10am eastern on the NYSE are much more close to the perceived inside bid/ask, but I've also seen some really delayed, terrible fills.

    Other question I have is -- can you re-run the study and look for the median slippage? Also, one more question. How many shares per order? I'm assuming slippage on a 100 share order is going to be a lot less than slippage on a 1000 share order or more. Since I primarily trade lots of small sizes, I'm trying to figure out just how terrible my slippage will really be.

    Obviously, don't re-run the study if it takes you too far out of your way or wastes time. I'm just trying to get a feel for what's appropriate, because it's just heart-breaking when I see what looks like a beautiful edge get vaporized by commissions and slippage.
  4. I did some checking and the raw data for the study was deleted back in 2008. Apparently I thought in was no longer current so I trashed it.

    Some of the other info I still remember. My automation is for 80% swing trading in the Daily time interval with the remainder intraday. So the vast majority of the trades are at the opens on daily trades. The number of shares started at 200 to 300 with higher priced stocks, many lower price stocks started at 500 with a few up a 700. I roll over profits so most share counts rose as profits were greater.

    I was planning on doing this again in a year

  5. Ok, now the data all makes sense. Thanks for your help, though. Was very enlightening.