Fading the opening gap

Discussion in 'Strategy Development' started by guy2, Jun 2, 2004.

  1. "There is no measure of the number of days it took to fill the gap. If the gap did not fill that day then it is regarded as a “no-fill” day."

    How about a simple example, correct me if I'm wrong. Close 1120. Open (as you describe it, which is simply where the E's are trading before the market opens)..at 1129. You are short at 1129. If it doesn't fill the gap, you just stay short? For days? Weeks?

    Have you tested the difference between opening price and closing price (9:30-4:00)? Any upward or downward bias?

    #11     Jun 2, 2004
  2. No, Bright's comment is inaccurate. Futures do have gaps. Especially futures that only trade during limited time periods like Nikkei, HSI, DAX, etc. Even the the US index contracts have gaps but they are obviously much smaller on average. (But he can't get you to sin city to trade those so you won't hear it from him.)

    I will also argue w/ db's second point. The ES and NQ futures are essentially traded 'all the time.' After hours the futures prices represent the best estimate of where the underlying indices should be trading were the underlying markets open to trade. IMO, this is where the opportunity lies. If one can better gauge where the underlying equity index will trade during the next trading day, one can take a position in the overnight futures market. Also, index arb is no longer taking place overnight so the futures price is thus more speculative relative to the underlying equities. Upon the open of the underlying market, arbs will bring the underlying and futures markets in line. No need to consider esoteric behavioral traits of traders.
    #12     Jun 2, 2004
  3. guy2


    Hi Don,

    If it doesn't fill the gap then the trade is closed at the closing price for the day. This sometimes produces a profit if part of the gap filled but as you would suspect it produces a loss most times. I have mentioned this in other parts of the study but I should have also stated this here in the introduction - thanks for pointing it out - I will be ammending it shortly.

    In answer to your second question: No, not yet. This I intend to do in the second study. The next study will be less concerned with generating statistics and probabilities and more focussed on applying the theory. I'm using a finer data set for that and will be running back tests which include scaling in and out of positions, improvement on entry prices and situations in which partials can be targetted beyond the gap fill.

    #13     Jun 2, 2004
  4. ES RTH from 9:30 to 4:15?? Ugggh.

    You might try trading gaps on small scale & discretionary for a couple of months. Take copious notes and use stops. Then reconsider the parameter of this study and consider starting over.
    #14     Jun 2, 2004
  5. dbphoenix


    His original remark was that futures trade "all the time". Therefore, I assume he was referring to those futures that trade all the time, not those that trade sporadically.

    As to my second point, if a trader holds a long position at some level when the equity markets close and the futures open 30 points lower the following morning, I suspect he's going to consider that to be a gap, whether it got there by an uninterrupted overnight process or not. Nothing esoteric about it. Just observable and predictable behavior.
    #15     Jun 2, 2004
  6. guy2


    In a virtually 24 hour market in order to implement a gap fade strategy you need to pick a start and end time. The most liquid times for the ES are when the pit is trading which is from 09:30 to 16:15 (I believe).

    I have tested other gap fade strategies with 1 minute data and ran the tests from 08:30/09:00/09:30 to 16:15 with the 08:30 start showing very profitable results. The data set I used for the study was, however, around 5 months of ES data and I did not consider this to be sufficient data from which to draw conclusions. I have subsequently obtained more data and will be re-examining the best entry time to use in this particular 24 hour market.
    #16     Jun 2, 2004
  7. Good. 8:30 isn't some random time. It is a time when a lot of economic news is released. Also consider when after hours earnings are released. Most of the better gaps are event driven. Many of these events are prescheduled like economic news and earnings, but others are not. Another event that comes to mind is the open of European trading.

    Thus quite detrimental to arbitrarily pick a single entry time. Better, IMO, to have a rules-based entry strategy that is time-flexible according to gap-driving events.

    Suddenly not so easy to program now is it? Profitable none-the-less.
    #17     Jun 2, 2004
  8. No reason for any argument whatsoever. I was only referring to the E's in this regard. And, of course .

    "".After hours the futures prices represent the best estimate of where the underlying indices should be trading were the underlying markets open to trade. ""

    This is one of the factors in determing opening only order pricing, using the futures markets" (SPM stops at 9:15, we use the E's until 9:30).

    Don :confused:
    #18     Jun 2, 2004
  9. damir00

    damir00 Guest

    that matches some gap studies i did on NDX/NQ. fading any and all gaps and closing at the end of the day was a net wash over the long run. the implication is that - on average - over long periods of time - the closing value on a gap day looks a lot like the opening value. that implies ignoring the gap and fading the gap breakaway or fill might be a better way to go - and indeed, that is what i found.

    i haven't yet had a chance to look over the original poster's data, but i'd like to thank him in advance for sharing like that.
    #19     Jun 2, 2004
  10. It sounds like most of the simulations described in this thread concentrated on just one vehicle (e.g., NQ). There seems to be some disagreement about whether fading the gaps is profitable. My experience says that it is.

    However, my approach seems different than most. Instead of looking at a particular vehicle, I did the following:

    1) Choose the top 20 NASDAQ % gainers @ EOD
    2) Choose the ones with average Vol >= 1M
    3) Short the picks 30 min after the next OPEN
    4) Exit before EOD
    5) Use discretion

    #1 works better on an "up" day on Wall St. #2 provides some liquidity, but it's not vital as long as the stock remains "in play." #3 limits your options b/c your broker probably won't have the stock anyway. Waiting 30 minutes smoothes out the ATR. #4 and #5 require some experience and common sense. It's crude, but it works. I don't see why someone would limit themselves to just one particular symbol.

    I've back-tested this strategy the hard way -- by hand picking these daily and recording the results in Excel for about a month. Does anyone know of a scanner that can determine the historical top % gainers for the universe of stocks? In other words, a scanner that tracks historical 'hot lists'?
    #20     Jun 2, 2004