S&P 500 Daytrade Sys: Opening Range Breakout

Discussion in 'Strategy Development' started by ewile, Mar 7, 2004.

  1. ewile

    ewile

  2. Sometimes you take 'em and sometimes you fade 'em -- and therein lies the rub. :D

    Okay, I'm feeling generous tonight. Use FVE or Chaikin.
     
  3. dbphoenix

    dbphoenix

    The narrower the range, the more difficult the ORB strategy. You may want to look at reversal strategies instead.
     
  4. I've tried to back-test this with the spy, qqq, and other etfs as well as with individual stocks and indexes

    I spent months trying to develop a strategy based on this well
    trodden idea but could not profitably implement it

    It could well be that my testing was flawed and that incorporating the S&P futures (or eminis) into the strategy would change things. Either way, I found this to be a tough one both to back-test and to trade.

    As Vic Neiderhoffer put it, "any testable statement should be tested"
     
  5. I coded this into Excel when Tripack first posted the link, and monitored the method - empirically - for about 6 months. During that time, it did well on strong trend days, and not as well on weaker days. That pretty much puts it in the same category as most range break methods.

    I have quoted below what I feel are the key points of that method.

    Earlier dbphoenix made an excellent point about our present market. The trend days are not as frequent as one would hope for this method to be successful. Narrow ranges, no follow thru, etc., etc.

    But as the first sentence of the quote... and the published method suggests, there is value in understanding the dynamics of the method, and how they are trying to accomplish the goal.
     
  6. mind

    mind

    some analysis on 30 minute data on the sp future since 1995:

    average firstBar:
    -0.0049%
    average dailyGain (Close/Open-1):
    0.0318%
    average GainAfterFirstBar:
    0.0377%

    number of days analysed: 2275

    before i did this i always thought that the market will on average make more in the very first bar. this is not true. [btw i just did that for the first time, maybe i have a mistake, but i do not think so].

    the thing becomes interesting once the result of the first bar is used as a filter. if the first bar is significantly positive, the rest of the day will be bullish as well, if the first bar is significantly negative the rest of the day will be too. the statistics seem to be robust.

    a while ago i looked at narrow range bars and inside bars. i did not find any evidence that they lead to tradeable non random movements the next day. even more narrow range days are a good forecast for ... another narrow range day.

    range as such forecasts itself. if we saw high range yesterday we have higher-than-random-chance to see big range today. this is already intuitive, but it is interesting that recent high range does not only shift expectations for range but for dailyGain as well upwards.

    the average range of the firstBar was 0.59%, the average movement thereafter (absolute number of dailyGainAfterFirstBar) was 0.92%. if the range of the firstBar was below average the movement thereafter reduces to 0.69%, if the firstBar range was high the average movement thereafter increased to 1.27%. this means that a day that starts volatile will show more follow through in terms of percentage until the close.

    just tried to add some numbers to the discussion.


    peace
     
  7. Excellent post.

    I trade stocks, not the futures and I'm wondering if your statistics apply to the spy (S&P 500 ETF) or the qqq also. Intuitively I would think yes. You have inspired me to look into this, I'll post whatever I find (if the project does'nt end up being too cumbersome as sometimes they become).
     
  8. mmillar

    mmillar

    It works much better on the Nasdaq then the S&P!!
     
  9. MIND: Nice work and interesting numbers.

    After reading about the work of Crabel, I decided to start looking at the futures intraday market a little differently than I first did. I started to collect data based on the time frame of 0930 ET to 1600 ET. I then looked at data relative to the 0930 ET open.

    Typically, the intraday market is skewed. In other words, the 0930 ET open will be closer to one extreme of the daily range than the other. I do not use the data as a strict mechanical trading system, but rather use it to help me know what my potential risk and reward may be intraday.

    For the March 2004 contracts, the average distance from the 0930 ET opening to each daily range extreme is as follows:

    ES Contract: 2.25 and 8.25 points.
    NQ Contract: 8 and 18 points.

    I also calculate the 10 day average for these numbers. These numbers not only give me a feel for risk and reward potential, but also help locate potential reversal points in the intraday range.

    Charles
     
    #10     Mar 14, 2004