Different Time Frames - Different Results

Discussion in 'Strategy Development' started by Burtakus, Dec 28, 2005.

  1. For those who have experience, if a strategy (entry, exit, and MM criteria) is successfull in backtesting in one time frame, should it be successfull in another time frame?

    I ask because I was doing some backtesting and found that the entry, exit, and MM criteria in the 1-hr time frame provided reasonable returns but when I put it to the 4-hr time frame it showed a loss. From ya'lls experiance, what if anything should I read into it. Obviously use the 1-hr and not 4-hr time frames but am I missing some basic concept somewhere?
  2. i would say it depends on the strategy you are backtesting. Some patterns are more dependable/probable in different time frames on different securities. Lots of variables involved....time frame is just another one of them.

  3. It should if you are using a trend following method and not strictly daytrading.
  4. If you can't reasonably expect a backtested strategy to work in REAL trading, why should you expect it to work in a different time frame?
  5. mogul


    I've found that each time-frame has different characteristics. You can see this by looking at the flavor of the charts. Some time-frames trend much better than others

    This is applicable both up and down, i.e. some instruments display much 'cleaner' waves on smaller time-frames than on larger, and vice versa
  6. The success depends of the range between the highest high and the lowest low in a timeframe. The wider the range the easier it will be to be profitable because there is much more room.
    Profitability increases when timeframes get bigger. The monthly range is always wider than the weekly range, the weekly range is always wider than the daily range etc etc...
  7. Thanks for ya'lls insight. It has been helpful
  8. cnms2


    Un caveat is that the price range increases at a lower rate than the time frame. It means that the reduction in opportunity is not compensated by the increase in price range. Also, you'll get steeper trends in lower time frames.

    With backtesting you have to check if inadvertently your results didn't benefit from lucky conjectures (curb fitting), and if so to eliminate them from the overall result.
  9. Since we don't know what type of strategy you've been backtesting...

    What RunTrade said is true.

    Some strategies are profitable regardless to the time frame but will have less trade signals as the interval increases.

    Also, although profitable, x profits could be dependent upon the trade management after entry, risk tolerance levels et cetera.

    Thus, it can get very complex using the exact same strategy on any time frame.

    It's that reason alone why some traders use one particular strategy for day trading on small time frame and a completely different strategy for swing trading on a larger time frame.

    (a.k.a. NihabaAshi) Japanese Candlestick term
  10. I took the december contract ES and calculated the range between the high and the low in different timeframes:
    weekly data: average range was 26.15 points
    daily data: average range was 12.04 points
    hourly data: average range was 2.03 points
    15 minutes data: average range was 1.04 points

    What i wanted to explain is that trades have more chances to be profitable in bigger timeframes than in smaller ones. This doesn't mean that the biggest timeframe is the best one. By increasing the timeframe the additional efficiency will, at a certain point, reverse in inefficiency. It's a matter of finding the optimal balance between timeframes and potential profits. But this will mostly depend on the specifics of the system that is traded.
    #10     Dec 29, 2005