Reinvesting

Forums › ProRealTime English forum › ProOrder support › Reinvesting

Viewing 6 posts - 1 through 6 (of 6 total)
  • 06/30/2016 at 7:05 PM #10050
    Brage

    Hello!

    Testing a 30-min breakout-system (found in the library) with a reinvest function (n = max(1,2+ROUND((strategyprofit-1000)/1000))). Backtested in the original timeframe 15-min but also tested in 2-min timeframe (with parameters for 30-min high and low adjusted) and I get very different results especially on the gain/loss ratio. Is this because of the limits that PRT only test once per bar? Is the smaller timeframe more similar to a real trading result? When testing on a fixed amount the difference is much less.

    Thanks!

    07/01/2016 at 10:19 AM #10077
    Nicolas

    Hi, yes you point it right. If the conditions are tested more often, then you should expect more precise results.

    At least, that’s the way how backtests are actually performed, as you may know this should be overridden by a new whole backtest engine in the next update.

    07/01/2016 at 3:19 PM #10088
    Brage

    Thanks Nicolas!

    New backtest engine will be interesting. Just another quick question. Do you have any idea why the difference between my backtest results are much bigger with the reinvest function? I can’t figure it out:)

    Have a nice weekend!

    07/01/2016 at 4:27 PM #10093
    Nicolas

    Surely because you reinvest your gain into each new trade?

    07/05/2016 at 10:11 AM #10221
    Brage

    What I meant is why the results are so different between testing on 2 minute and 15 minute timeframe with the reinvest function. Exactly same strategy settings and same time period?

    Don’t now if it it was this question you answered:)

    Thanks!

    07/05/2016 at 11:05 AM #10225
    Nicolas

    I don’t know what the strategy is all about, haven’t seen the code but as I told you already, because the conditions are tested more often, then you should expect more precise results.

    If you are not using any indicators, while there is still no multitimeframe support, you should launch/test your trading strategy in the less timeframe possible, even in seconds ones. This should give you more precise results, because of conditions checked only one time per bar.

  • Author
    Posts
Viewing 6 posts - 1 through 6 (of 6 total)

Find exclusive trading pro-tools on ProRealCode Marketplace

Create your free account now and post your request to benefit from the help of the community
Register or Login