Backtesting is useless

Discussion in 'Options' started by JCDST1979, Jun 30, 2020.

  1. userque

    userque

    Backtesting and "finding a profitable strategy" are two (or more, depending) separate operations.

    Backtesting a bad strategy won't help find a profitable strategy (except to the extent it tells you to pick another strategy to backtest).

    Backtesting a good strategy may make the strategy more profitable, or spin off different parameter sets that are geared toward different/specific types of market conditions.
     
    #61     Jul 1, 2020
  2. taowave

    taowave

    This thread is becoming the theatre of the absurd..

    Carry on..
     
    #62     Jul 1, 2020
  3. destriero

    destriero


    No, you're fine.

    Ask yourself how you arrived at ET. Was it via any of the following?

    -WSB Discord
    -Reddit
    -RH

    If so, then yes, you're stupid/ignorant/both.
     
    #63     Jul 1, 2020
  4. ironchef

    ironchef

    #64     Jul 2, 2020
  5. JCDST1979

    JCDST1979

    Maybe, but if you adjust your parameters to make it more profitable, isn't that a biased test?.

    How can you make those adjustment to your strategy if you don't know what the next market would be?
     
    #65     Jul 3, 2020
  6. JCDST1979

    JCDST1979

    Money management rules.
     
    #66     Jul 3, 2020
    SimpleMeLike likes this.
  7. JCDST1979

    JCDST1979

    How did you find proof that the markets are not random?

    Can you give an example of one of those patterns?
     
    #67     Jul 3, 2020
  8. virtusa

    virtusa


    Money management rules are applied to a system or strategy. Money management on its own does not generate money. It only helps in finding out how to protect the money invested.
    You first need a strategy and then apply money management rules on it. The rules can be different for different strategies, depending on the "behavior" of the strategy.
     
    #68     Jul 3, 2020
    777, d08 and _terminus_ like this.
  9. JCDST1979

    JCDST1979

    Indeed you need a strategy but you can't know for sure that it can make money by backtesting it. Are you familiar with the survival bias?
     
    #69     Jul 3, 2020
  10. virtusa

    virtusa

    I survive already 2 decades as a daytrader. So I have a fairly good idea about the "survival bias".

    Thanks to years of backtesting I have a fairly good idea about the expectancy of my trading and I know that I will always make money if the number of trades are statistically relevant.

    The problem is not the backtesting, the problem is the (bad/weak) strategy that is backtested.
    If backtesting does not give you comparable results as real trading, you should make a better strategy. Don't blame it on the backtest. You are the problem, as you created the strategy.

    A good strategy will in real trading give more or less the same results as in backtesting. The reason why is very simple: A good strategy works in ANY market condition.
    If your backtest gives different results it means it only works in certain conditions, so not reliable.
     
    #70     Jul 3, 2020
    Laissez Faire likes this.