On backtest we should trust

Discussion in 'Automated Trading' started by Prospergain, Dec 13, 2024.

  1. I did a, one-year simulation of my strategy to compare it with the results of my LIVE account to prove that a 16-year backtest is more reliable than a 1-2-3 years Live account.
    What would YOU trust?
    A backtest or a few years Live account?
     
  2. In God we trust
     
    faermark and EdgeHunter like this.
  3. even if true your observation is an N = 1 and we can place no reliance on a single observation. What we can rely upon is the active back testing in ninja trader is notoriously unreliable as shown below:
    https://ninjatrader.com/support/helpGuides/nt8/NT HelpGuide English.html?discrepancies_real-time_vs_bac.htm

    Trade station is also prone the corruption based on back testing ("Backtesting relies on historical data, and the quality and accuracy of the data used can have a significant impact on the results. Data may contain errors, gaps, or other inconsistencies, which can distort the backtest results and lead to inaccurate conclusions about the strategy's performance").

    Forward walk real time testing reigns supreme always. It is reality.
     
    Last edited: Dec 13, 2024
  4. maxinger

    maxinger

    I don't trust your 1000 years of backtesting.

    Ultimately, You have to do live trading in order to earn $$$$ (if the backtesting is reliable).
    But most people ended up having to top up their trading accounts.

    do note that you didn't factor in things like
    latency problem,
    bid-offer spread problem, slippage ....



    On backtest we should trust
    ----->
    On backtest we should distrust
    or

    On live trading, we trust
     
    MarkBrown likes this.
  5. MarkBrown

    MarkBrown

    can't trust backtesting alone need to match real time or big fail
     
    maxinger and Prospergain like this.
  6. Yes... Written On Every Dollar... :cool:
     
  7. Bad_Badness

    Bad_Badness

    Backtesting can show validity but often is un-sound. ( because of faulty data, inconsistent or magic fills)

    Only live can show validity and soundness. The two are not the same.

    What you choose to "trust" is up to you. These days most people don't know the difference. They think a validity = truth.
     
    McKoL and MarkBrown like this.
  8. Handle123

    Handle123

    Backtesting is only part of trading plan, should do 6-12 months of real time forward testing, and above all keeping STATS.

    NEVER DO optimizing cause the best past is not the future.
     
  9. Sekiyo

    Sekiyo

    But the truth is no unicorns :(
    Therefore I do trust the fairytales !
     
  10. Sekiyo

    Sekiyo

    Fake news
     
    #10     Dec 13, 2024