Same script but different results in demo and live IG CFD account

Forums ProRealTime English forum General trading discussions Same script but different results in demo and live IG CFD account

Viewing 11 posts - 1 through 11 (of 11 total)
  • #236358

    Hello everyone,

    On my IG CFD demo account, I’ve developed trading strategies for US Tech 100 ($20) contracts. However, when I apply the same script to my IG CFD live account, the results are different, sometimes significantly so. All settings, including time periods, are identical for both accounts.

    Does anyone have experience with this issue and any suggestions on what I should do? My theory is that historical data might not be perfectly identical, despite them supposed to be. I don’t feel confident starting trading until I can trust the results.

    #236361

    Your “significantly so” will emerge because of a different flow of trades (like more/less trades, or them entering at different times). Am I right ?
    If Yes, then this is the most normal thing in itself and “significantly” bears no value.

    The point would be why the trades enter at different times (thus that assumed) with the most profound reason that the previous trade exited earlier of later. And that assumed as well :

    This would be the most normal because in Demo an order will always be filled quite “instantly”, which in Live may go very different. Also, the – what I call – manipulation of IG with the trades (like a very different slippage from time to time) won’t be present in Demo. Thus, in this case the data is the same in both, but the processing of the orders goes different in Live. Especially with “News” (like Fed) and high volatility, things may not go as you want. In Demo you would not notice a thing of this.

    Ask away when still in doubt. 🙂

    1 user thanked author for this post.
    #236362

    It be useful for us to help you if you showed equity curve and positions for Demo and same for Live.

    Did you have ‘Tick by Tick Mode’ enabled when you used Demo?

    #236365

    Thank you for the feedback.

    The strategies are on a daily resolution, with signals based on the closing price and entry/exit at the next day’s open. Typically, positions are held for several days or weeks. The strategies are optimized, but not very aggressively because I don’t want a high risk of curve fitting.

    There is a 5-hour difference in when the positions are taken (even though I use the same time settings in the chart), which explains some of the difference. But the discrepancies are still quite significant.

    I did not use ‘Tick by Tick Mode’; perhaps I should?

    Attached is a document with some charts (real-time first) and the script for one of the strategies, please feel free to test the script on your own computer .

     

    #236367

    I did not use ‘Tick by Tick Mode’; perhaps I should?

    Yes without Tick by Tick mode enabled the results are meaningless.
    Enable Tick by Tick and you may have a shock?

     

    #236375

    But … are we talking Demo or are we talking Backtesting ?!?
    “Demo” is in the description. And with Demo “testing” there is no such thing as Tick by Tick mode. Or ?

    If this is about Backtesting, all is moot so far. So what is it ? Demo or Backtesting (or even Backtesting in Demo) ?

    #236376

    Peter, as per our ‘Agreement’ (re descriptors on another Thread) I assumed … Demo backtest.

    #236377

    Yea, but our agreements are strict secret. Haha.
    Anyway, dubbzie will let us know.

    And dubbzie, if Backtesting in Demo indeed, why would you do that ? (I’m always ready to learn)

     

    #236381

    I use ‘Demo backtest’ (mostly after optimising, but also a backtest on different chart timeframes) because then it’s easy to switch the backtested System to Forward Test on that Demo Account under ProOrder AutoTrading.

    Also I use Demo backtest when I want to re-optimise or amend code of a System that has been under Demo Forward Test (on ProOrder AutoTrading) … it’s easier as it is all kept on the same Demo Account.

    I do also optimise and backtest on my Real Live Account, esp when I want to keep my eye on a Real Live running System.

    #236382

    My working environment is actually always the demo account. I develop everything in demo, subject it to backtests in demo and start the system in demo to see whether it works as it should in “demo-live”. Only then do I transfer the system to the live account, subject it to further backtests there and then start it “live-live”.

    #236384

    Ah, totally clear.
    Would a Backtest in the Live account/environment always be the same as in the Demo account/environment ?

    My perception of things is a little different because a. I deem testing a system in Demo/Paper useless because of the differences incurred for by the broker presence, and b. because within PRT-IB I never could test Forex because the PRT system for Paper is different and the code can’t be the same (this is related to contract sizes).
    So my Backtesting is always on Live (for PRT-IB as well as PRT-IG) and testing a system in Paper (IB) or Demo (IG) I just never do. This is not how it should be of course – it’s just me. 🙂

Viewing 11 posts - 1 through 11 (of 11 total)

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