“Big difference between positions in backtesting and live trading.

• 04/26/2025 at 4:23 PM #246444
  dubbzie

  Hi everyone,

  I have a strategy for trading the US Tech 100 Euro 1 on daily charts, CFR, developed in ProRealTime. When I tested the strategy with automated trading in ProRealTime (ProOrdern AutoTradfing / broker IG Markets), both the entries and exits were different compared to the backtests, even though the data set was the same. The backtest results are on the left, and the actual trading results are on the right. There is a very big difference. Does anyone have an idea what the reason could be? 

  

  dubbzie

  Participant

  Topics: 8

  Replies: 20

  Been thanked: 5 times
  04/26/2025 at 4:47 PM #246446
  dubbzie

  Sorry, should be CFD not CFR. And a comment regarding the timestamp showing 19:00: I live in Norway and have set the chart to America/New York time.

  

  dubbzie

  Participant

  Topics: 8

  Replies: 20

  Been thanked: 5 times
  04/27/2025 at 11:09 AM #246457
  druby

  Not sure what you referring to with regards to ‘very big difference’.

  The left hand detailed report-order list  states a start and end date. Since it states ‘END’ this means that the back-test didn’t reach the current real time date/time.

  The right hand detailed report-order list states a start and end date different to back-test. Since it states ‘CURRENT’ this is up to current real time date and time, but started after back-test finished.

  Also, the strategy blue link I think states the strategy start date/time, again after the back-test ended.

  So this could mean the back-test doesn’t include the 24-25th trade.

  The difference between 22nd exit at 18707.1 and the 23rd entry at 18700.6 is 6.4 , which isn’t that much different, assuming that the other back-test trade were included in the strategy list, out of view.

  Scrutinising the lists of orders, may show were they deviated and maybe you could derive the cause.

  I think in general, back-test result are not forced to be the same when a strategy is run in demo, so don’t expect to be the same in live.

  The spread usually makes a difference, in back-test its estimated, if set, however this could change live as well as the time a order is actually executed.

  Anyway , that’s all I’ve got for now, I may be barking up the wrong tree, if that’s the case, be more specific so to narrow down.

  Regards

   

  3 users thanked author for this post.

  Iván, dubbzie, GraHal

  

  druby

  Participant

  Topics: 1

  Replies: 446

  Been thanked: 369 times
  04/27/2025 at 6:26 PM #246473
  dubbzie

  Thank you for the response, GraHal.

  Before I startet automatic trading, I checked if the strategy had a position. It didn’t according to the backtest I ran. Could it be that it actually had a position, but the backtest didn’t catch it right away? If not, maybe it’s a human error I made resulating in taking position 1 day late, and that’s the whole explanation?

  By the way, do you know if the time zone I selected plays any role in this context, or if the system always uses New York time. And what I choose for time zone only affects what’s shown on the chart and the timestamps in the position overview? 19:00 as shown is Norwegian time, and before I selected that, it used to show 01:00 as the time.

  

  dubbzie

  Participant

  Topics: 8

  Replies: 20

  Been thanked: 5 times
  04/27/2025 at 7:17 PM #246474
  GraHal

  dubbzie wrote:

  Thank you for the response, GraHal.

  Thanks need to go to druby (not me) as it is druby who replied above.

  dubbzie wrote:

  resulating in taking position 1 day late

  Have you got realtime data or are you on ‘delayed data’?

  Who is your broker and who provided your PRT Platform?

   

  

  GraHal

  Participant

  Topics: 219

  Replies: 8983

  Been thanked: 2278 times
• Author
  Posts