Possible bug using iBarshift with Strategy Optimiser
Posted: Mon Sep 26, 2011 12:33 am
Hi,
I have a strategy that works perfectly time after time when a standard backtest is run.
However, if I test it using strategy tester I get different results almost every time. The differences are not the same each time but seem random.
Also, the chart does not update smoothly. It goes blank for several seconds, strange dates appear at the bottom (when testing June for instance, Dec dates will flash on the bottom of the chart).
It also crashes FT a lot, I'd say almost 50% of the time (requiring task manager end process).
I believe this is all linked with the use of the iBarShift function in my strategy. I've not seen it happen with any other strategies.
I've created a cut-down version of my strategy (attached) for you to test yourselves.
Please run it in normal mode and take note of the $ result.
Then run it several (10+) times in Strategy Optimiser with the same parameters and see if you too get differing results. 6 months to a years worth of data should show the issue.
I have tested this a lot and am convinced it is related to the use of iBarshift. Even the simple SMA strategy with an added barshift function crashed on me several times.
I would appreciate you looking into this and if you need any more info please let me know.
Thanks,
Steve
I have a strategy that works perfectly time after time when a standard backtest is run.
However, if I test it using strategy tester I get different results almost every time. The differences are not the same each time but seem random.
Also, the chart does not update smoothly. It goes blank for several seconds, strange dates appear at the bottom (when testing June for instance, Dec dates will flash on the bottom of the chart).
It also crashes FT a lot, I'd say almost 50% of the time (requiring task manager end process).
I believe this is all linked with the use of the iBarShift function in my strategy. I've not seen it happen with any other strategies.
I've created a cut-down version of my strategy (attached) for you to test yourselves.
Please run it in normal mode and take note of the $ result.
Then run it several (10+) times in Strategy Optimiser with the same parameters and see if you too get differing results. 6 months to a years worth of data should show the issue.
I have tested this a lot and am convinced it is related to the use of iBarshift. Even the simple SMA strategy with an added barshift function crashed on me several times.
I would appreciate you looking into this and if you need any more info please let me know.
Thanks,
Steve