1 (edited by vidagig_13 2015-07-03 12:14:08)

Topic: Old good strategy retests bad in FSB, why?

Hi there

A strategy created on 2015-03-02 tested on FSB on 15-03-04 using my broker data from 2014-11-18 to 2015-02-26 using 80.0 points spread, had Annualized Profit 886.46% and Max Drawdown 19.52% on 36 trades.

When I rerun the test now, using the same broker's data on the same Data Horizon I get Annualized Profit -215.66% and Max Drawdown 81.50% on 47 trades.

When I rerun a backtest on MT4 (every tick f.w.i.w...) I get the identical good result as the backtest done on 15-03-03.

Why the difference?

It uses an hourly and a 5 minute CCI and trailing stop limit, trading 50% of the account balance.

I got an 'error bug report sent message' when re-opening the strategy yesterday, today it opens fine.

I remade the strategy with the same settings - identical poor result.

It retests equally poorly on 5 different data sets I have on FSB.

Just trying to understand if some FSB updates may have corrected things e.g. indicators or backtest method - that I was unknowingly exploiting in the strategy or something; and if an old strategy now retests poorly on the same data horizon, I need to discard the strategy as no longer being valid.

Perhaps FSB has become stricter in its requirements of a valid strategy.

Any input much appreciated.

Thank you.

Re: Old good strategy retests bad in FSB, why?

There was a bug with the LTF (longer time frame) indicators - incorrect "Use previous bar value" option. As a result, some strategies using LTF showed unrealistically higher result from the backtest. This bug was fixed on the beginning of May 2015.

If you reexport your EA with the newest FSB Pro version it must show  equal results in both FSB and MT testers. If the discrepancy still persists, please report.

3 (edited by vidagig_13 2015-07-03 16:32:55)

Re: Old good strategy retests bad in FSB, why?

Great, now I understand.
Thank you very much for your help to resolve this query.