Data Integrity in Optimization

AmiBroker
1

I am doing an optimization test with Equity data of about 2000 symbols using EOD data from Norgate.

I am seeing some peculiar results in a few of the optimizations.

I looked at the ratio of Avg Profit loss / Avg % Profit Loss as a measue of some consistancy. In this test run, the ratio is about 50 except for 4 out of 90 that are 11.9, 11.9, 14.5 and 17.4. Most others are 49 to 50.

The same numbers came up on a re-run, including number of trades
If I multiply Avg Profit x trades, the result is same as Net Profit, as expected, so the Avg % PL seems out of wack

Can not figure out for a reason for such a deviation from the norm of about 50.

Appriciate any feedback.

Thanks

2

Maybe if you provided the formula and screenshots and symbol names, someone could actually test that on their side. Otherwise it is just blue sky talk. Possible answer to your issue (without knowing any details) is position sizing anomalies (i.e. those problematic trades have very small sizes). Please follow this advice: How to ask a good question

2 Likes