Questions from a newbie: Strategies with very similar behavior.
2 replies
Lorenzo
1 year ago #280845
Hello everybody,
I’m new to Strategy Quant X. I started studying it and generating my first strategies.
I’m applying a first simple workflow: once I have obtained a historical series, I take a period for the in-sample and a first period for the out-of-sample. I set money management a spread that is double my broker average, slippage, and ranking based on Ret/DD Ratio = x * years, setting profit factor, and number of trades.
Once I have the first strategies, I run the following tests:
1. I run a second test on a second out-of-sample before the in-sample, adjusting the Ret/DD ratio.
2. I run a robustness test underlining strategies by increasing slippage on the initial build period, including the first out-of-sample time frame.
3. Then I try another poorly correlated market.
4. I randomly run a Monte Carlo test on the parameters (carrying out 200 simulations).
At this point I was left with a few strategies and ran them on a demo account.
I come to the question I wanted to ask you:
I set as money management the exposure for each strategy of 1% of the account balance. However, I observe that the strategies are very similar and practically all are activated almost simultaneously. Is this correct from a money management point of view? Could this very similar behavior depend on the way in which the stress tests are constructed?
I forgot I’m working on EURUSD H1 timeframes. I currently have a small group of 4 strategies, and they enter the market once a week is that a normal frequency considering the times?
Thanks and sorry for the silly newbie questions.
Lorenzo
tomas262
1 year ago #281043
Lorenzo,
you need to make sure strategies are not correlated. If one loses other should have different characteristics which is not an easy task to do when trading a single market.
See the clip attached how to check portfolio correlations in SQX
Lorenzo
1 year ago #281048
Viewing 2 replies - 1 through 2 (of 2 total)