Thank you for sharing your idea Michael. I can confirm that more users are saving results of every robustness test into separate folder for recovery purpose.
In the second OOS robustness test we are using for filtering Ranking&Filtering – Custom filters. You can choose if strategies which not pass these conditions will be deleted immediately or you will delete them manually in the databank.
Average slippage for ES Futures market is 1.5 tick. For slippage test you can use for example 2 ticks.
You can install MT4 into separate folders and then connect them to the one account. The only restriction from your broker might be number of simultaneously opened pending orders. Maybe discuss this issue directly with your broker’s support.
It will be probably GMT2 The best would be to contact Oanda support and ask them which exactly UTC timezone the broker is.
Which city is located your broker? DataManager allows to clone data including DST shift.
I suggest to use whole data sample which you have available.
You can check this tutorial about data: https://www.appdemostore.com/demo?id=6129955890003968
Data Manager in SQ don’t have tick data import. The only way is to export them as CSV from TickDataSuite and then import them into Data Manager.
Point value change: You can change instrument definition in the DataManager „Instruments“ tab. Point value for non USD symbols is changing over time but this affect only size of profit, Equity, Ret/DD ratio remains the same.
Here you can find table of current point values for Dukascopy symbols: https://pages.strategyquant.com/strategyquant-data-setting/
Adding new symbol definition: If you in DataManager click on „Dukascopy Data“ – „Add new symbol“ then you can select the symbol with defined setting.
there are two groups of users. First group is strictly following suggested parameters and the second one is treating this test as confirmation of robustness. So there is no exact answer for this. I am also adding attachment which discuss how to easily work with this parameters and export them.
- 8.1.2019 v 3:03 pm odpověď na: 02 Monitoring of strategies using FX Blue – part one #36636
1) Generally you could step-by-step try lower values until StrategyQuant start to generate strategies. Good practice is to obtain 20-30 trades per year for H4 timeframe and 5-15 for D1 timeframe.
2) You could test workflow where you will use less trades for verifying the strategy. Obviously more trades – better.
You can use IS/OOS but then you face mentioned problem with low number of trades. To obtain longer OOS you can start using 50% IS and 50% OOS. There is no the only right solution.
For D1 strategies high importance have test on a different markets. Logic of strategy should be robust enough to work well also on other similar markets. Also you can tray random generation which tends to produce less over-optimized strategies.