Because different teams have different caliberate period, for example, if the total backtest period is 2001-2002, some team has a longer caliberate period (3 month) then the corresponding trading period will be (4/2001-2002), others have 3days caliberate period, which hardly have any significant impact on the overall period. Would this difference result in our different scores? Or will the platform adjust this for us?
Hello - you need to work with the data you have and make sure your model can work for any kind of data (i.e. your data is not hard coded). I would suggest you should refer to examples that have been put as reference. In particular it is recommended you use the dataFeatures class for data, you use zoom / period argument for rolling data and you don’t have dates on the code. Make your code able to generalize, check : https://qlib.alphien.com/qlibv4/RFunctionEditor?open_functions=momentumSP500Select (you have example there).
Thanks.