Hi, I have 2 datasets of the same 50 stocks from 2 different sources from 2003 to 2022 (till date). I am backtesting both these datasets with the same parameters and conditions but I get different PnL at the end of the backtest. What are some of the common reasons why this may happen?
1 Like
There maybe some discrepancy in the dataset or the way it is implemented. Other than that i can’t think of anything else.
I am going through a few things wrt to datasets from both sources and I see there are discrepancies in the dates from both sources. 35 dates from each dataset are missing in the other. Since I have reindexed the data with one of the stocks index, this will be consistent throughout my dataset it seems.
Which dataset should I believe in this case? 
I guess you can go with the reputed source, and if both the sources are reputed, you can go with the one which has more history, which is used by many people, you can check reviews of what people say about the soruce and take a decision.
Thanks for the input 