I've been doing some testing on various automated systems and have been encountering a high level of discrepancy between the results of the backtest and market replay.
I am using MarketReplay.net for my replay data and Kinetick for my historical data using continuous contracts for markets such as ES ##-## and CL ##-##.
There appears to be a 2 hour shift in the data when using the global crosshair, I'm not sure if that is the cause of the problem or if the data is simply that much different. Even when running a very simple test strategy such as an SMA cross, there are still significant performance variations.
So I guess my real question is to which data and which results would I expect to be accurate?
Thanks!
Comment