My understanding of Market Replay is that it is literally a playback of actual market tick data that is recorded by user or retrieved from NT servers. I also assume that there is no compression or bundling (can't remember the nomenclature) of data into OHLC, 1-minute, or other larger-than-tick increments that might alter the way a chart is formed or strategy performs.
If so, I would also assume that if one ignores the non-ninja factors such as slippage, broker execution, etc, that the Market Replay data provides a tick by tick data stream with each tick retaining its original sequence, thereby eliminating problems like the "time machine" effect and other platform related problems associated with backtesting charts built on historical data. Is this correct?
My quest is to get an accurate (systemic, not real-world) way to test and gather statistics that would trigger stops, targets and other tick-by-tick events correctly. Including any chart type (renko, range, etc).
Again, I fully understand the non-ninja issues that exist when comparing any test to live trading, but I need to establish whether or not I am testing against an unfettered market data flow that is not manipulated by the testing platform in a way that could alter results. (also ignore the possibility of isolated technical problems with quality of the recorded data)
Sorry to overstate the question, but I recognize that you need to be careful to hedge your answer to avoid making any claims that the market replay is the same as live market conditions. So even trying to ask the question correctly can tend to drift off point.
I am experiencing an inarticulate moment here, so please let me know if I need to clarify or re-state the question.
Thanks.
Comment