Tickstory

Hi Tickstory,

I have exported the required tick data but I wish to test with control points for some basic optimizations. I read from a previous thread that 'The Tickstory launch process suppresses the generation of the FXT files which are used for the actual back-test in the selected modelling type.' My understanding is that this is only important for every tick testing so as to prevent MT4 from using its own fake ticks from M1 data. For control points, MT4 will have to generate its own FXT file in any case because tickstory only exports tick data, so it does not matter whether I launch via tickstory or not. Is this correct?

Secondly, for control point testing, will the data be based off the accurate Dukascopy data exported from tickstory? Or is it based on data downloaded from my own broker?

Hi Huangwh88,

You are correct - when you select the 'Control point' modelling method in the MT4 Strategy Tester, MT4 will automatically generate the FXT data from the historical chart data (HST files). This means that you will continue to benefit from the Tickstory data you have exported and sourced from Dukascopy. It is not necessary to use the Tickstory launcher to work with the Control Point modelling method.

Regards.

tickstory wrote:Hi Huangwh88,

You are correct - when you select the 'Control point' modelling method in the MT4 Strategy Tester, MT4 will automatically generate the FXT data from the historical chart data (HST files). This means that you will continue to benefit from the Tickstory data you have exported and sourced from Dukascopy. It is not necessary to use the Tickstory launcher to work with the Control Point modelling method.

Regards.

thanks for the great reply as always

tickstory wrote:Hi Huangwh88,

You are correct - when you select the 'Control point' modelling method in the MT4 Strategy Tester, MT4 will automatically generate the FXT data from the historical chart data (HST files). This means that you will continue to benefit from the Tickstory data you have exported and sourced from Dukascopy. It is not necessary to use the Tickstory launcher to work with the Control Point modelling method.

Regards.

hi Tickstory,

Just want to verify, say if I do H4 control point testing, MT4 will generate the FXT file containing H1 data (next lower timeframe) using the tick data exported from tickstory? Am concerned because control point testing seems to miss far too many trades, compared to tick mode

Hi Huangwh88,

MT4 will generated the FXT file for the time-frame you have selected (in your example, H4). This will contain the individual OHLC control points/ticks which are taken from the HST files that are generated by Tickstory. If you make any reference to a time-frame other than H4 in your code, MT4 will use the HST file to resolve the reference.

Hope this helps.

tickstory wrote:Hi Huangwh88,

MT4 will generated the FXT file for the time-frame you have selected (in your example, H4). This will contain the individual OHLC control points/ticks which are taken from the HST files that are generated by Tickstory. If you make any reference to a time-frame other than H4 in your code, MT4 will use the HST file to resolve the reference.

Hope this helps.

Thank you for replying. So since the HST file contains tick data from tickstory, does MT4 randomly select 12 data points from each period to create each candle in control point mode?

The HST file contains the candle-stick data for each timeframe. The HST file for M5, for example, will contain the OHLC information for all the 5-minute bars. When back-testing with the control point modelling method in M5 period, MT4 will construct an FXT file with the OHLC points of each bar (i.e. it will generate 4 "ticks" that it will pass to your EA).

Hope this helps.