At the moment when building and testing strategies, there are two spread choices.
Can I please suggest a 3rd option, which is when the bars are created for the test precision (e.g. selected timeframe or 1 minute bar precision), that the average spread for each period is used. This is only available when the source of the data is tick.
So for example, if 1 minute tick simulation precision is selected for the test, and the source is tick data, then for every 1 minute bar created internally, the average spread is used based on the 1 minute of tick data.
This means spreads will be more realistic at this precision. An additional spread offset and multiplier would also be useful to adjust the tick spread to more closely match the broker spread.
So, to summarise, split the tick precision from the applied spread and have:
Fixed
Real (from Tick Source)
as the spread options. If the user selects "Real (from Tick Source)" have an extra "Spread Addition" and "Spread Multiplier" setting.
So, in this instance, the user can select "1 Minute data Tick Simulation" and also select "Real Spread from Tick Data" and then enter "1.5" for "Spread Multiplier" and 0.5 for "Spread Addition", which would for every 1 minute bar created internally, calculate the average tick spread during that 1 minute bar, multiply this spread by 1.5 and add 0.5 pips to the spread.
Hope this makes sense.
Thanks,
Mike
PS Same would apply for the 1 second data Tick Simulation which I'm hoping you will add to the test precision too.
Assignee was changed
Milestone changed from None to Release Candidate 3