I 'm referring to trade that happened at 10.01.2017 23:00.
Description changed:
I 'm referring to trade that happened at 10.01.2017 23:00.
Attachment 4.png added
Here is the data and settings
https://www.dropbox.com/s/d1txbk8znicllnq/btcusd.csv?dl=1Status changed from Waiting for information to In progress
Status changed from In progress to Refused
Bid price at 10.01.2017 23:00 is 911.0, spread is 8, so Ask price is 911 + 8 = 919.
You see Bid prices on chart, not Ask prices.
I think your definition of instrument is wrong. Spear id in pips, so if you have Pip/Tick size = 1, then spread=8 means 8 full points. That doesn';t seem correct to me.
And Pip/Tick Step is also incorrrect I think, I can see that the minimal move in your data file is 0.01, and this should be your Pip/Tick step.
My exchange uses 5 significant digits size for bitcoin and there no such thing as step.
It looks like this as price grows:
1.1111 - min movement is 0.0001
11.111 - min movement is 0.001
111.11 - min movement is 0.01
1111.1 - min movement is 0.1
11111 - min movement is 1
I can't backtest old data (before 2016) because tick size between old data and new data is too different and strategy quant uses static instrument size.
I have chosen this sizer 1111.1 - min movement is 0.1 as the most relevant.
I expected spread to be in steps instead of sizes.
Spread is in pips, not in minimal steps
Attachment 5.png added
Attachment 5.png deleted
Attachment 5.png added
This is my new settings.
Does it look ok for you?
Because it is combined data.
So from my point of view :
bid should be imported - spread / 2
ask should be imported + spread / 2
Attachment 5.png added
And to use it only in single spread robustness test.
I would like to hear you feedback about this.
The way it is implemented right now seems wrong to me.
Because it can cause such impossible trades.
Hi CoolerNax,
I tried to retest your strategy on BTCUSD data downloaded from coinbase. I used exactly the same settings, but got different results.Can you attach your history data or specify which data sources you used?
Best regards,
Tomas