I found a bug when doing backtesting in TS & Mc.
I usually try to make a strategy with the same name.
I then paste the code I created in SQ.
Because it is fast to do this.
I expect many users will do a lot of strategic backtesting with the same name for convenience.
However, I found that some of the previous backtesting results remain if I do this.
I guess it’s a memory issue with TS and multi charts.
Here is the solution.
Open the strategy editor and clear all the all source code and compile it(F3).
Then paste and recompile your strategy.
I thought this is the problem of SQ so far.
I want other users to know this.
# bug 2 (SQx)
Additional build config
Strategy style Old SQ3 architecture
This is not clear.
But my feeling is that the backtesting results of old sq seem to fit better into trading programs.
#bug 3 (SQX)
There is continuous code bug in Strategy style Old SQ3 architecture.
A bug occurs when compiling on a multi-chart.
Thanks to you giving me the trial code again, I am doing a lot of backtesting.
Regrading your bug1, this is not SQX’s bug, this is TS and MC’s problem.
This is well-known problem for MC users.
Here is a solution for your reference:
When you need to test another strategy with the same strategy name in PLE, remove the strategy from MC chart.
And them copy and paste source code from SQX and compile, insert strategy in MC chart again.
This is real hassle to test mass strategies from SQX to MC, hope SQX is able to export mass source codes with .PLE file format.
You must be logged in to reply to this topic.