Hi,
there seem to be plans to end the concept of daylight saving time changes in the US. This would effect the way historical data is handled when backtesting.
Will MC get additional settings like: use GMT offset before/after daylight saving concept...
Chris
(If) Daylight Saving concept ends... Backtests?
Re: (If) Daylight Saving concept ends... Backtests?
Just make sure you use UTC data, it's the only real and permanent solution for professional purposes.
Re: (If) Daylight Saving concept ends... Backtests?
@c0ntango: I need to think that through LOL But first thing that comes to my mind, does`nt that just shift the problems of backtesting in context of time-changes, to a whole lot of considerrations to the code?
Ok, thats the only part I can fully controll, but it does make things way more complicated...
... I quess you`re right, even though I yet cant grasp the full extent of changes that would need to be done. I was/am hoping, that MC will compensate my lack of professional programming skills with a solution within the software
Ok, thats the only part I can fully controll, but it does make things way more complicated...
... I quess you`re right, even though I yet cant grasp the full extent of changes that would need to be done. I was/am hoping, that MC will compensate my lack of professional programming skills with a solution within the software
Re: (If) Daylight Saving concept ends... Backtests?
This is not a programming problem and has nothing to do with MC.
Re: (If) Daylight Saving concept ends... Backtests?
then explain please.
in my book, if I have a signal at 8oclock I dont have to think about time changes when doing a backtest, due to the built in settings for daylight saving. without those setting it meeds to be done in the code.
if that whole concept is dropped by the senat mext year, daylight saving settings like they are now is not enough, because there is before amd after daylight saving concept when doing backtest.
so how is that no programming/mc issue?
in my book, if I have a signal at 8oclock I dont have to think about time changes when doing a backtest, due to the built in settings for daylight saving. without those setting it meeds to be done in the code.
if that whole concept is dropped by the senat mext year, daylight saving settings like they are now is not enough, because there is before amd after daylight saving concept when doing backtest.
so how is that no programming/mc issue?
Re: (If) Daylight Saving concept ends... Backtests?
What I meant was - it's a problem with you sourcing the right data, not really a programming problem. Make sure the data you're analyzing is in UTC, and then set MC to UTC. No daylight saving issues. Hope that helps.