Since SkyzerFX brought up the summer GMT offset switch, now may be a good time to ask this question... (I have never dealt with EAs that are GMT sensitive, so I'm not sure the answer to this and how it may affect Geoff's EA, but I think it does)
As an example, I use MBTrading. MBTrading right now is GMT -4 because we are now in day light savings.
When day light savings goes away Nov 7 in the fall MBTrading switches to GMT -5
Daylight Saving Time dates for 2010:
Sunday 14 March 2010 2am local time
until Sunday 7 November 2010 2am local time
Eastern Standard Time (EST) = GMT-5
Eastern Daylight Time (EDT) = GMT-4
This brings up the question for backtesting: How do EAs interpret GMT offset during backtesting? I'm guessing if you set the GMT offset to a specific # and then when the time changes from standard time to daylight time or daylight to standard, then it will be off by one crucial hour. I think the EA may need logic so that backtests take into account this time shift, which may need hardcoded days and times of when the switch happens in 2010, 2009, etc since depending on the year, it is at different times. Any thoughts?
Geoff-> This change I'm sure will give you even more accurate and better backtest results! Who knows, maybe even give better results for previous years that some people have been wondering about.
http://wwp.greenwichmeantime.com/
Thanks,
jd
As an example, I use MBTrading. MBTrading right now is GMT -4 because we are now in day light savings.
When day light savings goes away Nov 7 in the fall MBTrading switches to GMT -5
Daylight Saving Time dates for 2010:
Sunday 14 March 2010 2am local time
until Sunday 7 November 2010 2am local time
Eastern Standard Time (EST) = GMT-5
Eastern Daylight Time (EDT) = GMT-4
This brings up the question for backtesting: How do EAs interpret GMT offset during backtesting? I'm guessing if you set the GMT offset to a specific # and then when the time changes from standard time to daylight time or daylight to standard, then it will be off by one crucial hour. I think the EA may need logic so that backtests take into account this time shift, which may need hardcoded days and times of when the switch happens in 2010, 2009, etc since depending on the year, it is at different times. Any thoughts?
Geoff-> This change I'm sure will give you even more accurate and better backtest results! Who knows, maybe even give better results for previous years that some people have been wondering about.
http://wwp.greenwichmeantime.com/
Thanks,
jd