As you correctly point out, the 'C-O' signal has inherent lookahead bias, allowing the trading of precise relative ETF quantities at the opening auction - without actually knowing the opening price.
Were you able to check strategy robustness using either a pre-market or post-open price such as 9:25am or 9:35am? I'd love to know, as the daily returns are very stackable in aggregate, but razor thin individually.
Thanks for sharing!
One question: when trading at the open, are you only looking at two of the prior three sessions?
Eg: If it is Wednesday morning right now, you can look back at
1) Mon close-to-Tue open
2)Tue open-to-Tue close
3) Tue close-to-Wed open
On Wed morning open, is my understanding correct that you are looking only at #1 and #2?
Thanks.
Thanks for reaching out!
If it’s Wednesday morning then the return we are looking at for the overnight signal is Tuesday Close to Wednesday morning.
Understood. Thanks!
Super interesting article thanks!
As you correctly point out, the 'C-O' signal has inherent lookahead bias, allowing the trading of precise relative ETF quantities at the opening auction - without actually knowing the opening price.
Were you able to check strategy robustness using either a pre-market or post-open price such as 9:25am or 9:35am? I'd love to know, as the daily returns are very stackable in aggregate, but razor thin individually.