Still looking for your feedback/response on my review of 2 of your articles and the discrepancies..I had posted this earlier..
“Hi .. could u kindly explain why your 10 SPY flips shown in article dated 9/19 is different from the one shown in the article dated August 1. The returns/some dates shown do not match. Creates the impression of the accuracy of the data . Thank you again”
Deepam, thanks again for this question. This is an excellent catch, and I completely missed the nuance until you pointed it out.
First, the last 10 flip dates are consistent in every single post; that's the most important part of the signal. From a practical execution point of view, the signals trigger after close, and I personally place trades at the next day's open, so there's always a one day execution lag between the signal and my real fills.
What caused the discrepancy was a data column mix up in my program: the 8/1 table was built from "Adjusted Close" (which restates history slightly every time a dividend is paid), while every issue since 8/8 has used the plain "Close," the actual traded price.
Since 8/8 and going forward, I'm standardizing on actual Close for all return calculations.
Thanks again for catching this. Appreciate the scrutiny.
Thank you for update..
Still looking for your feedback/response on my review of 2 of your articles and the discrepancies..I had posted this earlier..
“Hi .. could u kindly explain why your 10 SPY flips shown in article dated 9/19 is different from the one shown in the article dated August 1. The returns/some dates shown do not match. Creates the impression of the accuracy of the data . Thank you again”
Hi, thanks for pointing it out. Let me check. We are probably in different time zones in the world and my day just started.
Sure .. whenever you have the time to review and provide an update. 🙏🏽
Deepam, thanks again for this question. This is an excellent catch, and I completely missed the nuance until you pointed it out.
First, the last 10 flip dates are consistent in every single post; that's the most important part of the signal. From a practical execution point of view, the signals trigger after close, and I personally place trades at the next day's open, so there's always a one day execution lag between the signal and my real fills.
What caused the discrepancy was a data column mix up in my program: the 8/1 table was built from "Adjusted Close" (which restates history slightly every time a dividend is paid), while every issue since 8/8 has used the plain "Close," the actual traded price.
Since 8/8 and going forward, I'm standardizing on actual Close for all return calculations.
Thanks again for catching this. Appreciate the scrutiny.