The Golden Cross: Daily vs Weekly vs Monthly
18May10
Click for an UPDATED VERSION of this post.
Our apologies for the inconvenience.
Happy Trading,
ms
. . . . .
To stay up to date with what’s happening at the MarketSci Blog, we recommend subscribing to our RSS Feed or Email Feed.
Filed under: Trading Strategies | 7 Comments
Hey Michael, I have also tried various MA crossovers on a monthly basis, based on my style and objective, finally settled on a close above/below the 10-month MA. Would be interested in the results based on your data versus the 2/10; mine data only goes back to the early 1970′s.
First of all, thanks a lot for running this data!!
However, I’ve also been looking at the 10-month SMA and wondered for a while how it compares against the 50/200 SMA crossover as far as number of trades, annual return, and risk.
Also, how is your dividend-adjusted data different in this test compared to your previous similar test (where you did mention in the comments that dividends were included)?
http://marketsci.wordpress.com/2008/09/21/moving-average-crossovers-debunked/
Thanks!
Is the table you’re showing truly showing 1 trade per year on average for all frequencies of trade evaluation, 50/200, 10/40 week and 2/10 month?
RE to G$: yes. I would encourage you to run the #’s yourself if that seems odd. Note that if comparing to CXO’s post, remember CXO ran a close/200-day test rather than a 50/200-day test. michael
I read the original CXO post and like your expansion of it to the golden cross. Since the weekly and monthly tests have much fewer observations, I did wonder in reading both how sensitive the weekly and monthly results are to start date. In other words, would a test starting each day of the month/week yield the same results as the your calc date?
All the best,
Jerry
RE to Jerry: allow me to break your question down a bit…
This test wouldn’t be affected by “start date” because it’s trading on the final day of the week/month regardless of the number of days (and is not for instance trading ever 5 or 21 days).
But I don’t think that’s what you’re asking. I think you’re asking what about if the weekly strategy traded every 1st day or 2nd day of the week for instance rather than every last day (and same for the monthly test).
I would strongly suspect that because the Golden Cross has proven so insensitive to near-term lag (as evidenced by, for example, by the daily vs monthly test) that using some other day of the week/month rather than the final day of the week/month wouldn’t amount to any significant difference. Would need to test to say for sure of course.
Hope somewhere in that ramble I answered your question.
michael