This is interesting. I got on this track from Dogwood's site where he describes a battle royal going on between Joe Cit at Intelligent Investing and the owners/proprietors of INVESTools. I don't watch too much television normally and absolutely none during the day. Apparently INVESTools has a never ending infomercial regarding their amazing method of stock selection that will "get you rich quick".
Joe's position is simple - how do these people prove their claims other than anecdotally? My take is that every company that says this is the "best way to do things and send us thousands of dollars to show you how" should be made to prove their contention.
Turns out Dogwood had done a post some time ago that described the INVESTools method in all of its complexity.
I never pass an opportunity to test a new (or old) method so I developed a filter from that post that looks like this -
show stocks where close > ma(30)
and MACD fast line(12,26) is above 0
and Fast Stochastic(15,3) Fast %K crossed above Fast Stochastic(15,3) Slow %D within the last 1 day
and then I back-tested it.
You will note that there is no volume nor price constraints in this filter. I don't think any are called for in the INVESTools model. You know that I always use volume and price constraints in all of my filters - I never leave home without them.
Most readers of this Blog know my methods and they know that I use sound testing techniques and that I always present the true case regardless of the outcome. As a result I have negated several of my own closely held beliefs and opened minds to a lot of new ways of thinking. For example, buy when the MACD crosses into negative territory - not when it crosses out. So when I present the results here you know they are honest.
For this back-test I used two periods - the first period, 10/06/06 to 02/06/07, we have found in the past to be very fertile and profitable and the second period, 11/06/06 to 03/06/07, we know to contain February 27th.
Period 1 - 67% win percentage, 4-day ROI 34%, 30 day net .88%
Period 2 - 59% win percentage, 4-day ROI -67%, 30 day net -1.13%
And that is just awful. And the other problem with this approach is that it generates between 200 and 2000 signals a day.
Certainly if I cherry-picked the results I could come up with a whole slew of winners but I probably could have come up with all of them through other means as well including randomly throwing a dart at a newspaper or letting the magic coin select them or watching Jimmy Crack Corn Pone and getting rich.
Bottom line - I don't think this approach is worth the time and effort it would take to sort through the output looking for plums among the lemons.
Moral of the story - Rule 1 wins again - nobody knows nothing including me - So, if you insist on paying someone for advice don't ever forget that you are buying what that person knows.