10/02/2006

From concept to execution

The concept is simple to find unknown stocks having significant earnings or sales acceleration for the first time. There are many ways to make it work. Here is what you need. My approach is to keep it simple. There is always temptation to make it very complicated and add more factors. It might improve performance in some cases. After trying many ways of complicating it, now I prefer simple things.

Earnings Data
What you need is good quality, reliable and timely earnings data. There are several sources of earnings data like Bloomberg, Thomson, Zacks, Reuters, Wall Street Journal, Investor Business Daily, S&P etc.
I pull data from some of these sources in to TC2007 and my other database software.

Earnings Cut off
Now you need to decide on earnings growth rate cut off. Anything less than 25% is not worthwhile. Higher the better. I also do not look at earnings of less than 5 cents per share. If your cut off is say 25%, depending on market circumstances you will get 1500 to 1000 stocks. I use a higher cutoff. My quarterly database range from 300 to around 50 stocks.

Sales Cut off
This one is little tricky. There are many industries where growth of 25% may not be possible. But again it is safer to take higher cutoff. My sales database also eliminates all stocks with less than 10 million quarterly sales. I use higher cut offs and at best find 50 stocks meeting my criteria in a good quarter.

Once the data is built, you are on your way to just find any breakout strategy to find candidates. Three things will happen after a significant earnings 1) immediate breakout 2) no reaction continues in range (this might breakout later closer to next earnings) 3) reversal ( this typically happens after a string of earning surprises and significant price growth).

Once you do this for extended period of time you will learn more about how these stocks behave over long periods. You will also get most of the big movers before most people have even heard of them. You will be able to tell weeks in advance which stocks will make it to IBD100.

The most critical part in this is building your own databases. If you use scanners to get this data, you will not get same results. The most scan data is batch processed with significant time lags in many cases.

How is this different from the IBD EPS ratings. IBD EPS ratings are good but the way they are calculated they tend to lag on certain set of stocks. Especially stocks which had string of losses for many quarters and become profitable ( which is the case in many new companies or turnaround situation). Another thing is IBD waits for earnings momentum to build before entering. In many cases companies have outstanding earnings only for 1-2 quarters. They immediately react to the earnings. The IBD EPS and Relative strength rating increase , they form cup and handle and then breakdown. In no way I am saying IBD method is bad, I am just pointing out some of the things which I have observed. In fact I highly recommend IBD method to most traders.

Once I have the database I continuously rank it by price appreciation. The objective is to continuously look at top 25 price performers post earnings for opportunities. You will be surprised at the longevity and strength of some trends post a stock had an earning/sales acceleration. For example look at stocks like VPHM, or MT,or NTRI and see what happened post their first significant acceleration.

Coming next: Why I love virgins

No comments: