I was shocked, yes shocked, when the energetic Mr. Cramer revealed his secret stock selection method on CNBC. Naturally, this was too good an opportunity not to code it immediately into the StockCharts scan engine. However, you will run into the usual problems of running a mechanical scan versus Cramer's discretionary entry process. Anyway, here it is. Now, why was I shocked? Well, more on that later.
The Cramer Process
Apparently, Cramer begins this stock selection process with the new highs list. He is particularly interested in stocks that make new highs when the market is tanking. According to the CNBC article, a stock is on the new highs list because "it is part of a bull market, it announced fantastic earnings or the sector has tremendous sales momentum". Like every well-informed fundamentalist, he then delves into the stock's earnings.
But when does he actually buy? What is his entry rule? Here it is according to the article: "Cramer likes to wait for a pullback before pulling the trigger to buy a stock. The key is to buy on weakness and sell into strength. He recommends waiting until a high-quality stock falls at least 5 percent, as that will give you a solid entry point." But wait, there is more. You have to understand why it has pulled back. He admonishes, "The most important tip is to make sure you understand why a stock has pulled back. Make sure there is not a solid reason why everyone is selling the stock and driving the price down, otherwise the stock could turn out to be a disaster." Ordinarily, this is beyond the ken of mere technical analysts. Anyway, without further ado, here is the scan code.
Cramer's Secret Recipe: Scan Code: 5% Pull Back
Figure 1: The StockCharts ScanCode for the Cramer method. Another pathway is to set a lower bound for the pullback.
I tried to follow the 5 percent pull back rule in the sample scan code (see Figure 1). I first limited the search universe by using the usual volume rules and limiting the search to the S&P 500 stock universe. Naturally, you can easily broaden the search. I then tried to find the pullback in three steps. I first said today's high must be lower than the 253-day high, which makes sense, since we have a pullback. Then, I specified that today's high was greater than the 20-day ago value of the then 253-day high, so that the stock had not "pulled back too much" and was not in a down-trend (for another approach, see Figure 7 below). Lastly, I tried to find a pullback greater than 5 percent. The code seems to work reasonably well. I found only six stocks to match the pull back today (see Figure 2), and today was certainly a good day to look.
Figure 2: Results of the scan in Figure 1. With the Dow 30 down some 158 points, today was a good day to look.
I then took a closer look at Oracle, to check if the scan was producing the desired results (See Figure 3). Oracle had popped to a new high on earnings a few days back. The 51.85 high gave a 5% pull back target of 49.25 or so. The most recent close on the chart was below this threshold, so we can say with confidence that the code seems to be working in the right way. You can look up Oracle's earnings forecasts here. If you like it, you still have to decide how to implement Cramer's entry point suggestions.
Figure 3: Oracle has pulled back more than 5% from its post-earnings high, to meet the initial conditions of Cramer's recipe. You still have more work to do though! (A live chart is here.)
Variation: Pull back means close below 10-day half-width Bollinger band
You can define "pull-back" in very many ways. One simple way is to say there has been a satisfactory pull-back if the close is below the 10-day half-width Bollinger band. The code for this scan is in Figure 4. The results are in Figure 5 and I checked the Abbvie chart to confirm that the code is working as expected (see Figure 6).
Figure 4: Pull back defined as close below 10-day half-width Bollinger band.
Figure 5: The pull-back below bands found more stocks meeting this pull-back criterion.
Figure 6: Abbvie (ABBV) is first on the scan list from Figure 5. It has made a new high recently, and the close is below the 10-day, half-width Bollinger band, thus meeting our pull-back criterion. (A live chart is here.)
Why was I Shocked?
Way back, in 1992, in my book "Beyond Technical Analysis", I discussed this very concept in great detail (see reference 1 below pp. 101-112), and I called it the Channel Breakout Pull Back pattern or CBPB. So imagine my surprise when the great man revealed his secret recipe with much fanfare.
The biggest part of the puzzle here is when to get in. You can buy strength (new highs again) or weakness (the pull-back) or what ever criterion you choose. Given that you can define the pull-back itself in many different ways (see Figure 7), and choose a variety of entry points, a very large number of variations are possible.
Figure 7: Here I used a range of pullback values to explore another pathway to broaden the number of stocks picked up by the scan.
Thank you for checking in. I hope you will experiment with the scan code, and try out other variations. Also, please subscribe to the blog below, to get an email alert of my next post.
Chande, "Beyond Technical Analysis", (1992); ISBN: 0-471-16188-8
Chande, "Beyond Technical Analysis, 2nd Edition", (2001); ISBN: 0-471-41567-7