I suspect CANSLIM works better when the Dow Jones Industrial Average is near the low value of the prior 12 months. Stock price breakouts when the general market index values are near historic high values might not show trend continuation. My theory is that a selling phase creates a big pile of cash. Selling might be mostly of the poorer performing issues. If stock prices rally then some of this cash reenters the stock market and buys into the better performing issues. So the series of price rallys and reactions represents a kind of evolution where capital is reallocating out of the weakest and into the strongest. A kind of "Survival Of The Fittest". Probably another name for capitalism. I check the 52 week new high list almost every day. I look for stock symbols that I do not remember seeing before. Unfamiliar stock issues are sometimes newer companies beginning big price increases. That's how I found CROX. I check earnings growth at finance.google.com but reuters.com and yahoo.com also show fundamental data for stocks. I think William J. O'Neill is unwise in selling if price increases 20 % from the buy target price value. I observe some stocks that meet CANSLIM criteria continue increasing in price for years. All I need to do is hold on and use a trailing stop to guide my exit. CSCO was like that. CSCO went public about year 1990 showing big earnings growth and all I had to do was use a 300 day moving average as a trailing stop or maybe the lowest price in 15 months as an exit stop and I might make enough money to buy a house. The same thing is happening now except the stock symbols are BIDU, DRYS, CROX and some of the other symbols that I read on elitetrader.com. I also think Mr. O'Neill's 7 % stop loss is too tight. Lots of times I stop my losses when price closes less than the lowest price in a year, or 2 years, or the historic low price. I calculate my position size so that I only lose say 0.5 % of account equity. The idea is my profit potential is unlimited but my losses are limited and small. Eventually the big winner appears, offsets all the losses and shows a big profit besides. A similar method is described in Jesse Livermore's book "How To Trade In Stocks" by the way. Mr. Livermore does not describe stop losses and risk management is much detail though.
I've found several problems with IBD and CANSLIM: 1. IBD has a neo-Conservative editorial bias. You just hope that they are as rigorous with their data mining as they are sloppy with their cognitive biases. 2. They engage in historical revision of market calls in The Big Picture. It is definitely not quantitative. 3. The number of stocks O'Neill examined for "winning characteristics" only amounted to about 12 per year. There have been far more winning stocks each year than 12 in the same time period. 4. The screening revisions to the third edition of "How to Make Money in Stocks" garners worse performance than the first edition! 5. Stocks in the IBD 100 are only >=$15 and almost always are overbought, thus prone to failure. By the time a stock hits IBD 100, the majority of the move has already occured. MG