Not sure if this is statistically meaningful. Anyone ? 2.43% difference between the worst and the best result doesn’t look meaningful to me.
I've got one or more CANSLIM books in my library, and only read them out of interest of humor. Nothing else. Let's think about this for a minute. Do you honestly think that a simple solution such as CANSLIM really can out-beat the market? Considering, that even just one supercomputer owned by Jim Simons and his hedge fund carries out algorithms of magnitudes of orders higher than this simplistic system? Any possible repeatable gains would have been arbitraged away long ago.
Most likely, if you try it on a similar data set (say 1960-1989 instead of 1990-2019), you'll get very different results. The overall distribution of positive and negative outcomes may be similar, but the best MA pairs very likely won't be. Most likely the best pairs just happen to "fit" whatever data period was used, and won't be superior to other pairs in the futures.
You can’t arb the fact that institutions are pulling billions into a stock. What do you wanna do ? Okay you can front run them. Simons is certainly not making money out of complex algorithm otherwise he would have already blown up.
Almost all winners of the US Investing Championships are Canslim/modified Canslim traders. Maybe it's pure luck and being in the right place at the right time,but this has been a Momo players paradise.. Ryan and Kullamagi have put up astronomical returns using a modified Canslim system.. I'll be the first to admit I am not a directional trader,but a modified Canslim approach would be my approach... It employs money management, position sizing and leans towards technical factors with fundamental filters... What specifically don't you like about it??
I like O'Neil and think CANSLIM is one of the more promising methods out there, but like so many things, the results vary widely. Some individual traders have done amazingly with their variations during bull markets. How they've done over longer time periods (say 10+ year periods including bear/sideways markets) I'm not so sure. Have any reported long-term performance or just that they've won trading contests? AAII shows their screen using it has almost doubled the market's performance...but why have all attempts by CANSLIM mutual funds failed? I'd be curious how much of the AAII performance has been in real-time (not backtested) and how realistic it would be for retail traders to mimic their performance. Robert Lichello's AIM book's system from the 1970s is another strategy that still generates some attention and followers decades later. It's pretty much the opposite-a buy low and sell high system. Again, it's hard to find anyone who's followed it in real time for decades to compare with buy-and-hold. And Lichello changed the rules a bit in his books, eventually holding 80% stocks and only buying low/selling high with 20% cash reserves (making it pretty dang close to buy-and-hold).
The main idea as I know came from reverse engineering the very successful stocks to capture those profits for future stocks which should be found then by this custom made CANSLIM method. But I do not think you can rebuild such a winning formula for finding successful stocks just based on statistics like O'Neill have done it. There are too many other variables in play that can influence a stock/company their performance/success. Just do not believe that this can work out. Those successful CANSLIM traders I just might guess it is because of the law of big numbers, that some successful traders picking the right stocks at the right time are shown (in such competitions).