Any value to automated candlestick pattern recognition? Can it provide some edge? Has anyone gotten any real good results from it? No simulation results or academic studies please.
I know a lot about Japanese Candlesticks (automation and manual) pattern recognition via real-trading. Also, you can look up past info here by me about Japanese Candlesticks in general via my old user name NihabaAshi. I also wrote the thread called Trading Hammers revisited @ http://www.elitetrader.com/vb/showthread.php?s=&threadid=52880 With that said, there's no edge for trading purposes if trading automated pattern recognition alone as in not in combo with any other method. In addition, there's a lack of adaptation whenever market conditions changes and markets change many times every year. Mark
There is some possible edge in candle trading but you have to combine them with other indicators. At least, this is what one study indicated http://www.westga.edu/~bquest/2008/candlestick08.pdf Candlesticks are for the most part a small subset of price patterns and it is important to realize that because there are many more formations in that context that may be profitable. http://www.priceactionlab.com/Literature/article2/article2.html One must be very careful with either candle or price pattern formations because many random formations with high win rate can be the result of survival bias. Some formations are very robust and work for many years in daily bars. One must study them carefully before using them.
Thanks a lot for the links to the paper and software. I did some research and backtesting in the meantime and I found out that traditional candlestick formations offer a very small edge. The software is very interesting and I think I am getting it. I also like the indicator part of it. The idea that I can load it up with data and let it run through the weekend is intriguing.
http://www.iijournals.com/doi/abs/10.3905/jod.2005.580514 No evidence of any predictive ability for German futures.
You need to be careful about the academia testing because most of the stuff out there does not involve proper trade management after entry and many of the identifications are incorrect or a variation not typically used by traders. Simply, the only way you can determine the merits of the testing is if/when the author reveals the in-depth details of the criteria used and the trade management rules after entry. Good luck with that because most will not even if you personally contact them for such details. The few that do share such details is when I've discovered the flawed testing especially when such research had ignore market context. There's a discussion about such over at Traderslaboratory.com in the past. Market context is very important to someone actually trading the patterns whereas not important to the academia doing research. Lets put it this way, computerized identification doesn't have the ability to incorporate market context. Just as important, computerized identification discloses the name of the candlestick pattern and not the variation (see one of my prior threads here at Elitetrader.com with details about 15 variations of the Bullish White Hammer pattern...most have not edge whereas a few do...my prior user name is NihabaAshi). Last of all, there are other studies by academia that say candlestick patterns do have a small predictive value. However, just like those that say it doesn't, they don't reveal their criteria nor trade management rules other than to say it was "computerized identification". Simply, obviously different variations are being tested or there's too much discrepancies (errors). Recommendation: Learn a few Japanese Candlestick patterns for trading purposes and stay away from the computerized identification unless your intentions is to be an automated trader or academia research paper. Mark
This is true. Whenever I contacted a university professor I never got a reply. This is just amazing. You contact Bill Gates and you may get an answer. You contact Dr. Joe Crank of Nowhere Univ. and he plays big shot. I bet the study in that paper is completely wrong, from start to finish.
I think the biggest advantage of candlestick charts is there improved visibility vs the normal western display of high,low,open,close.IHMO
I've heard candlestick is not a system to base trade on and should only be used as a TA tool. Any thoughts?