Bulkowski has done some backtesting, see : http://thepatternsite.com/CandleEntry.html He has a book coming out on it: http://www.amazon.com/dp/0470182016?tag=bulkowskschar-20&camp=15041&creative=373501&link_code=as3 He's pragmatic about backtesting, actually trades for a living, and tends to give most of the valuable info away for free on his website.
I've heard such as well. However, I do know three Japanese retail traders... I met them in college, they were foreign students studying international finance and members of my TradeStation user group back in the old days when TradeStation was called Omega Research. They all do use Japanese Candlestick patterns but mainly as confirmation. Also, there were two blogs by Japanese traders (assuming they too were retail) that on a few occasions did talk about a candlestick pattern that had some impact in their trade decisions. Thus, that's 5 out of 6 traders that implied they are Japanese, living in Japan along with being traders. Simply, had learner2007 not said what he said... I would have said I've seen 5 out of 5 Japanese traders say they use candlestick patterns in some trade decisions assuming learner2007 currently doesn't use it. However, if he does use candlestick patterns, that would make it 6 out of 6. Last of all, I did come across a Japanese stock trading discussion forum back in year 2000 and they talked a lot about Japanese Candlestick patterns and used them. It was a very active forum but it eventually was suddenly closed for some unknown reason. Mark
Although I have question his trade management of candlestick patterns in the past... Any trader that knows a lot about volatility analysis will notice that the most reliable candlestick patterns tend to involve volatility expansion along and/or involves a candlestick pattern that consists of more than 3 intervals. I wrote about such from my own backtesting in 1999 about my observations of such for several TradeStation groups. There was a programmer (TA basher) here at ET that did his own backtesting of one sub-group of the Bullish White Hammer pattern. He tested it on what he trades (NYMEX Oil and COMEX Gold). He posted here at ET that his results on a 10 year history was 67% reliability. I asked him if that implied TA works. He responded only if it does the same accross all markets. I responded by asking him if he traded all markets and if he didn't...why be concern with stuff you don't trade nor watch??? He replied back with something like what's my point. Anyways, I've done my own testing and I can say I'm not in the camp that TA doesn't work. However, my testing included candlestick patterns combined with volatility analysis. Mark
They don't work for me. Particular candlestick chart patterns do tend to occur at major turning points. Then the same pattern will occur all over the chart and a reversal just doesn't eventuate. Other methods are easier and more reliable IMO. They are however better than boring ol' bars for visual effect.
This reminds me of the first time I went to Canada. Coming from Ohio, I walked into a restaurant and ordered some 'Canadian bacon' and eggs for breakfast. "...it was quiet, too quiet."
interesting, but people tend to then pick off the "top performers." But in any broad test like this, it is hard to separate a "top performer" from one that is there due to raw probability that some have to be at the top, some in the middle, some in the bottom. I saw other candlestick testing that was not so promising for candlesticks; cannot remember the source though???
>NihabaAshi No, I do not use candlestick patterns in my TA. And I am of the opinion that one should use what he finds best for him. So nothing against that form of TA whatsoever. I was surprised to find how popular candlesticks had become in the U.S. many years ago. As I remember it was around the time that the Japanese discovered W.D. Gann!! If you would go to a bookstore in Tokyo, out of the hundreds of stock market books you would find only a few pages in selected books covering candlestick patterns. I have been a member of the Japan Technical Analysts Association since 1979 and candlesticks are never a topic discussed. The current fad here seems to be indicators, moving averages, Ichimoku, Bollinger Bands, Elliott etc. Also, Renko, Kagi Ashi and most of the other charts that were developed here seem to have died out along with chandlesticks.
I would say again that candlesticks are a tool, and good for some jobs but not all, and you will need other tools. But even a good tool is useless if you don't learn to use it or use it for the wrong job. As for suggesting that candlesticks don't work and therefore you should restrict yourself to lines or OHLC bars, this is muddled thinking. Learn how all the TA tools work and use them all where you need to. Don't know about you but when I drive at night I use my headlights and see better: I could drive without them and just use the streetlamps but that does not allow me to say that headlights don't work, or that as a good driver I should do without headlights.
Here's a thought, let's define candlesticks. There are many types and variants of candlesticks. Several popular charting programs display candlesticks that show user defined, specific information that may be altered with automatic scripts. Personally, I've written a few scripts with the charting software I use and often toggle between the two on any given day. Personally, I prefer bar charts but they are what I cut my teeth on. I see patterns better with bar charts. For disclosure I must add that I use candlesticks mainly to see what the enemy is watching when actually trading. Every night I'll plow through several hundred charts and it's really the only time I'll have a moving average or a candlestick included. Actually it's a sub chart and once again it is often times a fade. I'm not degrading ma's or candlesticks, per se, just that they are a secondary decision process and one that adds insight to what others are thinking. Sometimes a tie-breaker. I know that new traders often times use (default) candlesticks, I suppose this is the underlying rational to my reasoning. edit: 200 sma for weekly/monthly/yearly perspective.