Candlesticks not profitable Massey University Research

[hanzam]

The Massey University Finance Dept. carried out research by Ben R. Marshall, Martin R. Young and Lawrence C. Rose, Dept. of Finance established from their research that candlesticks do not work and are not profitable, this research was released 16 April 2007.

We investigate the profitability of the quantitative market timing technique of candlestick technical analysis in the U.S. equity market. Despite being used for centuries in Japan and now having a wide following amongst market practitioners globally, there is little research documenting its profitability or otherwise. We find that these strategies are not generally profitable when applied to large U.S. stocks. Basing trading decisions solely on these techniques does not seem sensible but we cannot rule out the possibility that they compliment some other market timing techniques.

Download research link http://ssrn.com/abstract=980583

 

[TheBramble]

This research in my view is equivalent to saying blue line charts are not profitable, but black ones are.

It's all the same data - we're only talking about methods of representation. The chart 'patterns' in formal Japanese Candlestick Charting (JCC) may have different, fancier names to non-JCC methods and formations. And there are JCC formations which have names, where they may not have a direct equivalence in line or bar charts, but that's about it.

And don't forget, a lot of what we take as Japanese candlestick 'formations' are actually very recent, and Western, additions to the ancient panoply based on basic (media and mode independent) chart formations.

A more appropriate piece of research would be to assess the efficiencies of the various formations regardless of style or mode or data representation. This has been done on many occasions in the past and has shown, as you will already intuitively be aware, that the most basic patterns and formations are the ones that stand the test of time. X-MA, Support & Resistance, Range Breakouts.

This really is a very good case of Traders’ myopia gone academic – looking at the representation rather than what it really represents.

 

[nine]

A stinging attack but not based on fact.

If you read the research you'd find that they did test patterns. And although the patterns could be applied with any type of chart (well, duh) they are classic Candlestick patterns. Not classic bar chart patterns - you know, they're the ones promoted in books on Candlesticks.

The next thing you'll be telling me that people testing fibs aren't doing it right.

 

[TheBramble]

A stinging attack but not based on fact.

If you read the research you'd find that they did test patterns. And although the patterns could be applied with any type of chart (well, duh) they are classic Candlestick patterns. Not classic bar chart patterns - you know, they're the ones promoted in books on Candlesticks.

The next thing you'll be telling me that people testing fibs aren't doing it right.

I wouldn't have commented on the research had I not read it thoroughly. Additionally, I was offering a view, my personal view, and not a ‘stinging attack’.

If you art an expert on Japanese Candlestick charting, then you may want to consider widening your research into the non-JCC patterns and formations in order to round off your education and be able to appreciate the distinct similarities, which are by far the greater number of them, in both systems. If you are not an expert in Japanese Candlestick Charting then to simply buy into any bit of research which comes your way which appears to offer you more than you already have or understand is unlikely going to be sufficient as a default strategy in the long run to provide you with the appropriate degree of comprehension to engage with someone who is an expert in either or as in this case, both fields.

As for research into Fibs I understand a full 61.8% of all research in that area is seriously flawed.

 

[alexander]

The next thing you'll be telling me that people testing fibs aren't doing it right.

Some are not doing it right either. Fibs are just potential support levels. Some promote them as potential entry levels. Fibs can only be confirmed after the fact. You may get a 38.2% and then go down to 50% and then to 61.8% and so what? Fibs are not useful at all unless considered together with other indicators.

Just like with candle sticks. Viewed alone are virtually worthless. I spent a lot of time testing candle sticks in the 90's. I found that the promoted formations in conferences were amazing losers over a broad spectrum of markets. Still people paid to attend the seminars.

Alex

 

[alexander]

The Massey University Finance Dept. carried out research by Ben R. Marshall, Martin R. Young and Lawrence C. Rose, Dept. of Finance established from their research that candlesticks do not work and are not profitable, this research was released 16 April 2007.

No competent trader should wait for academics to tell him the obvious. It is then too, way too late.

Nevertheless, the study is far from complete from a quick glance I had at it. It has problems with profit targets and they try to circumvent that by considering a 10 days limit on open positions. Do you expect academics to know anything about trading and system development?

Alex

 

[hanzam]

Steve Nison's response to Massey University Research

The author used definitions of candlestick signals without validating how correct these signals are. Just because candle signals are on a software package does not make them correct. For example, for a hammer, there are four rules:



1. The color of the real body can be black or white.

2. A bullish long lower shadow that is at least twice the height of the real body.

3. It should have no, or a very short, upper shadow.

4. The market must be in a downtrend



Rules 1 and 2 are easy to quantify. But how was rule 3 defined (i.e., what was the maximum length of the upper shadow for this to be defined as a hammer)? Or rule 4? How was a down trend defined? A candle signal needs two criteria. The first is the shape of the line, the second is the trend. Thus, if there is no down trend there is no hammer—even if the shape of the line looks like a hammer.



The most serious error was basing the success of a candle signal on what happened the next session. Whether one trades based on a candle signal should be fully dependent on the risk/ reward of the potential trade. This means that one should have a stop and a price target before placing a trade, and not just buying or selling because there is a candle signal. For instance, a stop should be placed under the low of a hammer line. Unless one has a target that is a multiple of this risk, a trade should not be initiated and the hammer should not be used to buy.



At my seminars I have a series of pivotal rules. One of the most important is, “Not every candle signal should be used to buy or sell—always first consider the risk/reward.” And this article doesn’t take this vital risk/reward aspect into account. As such there are many times when a candle signal should be ignored. This study was based on just blindly buying or selling on every candle signal. Using a candle signal to trade without considering the risk/reward is like, as the Japanese proverb states, “leaning a ladder against the clouds.” This is why we have our DVD workshop series which shows how to combine candle charts with Western tools and most importantly money management. Using candles in isolation is the major mistake most traders make and that is why we emphasize the importance of correct education.



Steve Nison, CMT

President- CANDLECHARTS.COM

 

[alexander]

A candle signal needs two criteria. The first is the shape of the line, the second is the trend. Thus, if there is no down trend there is no hammer—even if the shape of the line looks like a hammer.

If one knows the direction of the short-term trend, there is no need to consider candle sticks or any other patterns. Just buy if the short-term trend is upward and sell if it is downward with appropriate stops.

Any attempt to introduce a trend bias of course manipulates the statistics on your favor. However, this makes the signals from candlesticks highly subjective formations.

Supposedly, candle sticks are short term signals. Nobody should care about the trend if they are to be reliable. By introducing the additional two criteria you are affirming the results of the study that pure candle sticks do not work.

As a matter of fact, your criteria point to complicated subjective filters involving candel sticsks rather than candel sticks themshelves, which was the objective of the study. It is probable that the filters can work as well without the full candle stick formations.

Alex

 

[jacinto]

Hanzam,

so you have some research. I dont know, I dont even use candlesticks, so dont care anyway. but I got an honest question for ya

what is your bottom line with this?

would be good to know such that we can either keep the interest here or move on to other threads.

j

 

[hanzam]

Jacinto

To have a debate about the applications and results about candlesticks with other traders and to get to the bottom whether candles work or not. At present we have conflicting views whether candles work or not. There is no other hidden motive I am a private trader I have been trading for five years.

 

[jacinto]

To have a debate about the applications and results about candlesticks with other traders and to get to the bottom whether candles work or not. At present we have conflicting views whether candles work or not. There is no other hidden motive I am a private trader I have been trading for five years.

thank you for your response.

what i can say is that candlesticks are another weapon in the arsenal and are not "the ultimate thing" or "the only thing". I change my charts to candles from time to time, but thats it.

By the way, I did come across some academic research that, if i remember correctly, had different conclusions and that certain candlestick patterns were actually quite accurate.....
thank you, now I got to go and dig it up but will post it when i find it.

j

EDIT: just wanted to add that in the same way as no language is better than another (i.e. French is not better than Spanish) , candlesticks should be seen as a language in which to interpret price action, bars are another, lines, etc. .

 

[jacinto]

took me ages. didnt read it completely. just went through the conclusions, so dont have current opinion. need to read properly, but you bu99ered my curiosity

j

http://www.pitt.edu/~caginalp/Paper65.pdf

 

[timsk]

To have a debate about the applications and results about candlesticks with other traders and to get to the bottom whether candles work or not. At present we have conflicting views whether candles work or not.

It's the trader who 'works' or 'doesn't work' - not candlesticks. Candlesticks are merely a graphic means of representing price movement and, in this respect, are no different from bar charts. A trader who uses charts as part of their trading is not going to be profitable using bar charts but unprofitable using candlestick charts. Said trader may prefer one type of chart over another but that's as far as it goes. So, candlesticks don't 'work' or 'not work' just the same as bar charts don't 'work' or 'not work'. Charts are just a tool to help traders make trading decisions. If used properly, they can be extremely useful. The same applies to bar charts, P&F, Ichimoku, Kagi, Renko etc. If used properly, a good bread knife is a very useful and highly effective tool. However, if one holds the blade and attempts to cut a loaf of bread with the handle, the result is likely to be a pile of crumbs drenched in blood.
Tim.

 

[barjon]

It's the trader who 'works' or 'doesn't work' - not candlesticks. Candlesticks are merely a graphic means of representing price movement and, in this respect, are no different from bar charts. A trader who uses charts as part of their trading is not going to be profitable using bar charts but unprofitable using candlestick charts. Said trader may prefer one type of chart over another but that's as far as it goes. So, candlesticks don't 'work' or 'not work' just the same as bar charts don't 'work' or 'not work'. Charts are just a tool to help traders make trading decisions. If used properly, they can be extremely useful. The same applies to bar charts, P&F, Ichimoku, Kagi, Renko etc. If used properly, a good bread knife is a very useful and highly effective tool. However, if one holds the blade and attempts to cut a loaf of bread with the handle, the result is likely to be a pile of crumbs drenched in blood.
Tim.

hear, hear

 

[guyroyle]

candlesticks

free cd check it out at ,candlestickgenius.com. just ordered mine so carnt coment on it sorry.

 

[Lightning McQueen]

It's the trader who 'works' or 'doesn't work' - not candlesticks. Candlesticks are merely a graphic means of representing price movement and, in this respect, are no different from bar charts. A trader who uses charts as part of their trading is not going to be profitable using bar charts but unprofitable using candlestick charts. Said trader may prefer one type of chart over another but that's as far as it goes. So, candlesticks don't 'work' or 'not work' just the same as bar charts don't 'work' or 'not work'. Charts are just a tool to help traders make trading decisions. If used properly, they can be extremely useful. The same applies to bar charts, P&F, Ichimoku, Kagi, Renko etc. If used properly, a good bread knife is a very useful and highly effective tool. However, if one holds the blade and attempts to cut a loaf of bread with the handle, the result is likely to be a pile of crumbs drenched in blood.
Tim.

hear, hear

A fine, well written piece Tim as always, I'm sorry to say im going to have to take issue with it.

The premise for the thread and the article in question being do candlestick patterns actually work or from our point of view can we expect a profit if we use them in the market. The article focuses on the profitability of one, two and three bar patterns and "any statistical significance to the profits from following candlestick signals", not about the representation of the price itself.

I personally have never held the view that chart patterns work well, its not an area of TA that I follow, so the article was of interest in at least giving us the opportunity of debating the issues at hand.

This could shatter a lot of belief systems out there and also statistically shows why its hard to make money in the markets, at least by following candlestick patterns. Am I risking a flaming by perhaps suggesting that other patterns on other price styles may statistically go belly up when put up under such scrutiny:cheesy:

Lightning

 

[alexander]

To have a debate about the applications and results about candlesticks with other traders and to get to the bottom whether candles work or not. At present we have conflicting views whether candles work or not. There is no other hidden motive I am a private trader I have been trading for five years.

It is easy these days to demonstrate whether they work or not. Any backtesting software will do. Since you have done a lot of work with candle sticks, it may be profitable at this point to show a couple of then that work for a major index while the results are statistically significant.

This is not a matter of liking or disliking a particular method. For technical methods the proof is in the pudding. It does not matter what the claims are, the method either works or does not when put to test.

I was surprised once during a seminar to hear a speaker claim a candle stick pattern worked well while it only had 5 trades in almost 6 years with 5 of them being winners.

Alex

 

[zupcon]

A good bread knife is a very useful and highly effective tool. However, if one holds the blade and attempts to cut a loaf of bread with the handle, the result is likely to be a pile of crumbs drenched in blood.
Tim.

Thats a very good analogy.

Ive read a few academic studies recently on the validity of various TA studies, and although Im in complete agreement with the formal statistical approach that they take, they all seam to suffer from the same types of problem.

They dont tend to analyse a signal in any type of practical context, for example, they'll assess the validity of trend continuation signal, in the absence of an underlying trend. A simplistic candle reversal pattern comprised of a couple of bars for most of the time will mean very little, I mean, theyre not predictive in any way, they just happen have the capacity to illustrate and highlight patterns in behavior that tend to repeat at points of capitulation or euphoria, like a big flashing neon sign pointing out to the dim witted whats actually occurring in front of their eyes. Its all a matter of context, you cant just pick these things out of a timeseries and seriously consider it to be a signal, yet time after time thats just what academics do. Id have a bit more time for them if they actually analysed a sensible set up

In order to provide repeatable experimental studies, academics will routinely use grossly oversimplified techniques, to construct a clunky framework over a highly complex non linear behavior, most of the research into fibs suffers from the problem of identifying swing points. Its a catch 22 situation, academics use a simple filter and prove fibs dont work, the fibs crowd, use the benefit of hindsight to pick out instances where they did work. Both approaches are equally bad scientific methodology.

The results from these studies are generally undertaken outside of the context of any form of trading system, trades are placed regardless of risk control, or risk reward, or without regard to behavior in higher or lower timeframes, a small black candle appeared after a large white candle, so its obviously an entry, you have to admit, this is unrealistic

The most bizarre aspect surely must be using a metric such as assessing the profitability after X bars. How many people do you know who operate a system, where candlestick patterns form any part of the analysis, where the exit strategy is to take profit after X bars ? , can you imagine that, finding yourself in a strongly trending market then thinking, oh dear, Id better get out now because 8 bars have elapsed !, or seeing your target hit within 4 bars, but staying in because you have to stay in the trade for 8 bars only to exit at a loss !.

What do these people think the markets are, a clockwork toy ?

 

[timsk]

A fine, well written piece Tim as always, I'm sorry to say im going to have to take issue with it.

Hi Lightning,
Thanks for the compliment and no need to apologize! I accept the points you make in your equally fine and well written reply! I also concur with your view about other chart patterns going belly up if put under the same scrutiny. But . . . this has nothing to do with candlestick patterns as such. It boils down to the ol' chestnut of Mechanical Trading V's Discretionary Trading and why - in my humble opinion - the latter will always prevail over the former in the long run.
Tim.

 

[Splitlink]

A candle shows high, low, open and close points. They don't "work". They, simply, provide information. What the trader makes of that information, if he uses it, is what makes him a successful trader, or not. There are quite a lot of patterns thrown up by looking at candle charts, but the same patterns can be seen on line, P&F and other charts and I would say that, if they do not give the trader useful information then what is the use of TA to him? A trader in such a dilema can only go back to FA. However, I am not saying anything new (I rarely do, I'm afraid ) when I say that patterns on a chart repeat themselves time and time again, notwithstanding the time frame. FA can show nothing of this. It can, however, throw up questions on the financial health of a company that TA does not do. Such questions could help traders decide whether to buy breakouts on the upside or sell on the downside. Frankly, I won't trade long any share that has too much debt, a high PE and earnings that do not decently cover the dividend it pays, no matter what the chart says. Nevertheless, there are stacks of wellknown shares that do just that and the daily share volume leads lots of people to, quite happily, trade them. Potential time bombs!

Split