Welcome to NexusFi: the best trading community on the planet, with over 150,000 members Sign Up Now for Free
Genuine reviews from real traders, not fake reviews from stealth vendors
Quality education from leading professional traders
We are a friendly, helpful, and positive community
We do not tolerate rude behavior, trolling, or vendors advertising in posts
We are here to help, just let us know what you need
You'll need to register in order to view the content of the threads and start contributing to our community. It's free for basic access, or support us by becoming an Elite Member -- see if you qualify for a discount below.
-- Big Mike, Site Administrator
(If you already have an account, login at the top of the page)
Are candlesticks patterns statistically significant?
Hi, I'm under the impression that analyzing price action based on arbitrary slices of time/range makes not much sense because you can find a slice size that gets you "rejection" wicks at "support" zones, but you can also find another slice size that doesn't show wicks at the same price level.
For example, a 1min chart might show a rejection candle where a 2min won't.
Or they work because many traders are looking at the same slice sizes (5min, 10min, etc.)? But this will make them extremely susceptible to be abused because "everybody" will know where their stops are.
Can you help answer these questions from other members on NexusFi?
I agree, to an extent. One time-frame that is not arbitrary, however, is the daily. Candle patterns on an intraday chart are less meaningful to me, but on a daily chart I will always pay attention.
It has been my observation in general that certain price-action trading concepts are far more effective on longer time frames than they are on very short time-frames.
If you really want to know whether candlestick patterns are statistically significant, then download a bunch of price data and do a study. You'll learn a lot more that way than you will by asking strangers on the internet what they think
They probably aren't, because I don't think you can quantify what you mean by a particular pattern well enough to run a statistical analysis on it. (Or, if you just decide that one pattern is described a certain way quantitatively, then probably you are being a bit arbitrary about how you classify them, which puts the whole question in doubt.) Patterns of any sort are largely a matter of human discretion and judgment in identifying them. If you can do it, then fine. But if it's a matter of human judgment, it's not going to be quantifiable well enough to really test their significance... something that is up to judgment is not that easily tested statistically.
So some simple advice is try it and see. We don't see many traders here using the patterns that are found in books all that much, at least not exclusively. If you're a discretionary trader, some of it may be part of your toolkit, but be aware that the more discretionary you are, the more individualized your trading is and the less verifiable by some statistical means. (Other than your profit/loss, that is.)
Bob.
When one door closes, another opens.
-- Cervantes, Don Quixote
One candlestick pattern that is easy to test is a hammer, or shooting star. To adherents of traditional candle pattern analysis, these are important reversal patterns, and just about all of the information required to identify them is contained in one bar.
To test the predictive power of a hammer, for example, establish the following criteria (or something similar; these would be the arbitrary conditions Bob was alluding to above):
New 20-period intra-period low = Low < minimum Low of prior 20 periods
Very small candle body = Absolute value (Open - Close) < 10% average true range
Long lower wick = (Close - Low) > 80% average true range
Then test what happens in the following period, or test what happens in the following 3 periods, or after a confirmation, etc. If you're ambitious, see if volume or range size affects the results.
Even if you determine that there is zero predictive power in these candles, you will have learned something.
By the way, anyone here old enough to remember InletCap's "retail bus" pattern?
Charles Bulkowski did a pretty thorough study of the most popular chart patterns and candlestick patterns. All of the stats are on his website. thepatternsite.com (no affiliation). He is really thorough and attempts to be as precise as possible, but the stats are still hard to interpret. There seem to be some promising patterns.
Did he quantify these patterns mathematically with programming ? OR is it some type of subjective analysis . In my experience standalone candle patterns offer zero edge BUT with some context results may be better .
Given the time he died and cannot see his probabilities on candle patterns to be robust in anyway . Without masses of data analyzed by code with powerful computers that he likely had no access to in his lifetime cannot see anyone developing anywhere near robust definitive probabilities ( edge) that you can take to the bank
I think patterns are extremely meaningful and important. The importance however is not in the patterns themselves but in what they represents. Patters are representations of the human psychology.
After years of trading you don't look at the pattern but you understand why it "works", because there is a mass psychology behind it.
In one way or another, all systems use patterns, all algos use patterns as well..... call them deep learning, call them time series analysis, call them technical analysis etc.... they are all forms of pattern recognition.
Concerning the second part of your question: the higher the time frame the more relevant the pattern is. For instance if we break above yesterday high, you have a green candle breaking the high of the previous day candle. This is very relevant.
The same patter on a 4H candles is less relevant, and on a 1H candle even less relevant.... but still relevant.
The reason why higher t-frames are more important is that big money mangers need to use higher t-frame.
This is my opinion, of course but I think many traders will tell you exactly the same.
