The Perpetuated Fraud of Candle Charting

[TraderZones]

Quote from Daal:

Yeah, like in the thread where I provided evidence the H&S pattern(as defined in 8 TA books) does not work in practice. then the mod went ahead and edited my OP(I had never seen that before, and I'm here since 2002) to write how he is not giving up in his TA religion because 'you are supposed to antecipate the H&S', so he essentially agreed the pattern doesnt work. When presented with contrary evidence, you want to close your eyes, that sounds like an awesome idea

Candle patterns do not work, and this is based on empirical tests performed by statisticians(check out Practical Speculation). Since they dont work one could say 'but I found/created some that do', if you keep testing long enough of course you will find stuff that works, its called curve fitting, doesnt mean it will work in the future although it can help in that regard.

The point is since most candle patterns dont work this shows that the idea that the 'japanese old traders were so wise and one can get rich from their stuff' is junk. So trying to find ones that do work might just get you fooled by randomness because you will be trading in a flawled trading theory

+1. I have seen candlestick patterns debunked in several studies. Believers in this stuff invariably claim "works for me", "I know someone who...", "They are self-fulfilling..." or other statistically meaningless testimonials or beliefs.

Unsubstantiated beliefs and opinions are not the tools of a serious trader. They are the of the religious, the superstitious or the uninformed. Testing (back, forward, etc.), validation, research and other methods are. Some traders mix several methods, and never really grasp that one or more of the methods may add no predictive value.

The only thing that matters is, does a particular method give POSITIVE EXPECTANCY - an edge. The overwhelming failure of most traders, and the lack of convincing proof of popular methods like candlestick patterns or technical indicators are not unrelated.

An edge is something that helps give you a provable profit factor reasonably above 1 in the situation it is applied. If not, you are wasting your time.

 

[dcraig]

Quote from JSSPMK:

BS

A quick glance over historical charts will show that a hammer works more often than not.

The eye is terribly deceptive in these matters. As I said I tested such things as hammers using a stock screener I wrote and couldn't find any edge. I rather like the use of a screener for this sort of thing because you are testing over thousands of instruments and statistically any edge should show up if it is real.

It is a sobering experience doing this sort of testing as you quickly find out that a lot of standard TA is bunk.

Which is not to say at all that an edge or edges cannot be found, and that it (they) may be surprisingly simple.

 

[JSSPMK]

Quote from dcraig:

The eye is terribly deceptive in these matters. As I said I tested such things as hammers using a stock screener I wrote and couldn't find any edge. I rather like the use of a screener for this sort of thing because you are testing over thousands of instruments and statistically any edge should show up if it is real.

It is a sobering experience doing this sort of testing as you quickly find out that a lot of standard TA is bunk.

Which is not to say at all that an edge or edges cannot be found, and that it (they) may be surprisingly simple.

You are looking to establish an edge from a standard matter? When you were testing, what rules have been applied for stop management? Mechanical systems do not allow room for wiggle. Also, what about re-entry after a loss based on same signal, was that part of the test? What rules were applied to position exits?

 

[Buy1Sell2]

Quote from JSSPMK:

Although I understand where you're coming from, what about the following.

For instance, what if you wish to trade a 5 minute chart's action, which is bullish, but which higher chart do you pick if?

30min bearish
60min bearish
240min bullish
Daily bulllish
Weekly bearish
Monthly consolidating

I don't trade a bullish signal on a 5 minute chart unless the 30 minute chart is bullish.

 

[R. Raskolnikov]

LoL, you people and your "testing". Did it ever occur to you that perhaps you don't have all the correct input to determine such things? And so drawing a blanket conclusion based on your very limited understanding is ludicrous?

Seriously, the 95% failure rate has got to be accurate. Between ego and laziness alone, it's easy to see why so many fail.

Quote from dcraig:

The eye is terribly deceptive in these matters. As I said I tested such things as hammers using a stock screener I wrote and couldn't find any edge. I rather like the use of a screener for this sort of thing because you are testing over thousands of instruments and statistically any edge should show up if it is real.

It is a sobering experience doing this sort of testing as you quickly find out that a lot of standard TA is bunk.

Which is not to say at all that an edge or edges cannot be found, and that it (they) may be surprisingly simple.

 

[Daal]

Quote from R. Raskolnikov:

LoL, you people and your "testing". Did it ever occur to you that perhaps you don't have all the correct input to determine such things? And so drawing a blanket conclusion based on your very limited understanding is ludicrous?

Seriously, the 95% failure rate has got to be accurate. Between ego and laziness alone, it's easy to see why so many fail.

The 95% failure rate is very easy to understand once you see the most traders dont have an edge. They buy typical trading books that put into their head that if they have discipline and know TA patterns, they will be millionaries

Extracting consistent profits from competitive markets in the information age is far from easy, no matter what the trading gurus that are selling stuff try to tell you. Saying stuff like 'use candle patterns in conjuction with other things', what is that other stuff and where are the test results from a statistical significant sample?Let me guess its 'bit discretionary'

 

[Daal]

Quote from TraderZones:


The only thing that matters is, does a particular method give POSITIVE EXPECTANCY - an edge. The overwhelming failure of most traders, and the lack of convincing proof of popular methods like candlestick patterns or technical indicators are not unrelated.

An edge is something that helps give you a provable profit factor reasonably above 1 in the situation it is applied. If not, you are wasting your time.

Agreed. The majority of trading books try to give the impression that getting an edge is easy, some of them go as far as saying one can make money with random entries(van tharp).
Please, if it were that easy capitalism would be broken, you see that all the time in the business world, a new type of business earn above average returns for a while, all the idiots flock there and returns do down. The same thing should happen with 'easy' trading returns, there is little reason why it would be different, specially after the internet and the faster dissemination of information

 

[R. Raskolnikov]

Are you nuts? Who would test AND reveal to the world an ultra consistent proprietary model?? I'm sure as hell not going to lay before you and everyone else all my hard work. And, no, it's not discretionary. Everything is based on reading sequences of bars a certain way (the same way every day) and then waiting for a few specific bars (ie bar formations) to form in the context I am in, in order to place a trade. Consistent application of the same process day in and day out (no gut feel or shooting from the hip whatsoever).

Do you honestly believe that it's just not possible to figure out how to read candles correctly in order to produce consistent profits?

Quote from Daal:



Extracting consistent profits from competitive markets in the information age is far from easy, no matter what the trading gurus that are selling stuff try to tell you. Saying stuff like 'use candle patterns in conjuction with other things', what is that other stuff and where are the test results from a statistical significant sample?Let me guess its 'bit discretionary'

 

[JSSPMK]

Daal, I have a question for you. If Blackjack paid 10:1 could you beat it a casino?

 

[NihabaAshi]

Quote from Daal:

Agreed. The majority of trading books try to give the impression that getting an edge is easy, some of them go as far as saying one can make money with random entries(van tharp).
Please, if it were that easy capitalism would be broken, you see that all the time in the business world, a new type of business earn above average returns for a while, all the idiots flock there and returns do down. The same thing should happen with 'easy' trading returns, there is little reason why it would be different, specially after the internet and the faster dissemination of information

Be careful whom you agree with because you've already shown a distaste for vendors. TraderZones is a vendor and he's been busted several times trying to market his trading system designs here at EliteTrader.com for $250,000 dollars (U.S. dollars) while getting rid of his competition (see feedback section).

Looks like you're still doing poor research.

You guys are funny, you use charts as part of your trade decision but YELL you don't use TA. You preach that other things are more important (money management, credit crisis) but can imagine how others using the same with candlesticks, TA or whatever is able to make a profit.

You quote poor backtesting resources and refuse to say what your own edge is while DEMANDING others to prove their edge to you.

In addition, your profitable trades suddenly aren't via randomless while others must be. Guys and gals...time to grow up.

Simply, the next time you talk about that profitable trade you've made while trying to sell the fact that you know what you're talking about...show proof. Show exactly what you demand of others. Without proof, your debates have no merits especially when you align yourself with those that represent some of the things you're attacking.

P.S. Gotta go and tie down some stuff outside...a big snow storm is about to hit here in Québec City.

Mark