Bar Sequence Patterns | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Bar chart patterns? 14 replies

3 Bar Fractal with mid-bar and bar on each side 11 replies

Pin bar, pin bar, pin bar 0 replies

Sequence of Order Creation - which one was first? 2 replies

Attachments: Bar Sequence Patterns
Exit Attachments

Bar Sequence Patterns

  • #1
  • First Post: Edited Dec 30, 2015 4:37am Dec 27, 2015 6:03am | Edited Dec 30, 2015 4:37am
The idea started small in Robots Lab, where I was initially interested in Up and Down bar sequence HERE, then Break High and Break Low bar sequence HERE.

In this thread I am willing to further expand the idea of Bar Sequence Patterns to cover ALL types of bars.

Terminology used for bar definition (could be changed as thread progresses),
these codes represent the status of the CURRENT closed bar:

Attached Image (click to enlarge)
Click to Enlarge Name: BSeq_Coding.PNG Size: 37 KB



Bar Sequence EA

  1. Bar Sequence EA v2
    Attached File(s)
    File Type: ex4 Bar_Sequence_02.ex4   30 KB | 1,122 downloads
  2. Bar_Sequence EA v3.


List of Found Profitable Patterns

Attached Image (click to enlarge)
Click to Enlarge Name: bar_sequences.PNG Size: 41 KB
No guts, no glory
  • #2
  • Edited 8:49am Dec 27, 2015 7:57am | Edited 8:49am
Quoting abokwaik
The idea started small in Robots Lab, where I was initially interested in Up and Down bar sequence HERE, then Break High and Break Low bar sequence HERE. Here I am willing to further expand the idea of Bar Sequence Patterns to cover ALL types of bars. Terminology for pattern definition (could be changed as thread progresses): + : Up bar,Close > Open, regardless of body size - : Down bar, Close...
Giovanni Maiani developed such a program, reporting his results in Stocks & Commodities Magazine in November 2002. He reports analyzing 575 American stocks over a period of 15 - 20 years. He found more than 6.14 million candlestick formations. In his words:
"The most reliable, long lower shadow, is present 1.9% of time, and anticipates a declining session 49.12% of time; a rising session 40.61% of time. For traders who are quantitatively based, candlestick patterns are not terribly useful"

Subsequently, Robert Colby, writing in SFO Magazine, Aug., 1996, found "No Major edge for Candlesticks". Further, he states: "Nevertheless, the candlestick patterns tested here failed to beat the simplest trend-following standard of comparison on a significant basis."

Steve Nison, the creator of Candlesticks, responded (S&C Jan 2003, p74) that Candlesticks are "A tool, not a system", and that you need to know the trend to use them.

An Aside on Technical Analysis:
In the Japanese Candlesticks technical method, the basic element is a cylinder with open and close as cylinder limits; with the high and low spiking above and below the open/close base. It is well known and consistently observed in profiles that the high and low prices are rejected quickly by the market. So it appears that candlesticks are giving most weight to prices the market quickly rejects. Consequently, it is no surprise that Maiani found candlesticks to be of little use for trading decisions. Candlesticks and other charts do have the value of offering a panoramic view of market history. While of little use for current, day-to-day trading decisions, a general view of one's market is worthwhile for understanding a market in general.
-Jones
Attached File(s)
File Type: pdf Giovanni Maiani .pdf   46 KB | 1,624 downloads
Markets are not efficient, rather they are effective - Jones
  • #3
  • Dec 27, 2015 8:16am Dec 27, 2015 8:16am
Quoting mzvega
{quote} Giovanni Maiani developed such a program, reporting his results in Stocks & Commodities Magazine in November 2002. He reports analyzing 575 American stocks over a period of 15 - 20 years. He found more than 6.14 million candlestick formations. In his words: "The most reliable, long lower shadow, is present 1.9% of time, and anticipates a declining session 49.12% of time; a rising session 40.61% of time. For traders who are quantitatively based, candlestick patterns are not terribly useful" Subsequently, Robert Colby, writing in SFO Magazine,...
Interesting ...

I agree that candlestick patterns aren't always usefull alone ... However combined with strong levels, the price reaction is often amazing.
MT4/MT5 EAs/Indicators/Alerts coder
  • #4
  • Dec 27, 2015 9:28am Dec 27, 2015 9:28am
Subscribed
  • #5
  • Dec 27, 2015 9:58am Dec 27, 2015 9:58am
Quoting FerruFx
{quote} Interesting ... I agree that candlestick patterns aren't always usefull alone ... However combined with strong levels, the price reaction is often amazing.
My post wasn’t meant to discourage, I believe in “Auction Market Analysis”, as TA does not make any considerations to which Auction Markets actually operate. The use of “candlesticks” is a non-permitted practice. TA methodology is based on the assumption that every price is somehow “fair value” (EMH).

While TA and candlesticks works for some, it definitely does not work for me.

I thought for those such as yourself, who do use candlesticks, might find the stats to be of some value. Stats allow one to form a personal opinion based on facts…………….

Good Luck & Happy Trading
Markets are not efficient, rather they are effective - Jones
  • #6
  • Dec 27, 2015 10:13am Dec 27, 2015 10:13am
Quoting mzvega
{quote} Giovanni Maiani developed such a program, reporting his results in Stocks & Commodities Magazine in November 2002. He reports analyzing 575 American stocks over a period of 15 - 20 years. He found more than 6.14 million candlestick formations. In his words: "The most reliable, long lower shadow, is present 1.9% of time, and anticipates a declining session 49.12% of time; a rising session 40.61% of time. For traders who are quantitatively based, candlestick patterns are not terribly useful" Subsequently, Robert Colby, writing in SFO Magazine,...
Thank you Jones for the valuable information.

I do agree (to a certain degree) with those findings.
However what I am presenting here (open for discussion) is not about candlestick patterns (or bar patterns).

It is about bar Sequence patterns (or candlestick Sequence patterns). Sequence is the magic word here.

What does this mean and how does it differ from known candlestick patterns is the subject of this thread.

The sequence patterns discussed here include all known candlestick patterns Plus any sequence you can think of, even if its 20 candles long or more.

In post #1 I posted simple examples composed of 4 and 2 bars.

What I am interested in are more complex (repeating) patterns.
Using the terminology laid out in post #1 (btw: the coding system is open to members suggestions): example pattern could look like this :

  1. LIo--H : Translated to Bearish Break followed by a bearish Inside Bar followed by a Bullish Outside Bar followed by 2 bearish bars followed by Bullish Break. With such a sequence the system will try to predict the next bar based on statistical analysis (i.e. what is the probability the next bar will bullish or bearish)

I am going to provide the tools that will enable members to test the winning rate of trading different sequence patterns (best set with 1:1 RR).
Any pattern that can generate >52% win rate will be considered for further analysis.

This is not a trading system, its a research which may evolve into a trading system .

All the best,

Khalil

No guts, no glory
  • #7
  • Dec 27, 2015 10:38am Dec 27, 2015 10:38am
Quoting mzvega
{quote} Giovanni Maiani developed such a program, reporting his results in Stocks & Commodities Magazine in November 2002. He reports analyzing 575 American stocks over a period of 15 - 20 years. He found more than 6.14 million candlestick formations. In his words: "The most reliable, long lower shadow, is present 1.9% of time, and anticipates a declining session 49.12% of time; a rising session 40.61% of time. For traders who are quantitatively based, candlestick patterns are not terribly useful" Subsequently, Robert Colby, writing in SFO Magazine,...
Thanks for the PDF file. It contains a very good statistical analysis on candlestick patterns. However, I'm still interested in the tool that Khalil is going to kindly share with us. It may not reveal any profitable pattern at the end of the day, but it is definitely an invaluable tool for backtesting the performance of different patterns that one may come up with.

Best,
Math
Trading is the hardest way to make easy money...
  • #8
  • Edited 12:55pm Dec 27, 2015 12:33pm | Edited 12:55pm
A candle is defined by four price points (open, high, low, and close) within a defined time frame. That time frame can be any standard length of time, and a great deal of thought and study has be done on which time frame is the best for different trading styles. Any time frame can also be defined as a group of equal and consecutive sub-time frames as well, so that, for example, a 30 minute candle can be defined as a group of 6 consecutive 5 minute candles. The difference is that instead of just four price points for this 30 minute candle, there will be 24 different price points (4 price points X 6 candles) that define and make up the candle.

With that in mind, this ambitious project will hopefully define a set of significant candles, and their inner workings, that can give the astute trader an edge in their trading. I look forward to following your progress.

As we are close to greeting a brand new year, I hope that everyone can take a moment and reflect upon all the lessons learned in the past year, and that GREEN pips fill your accounts and lives in the new year.
  • #9
  • Dec 27, 2015 2:00pm Dec 27, 2015 2:00pm
Quoting abokwaik
The idea started small in Robots Lab, where I was initially interested in Up and Down bar sequence HERE, then Break High and Break Low bar sequence HERE. Here I am willing to further expand the idea of Bar Sequence Patterns to cover ALL types of bars. Terminology for pattern definition (could be changed as thread progresses): + : Up bar,Close > Open, regardless of body size - : Down bar, Close...
Hi Khalil,

I suggest that we add two more patterns:

1. G: (High > previous High) and (Low > previous Low) regardless of being Up/Down bar.
2. W: (High < previous High) and (Low < previous Low) regardless of being Up/Down bar.

Best,
Math
Trading is the hardest way to make easy money...
  • #10
  • Dec 27, 2015 8:37pm Dec 27, 2015 8:37pm
Trading should be simple. If not, dont trade. It will confuse you, if complicated.
  • #11
  • Dec 27, 2015 8:41pm Dec 27, 2015 8:41pm
Quoting antonsatu
Trading should be simple. If not, dont trade. It will confuse you, if complicated.
If you understand what you are doing, nothing is complicated.
MT4/MT5 EAs/Indicators/Alerts coder
  • #12
  • Dec 27, 2015 10:28pm Dec 27, 2015 10:28pm
Subscribed!

Would also be interesting to test the sequence forecasting in relationship with a single anchor candle.
Solutions ONLY!!!
  • #13
  • Dec 27, 2015 11:44pm Dec 27, 2015 11:44pm
Quoting MathTrader7
{quote} Hi Khalil, I suggest that we add two more patterns: 1. G: (High > previous High) and (Low > previous Low) regardless of being Up/Down bar. 2. W: (High < previous High) and (Low < previous Low) regardless of being Up/Down bar. Best, Math
Great Math , these will clearly represent Higher Highs & Higher Low in up trend, and Lower Highs & Lower Lows in down trend.
I will certainly add them, but I will change the coding to :
W : for HH, HL (uptrend)
M : for LH, LL (downtrend)

For better visualization, I hope its OK with you.
No guts, no glory
  • #14
  • Dec 27, 2015 11:51pm Dec 27, 2015 11:51pm
Is it meaningful to add also :
w: for Higher Lows followed by Higher Highs, and
m: for Lower Lows followed by Lower Highs


???
No guts, no glory
  • #15
  • Dec 28, 2015 1:13am Dec 28, 2015 1:13am
Quoting abokwaik
{quote} Great Math , these will clearly represent Higher Highs & Higher Low in up trend, and Lower Highs & Lower Lows in down trend. I will certainly add them, but I will change the coding to : W : for HH, HL (uptrend) M : for LH, LL (downtrend) For better visualization, I hope its OK with you.
Thanks Khalil!
And the coding is fine with me :-)
Trading is the hardest way to make easy money...
  • #16
  • Dec 28, 2015 1:15am Dec 28, 2015 1:15am
Quoting abokwaik
Is it meaningful to add also : w: for Higher Lows followed by Higher Highs, and m: for Lower Lows followed by Lower Highs ???
I think I didn't get the logic behind "w" and "m". They seem to be the same as "W" and "M"
Trading is the hardest way to make easy money...
  • #17
  • Dec 28, 2015 1:26am Dec 28, 2015 1:26am
Quoting MathTrader7
{quote} I think I didn't get the logic behind "w" and "m". They seem to be the same as "W" and "M"
my bad, they are the same.
No guts, no glory
  • #18
  • Dec 28, 2015 1:47am Dec 28, 2015 1:47am
Quoting abokwaik
{quote} my bad, they are the same.
And I have a question regarding the terminology you are using. What is the difference between an Outside bar and an Engulfing bar in your definition? Here and Here there are two pages explaining Engulfing bars where there is no difference between them and outside bars to my knowledge. However, I guess you are using your definition of Engulfing bars as you did in your Engalfer EA.
Trading is the hardest way to make easy money...
  • #19
  • Dec 28, 2015 1:56am Dec 28, 2015 1:56am
Quoting MathTrader7
{quote}What is the difference between an Outside bar and an Engulfing bar in your definition? Here and Here there are two pages explaining Engulfing bars where there is no difference between them and outside bars to my knowledge.
The difference in the engulfing is the body must engulf the previous body ... while it is not necessary in the outside bar.
MT4/MT5 EAs/Indicators/Alerts coder
  • #20
  • Dec 28, 2015 2:01am Dec 28, 2015 2:01am
Quoting FerruFx
{quote} The difference in the engulfing is the body must engulf the previous body ... while it is not necessary in the outside bar.
That difference makes sense. However, it seems that outside bar is also known as engulfing bar in the terminology of some traders.
Trading is the hardest way to make easy money...
0 traders viewing now