There are already some threads on Fibonacci methods and harmonic trading. I am starting this one to focus on application of these methods to trading.
"Sure, if you draw enough lines on your chart, price is going to hit one of them"
Fibonacci Retracements, Expansions and Extensions are based on the Golden Ratio. Is this just something in vogue? A buddy of mine, who does not believe in Fibonacci lines, says: "Sure, if you draw enough lines on your chart, price is going to hit one of them" . So for him Fibonacci based trading is either collective self-delusion or at best a self-fulfilling prophesy feeding on herd behavior.
Does Fibonacci Trading give you an edge?
To find out whether the Fibonacci levels provide an edge, I have divided swings into equal intervals of about 2.2% and calculated a probability distribution for the retracements, extensions, etc. of the following swings, to find out whether there are some high probability levels. The absolute count did not yet give me a probability. As an example, to calculate the conditional probability for a reversal near a 127.2% extension line, I have to divide the reversal count by the number of all reversals that touch the 127.2% interval or exceed it.
Results from evaluating distribution of swing sizes
The 100% line has the strongest impact both in absolute and relative terms. This reflects conventional wisdom, as prior swings are known to create support and resistance. In absolute terms this is followed by lower retracement levels of 61.8%, 50% and 78.6%. However, in terms of relative probabilities, the extension levels have a higher success rate than the retracements. Do Butterfly patterns have a higher probability than Gartley patterns?
Combinations of different Fibonacci Levels increase the edge
Now if an extension is close to a retracement on a higher time frame, these probabilities can be reinforcing. After running the statistical distributions, I believe that fibonacci cluster do provide an edge that can be exploited. Fibonacci cluster are leading indicators, they establish support or resistance levels, before the price approaches them. So they belong to the same family as pivots, trend lines, trend channel lines, moving averages channel lines and moving averages (these are usually lagging, but leading if used as S/R). Cross support or resistance is when several of these methods reinforce each other. So if you use fibonacci cluster together with trend lines or tredn channel lines, you will get something like a cross-cross-support or resistance.
Volume Analysis is needed to confirm valid cluster zones
Volume is the key indicator to verify whether S/R lines will be respected or ignored by the market. So S/R indicators should be combined with a BetterVolume or VolumePattern indicator to check their validity. Volume is not only about smart and dumb money. Higher volume indicates the "Other Timeframe Participation" - a term I borrow from James Dalton. Volume indicates that large or institutional traders, who are watching large swings are participating in the game, when smaller daytraders are focussing on smaller swings. The larger swings cannot be observed on your 5 min chart, they took place during the last weeks and months, but you should know where these points are.
Timeframe transfer, at what levels do larger traders enter the game?
This means transferring larger timeframe signals to your intraday chart and display them there, because that is where the larger traders will come into play. My first attempt to do this was the indicator that I attach below. It uses first and second order swing lows and highs, which is definitely not an elegant approach, but it works. It collects Fibonacci retracements and extensions from up to one year back and displays them on intraday charts. I am currently working on a more advanced concept that uses a zigzag indicator as a starting point, and makes an analysis of 6 different types of harmonic lines over 10 different time frames, but this is still work in progress.
How to trade Fibonacci Cluster?
Here are some suggestions, I will detail this later.
(1) Do not use them to catch a falling knife!
(2) Trade them as support and resistance on boring days with no important news
(3) To etablish potential exits for a 3rd wave
(4) To trade 2B Patterns (Victor Sperandeo) after news announcements: the Sting
(5) To locate the end of the two-legged correction within a Gartley pattern
(6) To establish Risk-to-reward ratios, when entering trades.
Cluster Area 2 was ignored by the market. How do youfind out that the cluster is ignored by the market? The key is volume. If Cluster 2 was a valid support area, you should have noticed some buyers stepping in creating at least one high volume churn bar. However the chart only shows the expansion bars (light blue volume bar), but no churning that follows. So you would qualify the blue bars as a breakaway and not as climax bars. Cluster Area 2 is then too close to the original trading range.
Cluster Area 3 stops the main price action. Just above this cluster some churning occured (yellow vollume bar) after a weak climax (dark blue volume bar). The churning often is the starter for the accumulation phase, so it is the first signal that is generated and leaves you lots of time to watch the market. After the churning expect a retest of the low or high. On the chart you will see a Rickshaw Man (long leeged doji) climax bar that closes right at the 76.4% retracement line. The high volume (magenta bar) and the range indicate a climax churn bar. This is the bar that defines the final value range of the move prior to a reversal.
Cluster Area 4 was only identified by the confluence indicator (work in progess). This indicator, however did not catch the trading range of the morning. For those who want to know, how the last support line was generated: It is basically the 1.382 downward projection of the trading range that formed during June and July last year, reinforced by some other lines to produce the necessary weight to act as a confluence zone.
Last edited by Fat Tails; March 21st, 2010 at 02:57 PM.
The following 55 users say Thank You to Fat Tails for this post:
Hey Fat Tails (is that a description or a preference...? )
I like what you are doing here. Take a look at this indie. You can use multiple instances of this to accomplish what you are doing. This has been a favorite fib indie of mine for awhile. For NT6.5 - haven't tried it on NT7. You can easily add one for the weekly, one for the daily, monthly, etc. - Maybe the code will give you some other ideas.
“Be who you are and say what you feel because those who mind don't matter and those who matter don't mind.” - Dr. Seuss
The following 10 users say Thank You to websouth for this post:
tried your indicator. If I apply it 10 times and select different time frames, I will get about the same as for the cluster indicator, except that it does not catch hidden highs and lows. My main intention was to automate the search for prior swing highs and lows. I am now working on an indicator creating a collection of zigzags, each having a different minimum deviation and scanning the price series for different Fibonacci levels. To name them:
(1) Prior Highs and Lows
(3) Extensions: Retracements of more than 100%
(4) Alternates: Measured moves after a correction, the target swing is a Fibonacci multiple of the first swing
(5) Projections: Extensions of the range generated by a prior swing
I have attached some charts to show again what these are, because it is easier to recognize them from a chart than to explain what it is.
Now, if you draw all the different levels for ten different time frames, you will get up to 540 lines. Because it is impossible to trade 540 lines, I want to calculate probabilities for each line based on the disstribution above, and based on a weight attached to each timeframe -> larger timeframes will be more important than smaller timeframes. Then you add up all the weights for the lines within a given window. This window is then verrtically moved over the chart to identify zones, where severak major lines add up. These are the confluence zones. They are high probability zones to provoke a reaction. To catch these, it is important not to leave out any timeframes by defining appropriate values for the minimum deviatations of the zig zag.
Last edited by Fat Tails; March 21st, 2010 at 03:50 PM.
The following 15 users say Thank You to Fat Tails for this post:
my exp suggest fib extensions works much better than fib retracements. fib confluence/clusters are definitely tradeable and its better to short the resistance rather than buying the support (catching the falling knife).
A reversal is like a relationship, you need to go through a number of stages before getting married to a position. Reversals take more time than pullback trades. I believe that many traders - including myself - suffer from unnecessary losses by trading reversals too early. Our brain is trained to complete fragmentary patterns and unfortunately has a clear preference for reversal patterns as opposed to trend continuation patterns. Markets usually have the opposite preference, as by design they will trap a maximum number of traders.
First of all, an analysis of the larger picture is needed and an area that might trigger a reversal should be spotted. For FDAX this morning there was a Fibonacci Confluence Zone. On the chart below you will find the different stages through which a reversal needs to go.
This is the screenplay with its 10 chapters.
The Reversal: A Hate and Love Story around a Fibonacci Line
The Churn Bar - The Climax Bar - The Final Trading Range - The Trap - The Breach of The Trendline - Filling the Trading Range - Retesting The Trendline - The Expansion Bar - Conquering the Moving Average - Triggering the Stops - A Trend is Born
The Churn Bar
First thing to wait for is a churn bar. The churn bar is the harbinger of the reversal. For a reversal to become effective, confirmation by other signals is required. For FDAX below the churning occured just above the midpivot level S12, when some traders closed their short positions early and new longs stepped in.
The Climax Bar
The climax bar is the final extension of the value area of the auction. If the climax bar is the last one of the prevailing trend, it is likely to create a final trading range that will last for quite a while. For FDAX below the climax bar shows that the last old longs have surrendered.
Sometimes the churn bar occurs before the climax bar, sometimes it is the other way round. Under any circumstances, do expect a false breakout from the trading range in the direction of the old trend.
The Final Trading Range
It is now the duty of the price to test the boundaries of the trading range created by the climax bar. Within the trading range the following events may occur to confirm the reversal.
The trap is false brekaout from the trading range on medium high volume. It often appears as a Sting. For FDAX it can be easily identified. Its purpose is to trigger the stop losses of weak new longs and equally trap new shorts. The Sting cleans the market from bugs and vermin. The Sting may also occur at a later stage, even after the retest of the high or low.
The Breach of the Trendline
As long as the trendline has not been penetrated, the old trend is still intact. I use a linear regression indicator as a substitute for a trendline. For FDax below the trendline breach occurs, when the blue line is penetrared.
Filling the Trading Range
Prices must first move to the opposite side of the trading range generated by the climax bar. This often generates a nice pullback trade near the EMA(20). FDAX below has a low volume doji bar, which indicates that prices are going to retrace. The price stopped just 1 tick below the high of the climax bar. This indicates weakness, as nobody bought the market to trigger the stops that lie just above the climax bar.
Retesting the Trendline, Low or High
For a retest of the high or low, prices will likely move back to the linear regression line. if the retest does not stop with a reasonable margin of the prior high or low, the potential reversal was not confirmed. For FDAX below the test is indicated again by a yellow high volume churn bar and a climax bar. Both, churn and climax bars have lower volume than the original churn and climax bars. This is required for a successful test.
The Expansion Bar
The expansion bar confirms that the market accepted the test as successful. For FDAX the expansion bar was generated just after the Doji formed on the lower end of the trading range.
This is by no means a reversal, all the price action observed until now could also describe transition from a trend to a trading range. For FDAX the bears still exerce control, although they have suffered a severe hit, after the successful test of the prior low and the expansion bar. There are still two defense lines that need to be overcome, the EMA (period between 20 and 30) and the opposite end of the trading range.
Triggering the Stops
Once the trading range range has been conquered, the stops need to be triggered. For FDAX the bulls will only be in control if they shift the auction from the value area, where it is established now - the trading range defined by the climax bar - back to higher levels. We wait for a green auction bar to print. The chart shows all bars that shift value to a new value area as filled candles, those bars that close inside a preestablished value area are shown as hollow candles.
A Trend is Born
The first auction bar that shifts value away from the final trading range starts the new trend.
FDAX below did not manage a reversal prior to 2 PM CET. In the afternoon it will be influenced by the US markets. Consider this as a separate trading day. I am afraid that there is no happy end for this screenplay today.
Last edited by Fat Tails; March 22nd, 2010 at 09:12 AM.
The following 38 users say Thank You to Fat Tails for this post:
Well put , thanks for this post . Im with you 100 % about the focussing on reversals thingy . Its a psychological flaw linked to wanting to be right and first in line IMHO . Cant ignore reversal signs and cant jump on every reversal signal and miss the stronger continuations . Too bad you can only make one decision at a time and stick with it but Im learning to ignore reversals and only validate them when they are confirmed on several levels .
The following user says Thank You to Eric j for this post:
Nice posts, I like the what you're doing -- keep it up
Due to time constraints, please do not PM me if your question can be resolved or answered on the forum.
Need help? 1) Stop changing things. No new indicators, charts, or methods. Be consistent with what is in front of you first. 2) Start a journal and post to it daily with the trades you made to show your strengths and weaknesses. 3) Set goals for yourself to reach daily. Make them about how you trade, not how much money you make. 4) Accept responsibility for your actions. Stop looking elsewhere to explain away poor performance. 5) Where to start as a trader? Watch this webinar and read this thread for hundreds of questions and answers. 6) Help using the forum? Watch this video to learn general tips on using the site.
If you want to support our community, become an Elite Member.
The following 2 users say Thank You to Big Mike for this post:
The reversal finally came, but much later than expected. Fibonacci Lines are only half of the truth, the final reversal point was the pivot level S2. I did not have pivots on the first chart, but the picture would be somehow incomplete, if I did not mention S2.
S2 is an obvious point to look for reversals - see also the Reaction Trend Rules in J.Welles Wilder's book "New Concepts in Technical Trading Systems" from 1978.
So the final range occured one step lower, between the Fibonacci Cluster line and the Pivot level S2. The reversal was only confirmed, after price broke through the Fibonacci Line showing the first auction bar up, which confirmed an upward shift in value.
Since this morning FDAX has travelled back to the upper confluence zone and closed the gap from Friday. A collection of churn bars indicates that a distribution pattern may have started.
The following 5 users say Thank You to Fat Tails for this post: