Fibonacci Clusters & the Power of Confluence
by: Bernard Mitchell

Leonardo Pisano, better known by his nickname Fibonacci, published Liber abaci in 1202 in which he posed the following problem: fibonacci sequence in nature

A certain man puts a pair of rabbits in a place surrounded on all sides by a wall. How many pairs of rabbits can be produced from that pair in a year if it is supposed that every month each pair begets a new pair, which from the second month on becomes productive?

The resulting sequence is 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, ... each number in the series is the sum of the two numbers preceding it. He also discovered that each number in the sequence is 1.618 times the preceding number and .618 of the next number.

Many market technicians use the Fibonacci sequence numbers when trying to determine support and resistance, and commonly use .382, .50, and .618 fibonacci retracements.

It is commonly believed that .382 fibonacci retracement lines from a trend move will tend to imply a continuation of the trend. A .618 retracement implies that a trend change may be in the making.  Market technicians have adopted many such rules. My approach is different in that I do not use Fibonacci retracement numbers to determine the trend.  When trading trends I use my long-term which is determined by a simple 89-period moving average. It is effective from a one-minute chart to a monthly chart, any market or time frame trend trade. This moving average is the literal battleground between the stock market bulls and bears.

The market technician who uses Fibonacci Clusters will determine support and resistance in the direction of the main trend (which is always determined by prices either above or below the long-term moving average)—it is what I call my "Main 89 Line." Most charting packages will enable you to create a simple 89-period moving average. I encourage you to look at the market this way in order to trade trends efficiently.

What are Fibonacci Clusters and how are they different in calculating support and resistance than the standard ratios with which most traders are familiar? Simply put, Fibonacci Clusters are more accurate because of the use of more data points and the way the Fibonacci ratios are calculated. I include as many as seven swing highs and swing lows in addition to the significant high points and significant low points in the calculations.

Let’s review the approach I use to calculate support and resistance levels. The first thing I need to identify is a trend move, i.e. a significant high point to a significant low point or vice versa. These points are the extreme boundaries of the trend.

Within these extreme points are contra-trend swing highs within an overall downtrend and swing lows within an overall uptrend. In other words, I look to "capture" the high/low points of a contra-trend move within the longer trend.

These contra-trend high/low points are swing points and are critical to my calculation of Fibonacci Clusters.

The normal retracement levels of .382, .5, and .618 of the main trend, are simply not accurate enough when the pinpointing of support and resistance becomes crucial to pulling money out of the markets.

First, it is important to understand the concept of support and resistance. In this writer’s opinion, unless you can determine precise support and resistance, most traders will achieve mediocre results at best.

Simply put, support means there is sufficient buying pressure (stock market volume) at a given price level to halt a downtrend and resistance means there is sufficient selling pressure (stock market volume) at a given price level to halt an uptrend.  These market volumes place boundaries in the market place, support places a floor under the market and resistance places a ceiling above the market.

The higher the trading volumes, relative to the range of the price bar, the more significance is given to the price level as support and resistance. I measure this in terms of ticks in a price bar. A greater number of transactions (ticks) within a given time frame (and this can be anything from a one-minute price bar to a monthly bar) when divided into the range of the bar and compared with the previous two price bars, with the mathematical result being lower, will produce a robust squat with high predictive power for market reversals.

The squat bar is my trigger to enter a trade. The only thing I now need to know is where support and resistance will likely take place.

This is where my calculation of Fibonacci Clusters comes in. The resulting output is four levels of support and four levels of resistance for indentifying changing trends. They are:

• Lower Resistance
• "Stopping Point" Resistance
• Upper Resistance
• Maximum Resistance
• Upper Support
• "Stopping Point" Support
• Lower Support
• Maximum Support

Each of first three levels of resistance and first three levels of support are tradeable levels, as long as a squat is generated on the entry bar. The maximum resistance and support levels are normally not used for entries—unless they are a "confluence" number from a larger or smaller trend. There will be more on this later.

Prior to the development of software, it was time consuming to create these levels of support and resistance, except on daily and weekly data. We can now produce reliable, tradeable numbers on intra-day data even when highs and lows are constantly changing.

How to calculate support and resistance

To Find Resistance
An example of resistance would be that the market has made a significant high point (SHP). Now the market proceeds to correct, contrary to the main trend and puts in a series of swing highs (SH), finally registering a significant low point (SLP).


Observing the chart, you can easily identify the highest high (SHP) and the lowest low (SLP) in the trend and a series of swing highs (SH) that fall between the two extreme points. The selection of swing highs is straightforward. There needs to be a minimum of three bar highs.

To Find Support
An example of support would be that the market has made a significant low point (SLP). Now the market proceeds to rally, contrary to the main trend and puts in a series of swing lows (SL), finally registering a significant high point (SHP).

Observing the chart, you can easily identify the highest high (SHP) and the lowest low (SLP) in the trend and a series of swing lows (SL) that fall between the two extreme points.

The selection of swing lows is straightforward. There needs to be a minimum of three bar lows.

We will now look at some examples of Fibonacci Clusters and apply the calculations to different time frames (i.e. weekly, daily and intra-day).



Weekly—ATML CORP (ATML) - - calculating support and resistance
This stock had a huge bull run and peaked on 3/24/00 at 30.688. It then proceeded to correct, making several important lows, down to its recent low of 7.625 registered on April 6, 2001. Prior to that, a low was made on December 1, 2000 at 9.375. To project resistance from that point, I selected a series of three strength weekly highs starting from 24.406, 21.938 and 16.625. Using SHP 30.688 and SLP 9.375, along with the three swing highs, the result for the changing trend was:

Lower Resistance: 16.26
Stopping Point: 18.62
Upper Resistance: 21.51
Maximum Resistance: 25.75

Note that the market then proceeded to stage a choppy rally over 10 weeks, peaking at 18.438; just shy of our stopping point, 18.62. Over the next 10 weeks a new low was made at 7.625.

Using the new low as my new SLP, and using the 18.438 rally high as my new SH3, a new series of resistance numbers have now been projected. Since this is in the future, we need to wait and see how these numbers eventually work out.

Observing the output from the two trends, each with a different SLP, i.e. 9.375 and 7.625, a series of confluence resistant points have been established. They are:

16.26 and 16.65—Lower Resistance
18.62 and 18.64—Stopping Point
21.51 and 21.07—Upper Resistance

At these three levels, shorts or exits to previously held long positions can be taken, provided a squat bar is generated on the daily charts.



Daily—Texas Instruments (TXN) - - calculate support and resistance levels
After rallying off a low, the market peaked at our daily 89-period moving average at 42.91 on May 22, 2001. It then broke sharply to 28.25 where it created a SLP. Between the two extremes we find SH1 at 39.39 and SH2 at 32.95. Thus, using the four inputs the result is:

Lower Resistance: 32.94
Stopping Point: 34.22
Upper Resistance: 35.76
Maximum Resistance: 39.77

Going one step further in attempting to find confluence, I used the SH1 as my SHP. The result:

Lower Resistance: 31.55
Stopping Point: 32.49
Upper Resistance: 33.54
Maximum Resistance: 37.01

You will note that the lower resistance (32.94) on the longest trend comes close to the stopping point (32.49) on the shorter trend and stopping point on the longer comes fairly close (not perfect) to upper resistance on the shorter trend. Observing the chart, the high point of 34.01 falls between these two numbers, before the market started a significant sell off.

WorldCom (WCOM) - - calculate support and resistance
Starting from a SLP on March 14, 2001 at 15, the market rallied establishing an SHP on April 26 at 21.52. The market then proceeded to correct off the highs creating a 15-day low at 18.10, penetrating our stopping point by just 13 points (18.23 - 18.10 = 13).



To find confluence, I used the SLP of 16.125 with the same SHP. This produced important support between 17.37 to 17.28. As you can see, the market spent 26 days trying to hold this support level, before finally succumbing in a waterfall decline to 13.



Intra-day—September S&P Futures 135-minute Chart
The SLP was established on June 26 at 1211. The SHP was established on June 5 at 1297.80. Using the SH1 at 1289.50, SH2 at 1272, SH3 at 1249.30 and SH4 at 1240, the calculations are as follows:

Lower Resistance: 1244.84
Stopping Point: 1254.03
Upper Resistance: 1265.42
Maximum Resistance: 1279.23

From the SLP low, the market rallied to 1244.90 eight bars later, just shy of lower resistance. A series of squat bars showed up as the market spent nine more bars before finally tanking. Confluence numbers were also created showing important resistance around 1244. Now, (July 6) with new lows being made, a new series of resistance levels can be calculated.



-E-Mini S&P—610 ticks
Fibonacci clusters are most advantageous for day traders looking to scalp profits on small moves using the benefits of electronic execution. I have created resistance levels from the huge downtrend that started with the SHP at 1247.25 on July 3 at 6:00 AM E.T., culminating in a SLP at 1189 on July 9 at 10:04 AM ET

Using seven swing highs as my inputs along with SHP and SLP the resistance levels are as follows:

Lower Resistance: 1214.04
Stopping Point: 1219.81
Upper Resistance: 1226.45
Maximum Resistance: 1234.79

In addition, I have attempted to find confluence numbers by using two different SHPs, thus calculating from smaller trends. Those resistance levels come in at:

Lower Resistance: 1209.43
Stopping Point: 1214.44
Upper Resistance: 1220.03
Maximum Resistance: 1229.87

and:

Lower Resistance: 1205.13
Stopping Point: 1209.52
Upper Resistance: 1214.35
Maximum Resistance: 1222.60

The very first thing to notice is numbers have confluence around 1214.04 (lower resistance on the largest trend). They are 1214.44 (stopping point next largest trend) and 1214.35 (upper resistance smallest trend). Average these three numbers and arrive at 1214.27.  Trends moves

The series of confluence numbers on the big moving trend come in around 1220.75. They are 1219.81 (stopping point largest trend). Those numbers average out to 1220.82.



The third series of confluence numbers on the big moving trend come in around 1228.25. They are 1226.45 (upper resistance largest trend) and 1229.87 (maximum resistance smaller trend). They average 1228.16.

Shorts can be taken at all three levels as long the entry bars are accompanied by squats. The point to remember: this resistance will be constantly updated in real time as new lows are made.

After more than 20 years of working with the natural numbers that make up the Fibonacci sequence, Fibonacci clusters with the power of confluence come closest to pinpointing support and resistance.

— end


Copyright © 2004 SFO Magazine All rights reserved.
Reproduction in whole or in part without permission is prohibited.