AUTHOR

johnehlers.jpg
Name:
John Ehlers



John is an Electrical Engineer. He received his BSEE and MSEE from the University of Missouri and did his doctoral work at The George Washington University, specializing in Fields & Waves and Information Theory. He has retired as a Senior Engineering Fellow with one of the largest Aerospace Companies. He has been a private trader since 1976, starting with fundamental analysis. With his engineering training he quickly gravitated to technical analysis of the market. He originally questioned what was magic about a 14 day RSI, or any other period. He concluded there was no unique answer and that one should adapt to current market conditions by using the measured cycle. He popularized the measurement of cycles with his MESA program. John has written extensively about technical trading and has spoken internationally on the subject. His approach is unique. Any technique must first work on theoretical waveforms before testing against real-world data is attempted.

Categories

July 23, 2008

Trend / Cycle Heatmap

23 July 2008
Look at all that red! The recent Vee bottom is interpreted by the heatmap as a piece of a short term cycle.

TCHeatmap.jpg

Trend/Cycle Heatmap
The Trend/Cycle heatmap is similar in concept to the Hurst Coefficient with the added advantage that it shows uptrends and downtrends. This indicator measures the strength of the trend across the lookback period compared to the cycle amplitude within that same period. Uptrends are shown in green, downtrends are shown in red, and blue denotes the cycle amplitude is stronger than the trend. This indicator varies with the length of the lookback period. In the display, a horizontal slice through the heatmap denotes the relative trend and cycle amplitudes at a fixed lookback period. A vertical slice through the heatmap shows the relative amplitudes over a range of lookback periods for the particular selected price bar.

July 22, 2008

Signal to Noise Ratio

22 July 2008
The Signal to Noise Ratio is so low that any cycles-based signals should be ignored.

Amplitude.jpg

Signal to Noise Ratio
The amplitude of the dominant cycle is normalized to the amplitude of the noise. Noise is defined as the average daily trading range. Unless the cycle amplitude is much greater than the noise, the use of cycle-based entries and oscillators should be avoided because the uncertainty of getting a good entry and exit point during day can negate any potential profit realized from the cyclic swing.

July 21, 2008

Spectrum

21 July 2008
The turning point from the end of the long downtrend is probably interpreted by the software as the low of a long cycle. You will see tomorrow that the Signal to Noise Ratio is still very low, and therefore very little credence should be place on the spectral measurement.

Spectrum.jpg

Spectrum
We measure all the cycle components in the market data over a range from 6 bar cycles to 30 bar cycles. Continuity of the cycles can be seen in the heatmap that is in sync with the barchart above it. Amplitudes of the cycles are colorized from white hot (yellow), through red hot (red), to ice cold (black). Knowing the dominant cycle period is useful for making conventional indicators adaptive to current market conditions by setting their computational window to be the dominant cycle period or half the dominant cycle period, depending on the indicator being used.

July 18, 2008

Hurst Coefficient

18 July 2008
The recent upturn from the long downtrend has caused the data to be random over the shorter lookback periods.

Hurst.jpg

Hurst Coefficient
The Hurst Coefficient is a measure of randomness in the market, and varies over a range from 0 to 1. 1 signifies the data are persistent. 0 signifies the data are antipersistent (cyclic). A value of 0.5 signifies the data are random. I have colorized the Hurst Coefficient to be green at a value of 1, red at a value of 0.5 and blue at a value of 0. The colors are mixed to show the gradual transition of the Hurst Coefficient. The Hurst Coefficient is a function of the lookback period used to compute it. In the display, a horizontal slice through the heatmap denotes the Hurst Coefficient at a fixed lookback period. A vertical slice through the heatmap shows the Hurst Coefficient over a range of lookback periods for the particular selected price bar.

July 16, 2008

Signal to Noise Ratio

15 July 2008
Since the market has been in a pure downtrend, it is no surprise that the cycle amplitude is low to nonexistant. Therefore, the Signal to Noise Ratio is very low - the Noise has remained about constant.

Amplitude.jpg

Signal to Noise Ratio
The amplitude of the dominant cycle is normalized to the amplitude of the noise. Noise is defined as the average daily trading range. Unless the cycle amplitude is much greater than the noise, the use of cycle-based entries and oscillators should be avoided because the uncertainty of getting a good entry and exit point during day can negate any potential profit realized from the cyclic swing.

July 14, 2008

Spectrum

14 July 2008
Since the market has been in nothing but a downtrend recently, the cyclic content is virtually nonexistant. A few mixed spectral lines are showing, but there is no real dominant cycle.

Spectrum.jpg

Spectrum
We measure all the cycle components in the market data over a range from 6 bar cycles to 30 bar cycles. Continuity of the cycles can be seen in the heatmap that is in sync with the barchart above it. Amplitudes of the cycles are colorized from white hot (yellow), through red hot (red), to ice cold (black). Knowing the dominant cycle period is useful for making conventional indicators adaptive to current market conditions by setting their computational window to be the dominant cycle period or half the dominant cycle period, depending on the indicator being used.

July 11, 2008

Hurst Coefficient

11 July 2008
The Hurst Coefficient joins our Trend/Cycle Heatmap as showing nothing but a downtrend at all lookback periods.

Hurst.jpg

Hurst Coefficient
The Hurst Coefficient is a measure of randomness in the market, and varies over a range from 0 to 1. 1 signifies the data are persistent. 0 signifies the data are antipersistent (cyclic). A value of 0.5 signifies the data are random. I have colorized the Hurst Coefficient to be green at a value of 1, red at a value of 0.5 and blue at a value of 0. The colors are mixed to show the gradual transition of the Hurst Coefficient. The Hurst Coefficient is a function of the lookback period used to compute it. In the display, a horizontal slice through the heatmap denotes the Hurst Coefficient at a fixed lookback period. A vertical slice through the heatmap shows the Hurst Coefficient over a range of lookback periods for the particular selected price bar.

July 10, 2008

Detrend

10 July 2008
The downtrend is showing signs of weakening. Look for a reversion to to mean.

Sigma.jpg

Detrend
The detrend indicator is created by filtering the price data in a high pass filter, thereby removing the low (trending) frequency components. The detrended data is rescaled relative to the +one Sigma and –one Sigma variations over the last dominant cycle. A price swing outside the one Sigma points usually signals a reversion to the mean.

July 9, 2008

Trend / Cycle Heatmap

9 July 2008
There is no surprise that the Trend/Cycle Heatmap only shows a downtrend for all lookback periods.

TCHeatmap.jpg

Trend/Cycle Heatmap
The Trend/Cycle heatmap is similar in concept to the Hurst Coefficient with the added advantage that it shows uptrends and downtrends. This indicator measures the strength of the trend across the lookback period compared to the cycle amplitude within that same period. Uptrends are shown in green, downtrends are shown in red, and blue denotes the cycle amplitude is stronger than the trend. This indicator varies with the length of the lookback period. In the display, a horizontal slice through the heatmap denotes the relative trend and cycle amplitudes at a fixed lookback period. A vertical slice through the heatmap shows the relative amplitudes over a range of lookback periods for the particular selected price bar.

July 8, 2008

Signal to Noise Ratio

8 July 2008
The reason the Signal to Noise Ratio is so low is that there just is no cycle amplitude to speak of.

Amplitude.jpg

Signal to Noise Ratio
The amplitude of the dominant cycle is normalized to the amplitude of the noise. Noise is defined as the average daily trading range. Unless the cycle amplitude is much greater than the noise, the use of cycle-based entries and oscillators should be avoided because the uncertainty of getting a good entry and exit point during day can negate any potential profit realized from the cyclic swing.