Chart Interpretation

Tom’s Tool Box

Every technician needs to have more than one tool in his toolbox. This is because some tools only work in certain types of markets, and every technical indicator will be wrong some of the time. So it is useful to follow multiple indicators to get the best sense of what the market is really doing.

We have evaluated many of the indicators that have been created by others, and we have innovated several of our own. Here is a list of the ones we count on the most in our analysis.

Want to know more? Our extensive 3hr Advanced Topics and Indicators Seminar is now available on DVD.

Conventional Bar Chart Analysis
Moving averages
Oscillators
Bollinger Bands
Trend Indicator
Correlation
Relative Strength
Permutations of other indicators

Ratio-Adjusted Breadth Numbers
Seasonality
Cycles
Sentiment Indicators
Price Pattern Analogs
Liquidity Waves
Timing Models
Goldollar Index
Fibonacci Ratios

 

Conventional Bar Chart Analysis

  • uptrends and downtrends
  • rounding bottoms
  • head and shoulders
  • flags and pennants
  • volume confirmation
  • support/resistance levels
[ top ]

Moving averages

  • Simple moving averages (SMAs) are better for some things.
  • Exponential moving averages (EMAs) are better for other tasks.
  • Weighted moving averages (WMAs) are limited in application, but are used for calculating the Coppock Curve (see below).
  • EMAs are a type of WMA, because the more recent data count more than older data. But when technicians refer to a WMA, there is a specific method of weighting each day's data that is usually implied by that reference.

Quasi moving averages

These are things which behave like MAs and which can be used for chart analysis as one would use an MA, but which are based on other math. Examples are the Price Oscillator Unchanged line, the Coppock Unchanged line, and the Sum/10 line, discussed below.

[ top ]

Oscillators

Background

All oscillators share 3 main properties:

  1. A neutral level, toward which the oscillator will return if the data stop changing
  2. Overbought/oversold indications, or perhaps better stated as "extended readings"
  3. Divergences relative to prices

 

The McClellan Oscillator is an oscillator which is based on measuring the acceleration in breadth data, e.g. Advances minus Declines, or Up Volume minus Down Volume. By the definition above, the Summation Index is also an oscillator, even though it is calculated by totalling all previous values of the McClellan Oscillator.

Oscillators we use:

  1. McClellan Oscillator (A-D, and Volume, plus newer permutations such as dollar volume)
  2. Summation Index (again, with different permutations)
  3. Stochastics
  4. Up-Down Oscillators (how many up closes out of the past XX days?)
  5. Trend Indicator (percent deviation from Sum/10 line)
  6. Percent deviation oscillator (how far from a moving average)
  7. McClellan Price Oscillator (10% Trend minus 5% Trend of closing price)
  8. Short Term Price Oscillator (STPO), 50% Trend minus 20% Trend of closing price, or of intraday high or low
  9. Coppock Curve
    • Developed by E.S.C. Coppock in the 1960s, who called it his Very Long Term (VLT) Momentum Index.
    • Looks at the 11-month and 14-month rate of change (ROC) in the DJIA, smoothed with a 10-month WMA.
    • Based on the insight from experts in the Episcopal Church (for which he managed assets) that the period of grieving over a lost loved one is typically 11-14 months. Grief over lost assets works on the human mind in a similar way.
    • Coppock Unchanged Line is the theoretical price, above or below which a monthly close would cause a change in direction of the Coppock Curve.
  10. Sum/10 Line
    • Calculated by first creating a Summation Index of all previous Price Oscillator values, then dividing the result by 10 to it put back into the realm of where current prices are.
    • Acts like a long term moving average, but with some interesting properties
[ top ]

Bollinger Bands

Useful for indicators whose range tends to change over time, such as VIX (50-day, 1-sigma works well) and some options data.

[ top ]

Trend Indicator

(see Oscillators above)

  • Measures percent deviation of price away from the Sum/10 line.
  • Useful with a 5.8% Trend of the Trend Indicator.
[ top ]

Correlation

DJIA Correlation Index examines the Pearson's Correlation Coefficient of each DJIA stock against the DJIA itself, then averages together each of the 30 stocks' values into one indicator. This index rises as the DJIA is moving in the direction of the dominant trend, either up or down. If we see what direction prices are moving in during a rising or falling index phase, that tells us which is the dominant trend direction.

[ top ]

Relative Strength

Take one price series and divide it by another, to see which is stronger. For example, take the share price of Microsoft, and divide it by the Nasdaq Composite. If the relative strength line rises, then that means Microsoft's share price is acting relatively stronger than the overall Nasdaq market. This can be because it is rising faster, or falling more slowly.

[ top ]

Permutations of other indicators

  • 50-day rate of change (ROC) of the Summation Index (10-day ROC is also useful)
  • DJ Oscillator Rising Index
    • What percentage of the 30 DJIA stocks have a rising Price Oscillator?
    • A positive Price Oscillator?
[ top ]

Ratio-Adjusted Breadth Numbers

The changing number of issues traded on the NYSE, and rising trading volume, necessitates doing some type of adjustment to the data to make meaningful long-term comparisons.

Ratio-Adjusted Summation Index (RASI) is particularly meaningful for confirming or refuting "escape velocity" of a new uptrend. If it moves from an oversold reading to well above +500, that shows the requisite strength for a new utprend to continue higher. Rolling over before reaching +500 says that there is not enough energy, and that a retest of the prior low is likely.

[ top ]

Seasonality

  • Annual seasonal patterns, as described in great detail by Yale Hirsch, Sy Harding, and others seasonal pattern.
  • January barometer
[ top ]

Cycles

Analysis of cyclic behavior of markets is inherent in all other indicators

9-month (AKA 40-week)

The 9-month cycle is important for stock prices, as is its half-period harmonic the 20-week cycle. The actual length of the "40-week" cycle has historically been more like 38.5 weeks, and most recently appears to have contracted to less than 35 weeks, for reasons yet to be discovered.

13-1/2 month cycle

Important for gold prices

4-year cycle

Important for stock prices (see Presidential Cycle Pattern)

 

40-year cycle

40-Cycle appears in:
  • stock prices
  • real estate bubbles
  • economic wars
  • gold rushes
  • social movements
  • and other phenomena
[ top ]

Sentiment Indicators

Commitment of Traders data (COT)

Collected and distributed by the Commodity Futures Trading Commission, COT Measures the numbers of contracts held by traders who fall into each of 3 groups:

  1. Commercials (large firms, use futures for hedging)
  2. Non-commercials (smaller than commercials, though might be thought of as large speculators)
  3. Non-reportables (small speculators, people whose positions are not big enough to merit individual reporting to the CFTC Commercials are thought to be the "smart money"

Options data

  • Put/Call Volume Ratio
  • Put/Call Open Interest Ratio
  • Index options versus equity options activity
  • VIX

ETF Volume versus volume in member stocks

  • Jason Goepfert's "Liquidity Premium" indicator
  • Also raw volume in QQQQ, which goes to a low value at tops

Rydex fund assets

Not as useful as when Rydex family was new, and number of funds was small. There's been a proliferation of competing ETFs as well, which may help explain the diminished usefulness.

 

Contango Backwardation

How does the price of a near month futures contract compare with those farther ahead in time?

Total Open Interest

For some futures contracts, open interest will rise or fall with prices as the prices move with the trend, or opposite to the trend. Stock index futures do not work this way, due to the nature of the contract specifications, but for T-Bonds and gold especially the confirmation or refutation of a move by open interest has important messages.

Total assets invested

Total assets in certain investment articles, e.g. gold ETF GLD.

[ top ]

Liquidity Waves

The principle that a price structure seen in one market or price series will be repeated after a period of delay in another market. Several such relationships have been found. 

  • Gold leads inflation by 14 months
  • Inflation leads crime rates by 1 year
  • Inflation leads unemployment rate by 2 years
  • Gold leads commodity prices by 4 months;
  • Lumber prices lead Fed Funds rate by 1 year
  • Net position (long or short) of commercial traders of Eurodollar futures traders leads T-Bill yield by 1 year
  • 4-month ROC of M1 leads stock prices by 8 months
  • Stock prices lead presidential approval poll numbers by 6 trading days
  • Stock prices lead federal tax collections by 1 year
  • T-Bill yields lead the VIX by 2 years
[ top ]

Price Pattern Analogs

Price patterns seen before tend to show up again, reflecting a repetition of the "dance steps" as prices respond to similar market conditions. The conventional economic indicators might show vastly different situations, yet prices still seem to portray repeating price pattern behavior.

[ top ]

Timing Models

Proprietary technique of using values for certain indicators described above for calculating future turning points (tops and bottoms).

[ top ]

Goldollar Index

When multiplying the spot price of gold by the US Dollar Index, we get a unitless measure of gold's real value irrespective of currency fluctuations. Comparing the Goldollar Index to the price of gold itself can point out confirmations or non-confirmations of a move.

[ top ]

Fibonacci Ratios

The standard Fibonacci Ratio is 1.618 to 1, and it shows up in nature in a variety of ways, including plant growth patterns, human body symmetry, and the arms of spiral galaxies and hurricanes. It also shows up in the growth and contraction legs of price movements. Related ratios are 0.618, 0.382, and 0.236.

Fibonacci Ratios are most easily and correctly used in identifying retracement targets within price corrections. They can also be found in the magnitudes of expansions of prior moves, but are more difficult to use for this. There is some evidence that Fibonacci Ratios can appear in time cycles, i.e. horizontally on a chart, but this too is much harder to use successfully than with price retracements (vertically on the chart).

Two good books on the subject of using Fibonacci ratios

 

[ top ]