Forecast Arrays

Princeton Forecast Arrays

The Princeton Forecast Arrays are only available through our monthly publications and daily FAX services. These arrays provide a quick graphical representation of a number of independent forecasting models including Composite and Empirical Timing Models, Directional Change, Trading and Volatility Models. The Forecast Array will enable you to quickly see when our computer models are looking for ideal highs or lows in addition to important changes in trends and volatility.

Each model includes numerous variations. For example, Empirical Timing Models reflect the total sum of at least 20 separate timing frequencies. Each frequency is then plotted. The greater the number of individual timing frequencies that converge during a specific time period results in a greater amplitude. Therefore, the tallest bar reflects the greatest number of independent targets.

The proper use of our Forecast Array is fairly straight forward. The plots within each model provide a guide as to where a turning point in price or volatility should unfold. With the exception of the Trading Cycle indicator, each model is designed ONLY to provide an indication of when the market will change trend at a specific point in time. Turning points in price (high or low) unfold on BOTH the highest and lowest bars. There is NO direct relationship between turning points and highs and lows in the array. A low in price may unfold with the highest plot and a high could form on the lowest plot.


The Combined portion of our array provides a general view to over sixty independent models including timing and volatility. Each separate model from timing to volatility is taken as a sum and reflected in this Combined portion of the array.

Composite Models

This is a representation of our Composite Timing Models which expand and contract in time. The cyclical frequencies are generated solely by our Artificial Intelligence computer system without human interference or “hard-wired” rules. We merely taught our computers the methodology of timing theory and it provides its own analysis, much as if it were a human analytical researcher.

Empirical Models

The Empirical portion of this array refers to our Empirical Timing Models. Again, the greater the amplitude indicates an increased number of independent timing models converging during that time period. Typically, this portion of our array represents a minimum of 20 different timing frequencies. In some markets, the total number of frequencies may reach as many as 64. Empirical Timing models are what we call “hard-wired.” These are frequencies of a fixed duration that have been determined manually through research over the years.

Directional Change Models

A Directional Change differs from a Turning Point in that Directional Change targets need NOT be the period when an intraday high or low forms. Instead, a Directional Change is the target where the market actually begins to make a decisive move. For example, it is possible to find the intraday high or low take place 1 to 3 time units (days, weeks, months, etc.) preceding the Directional Change target. On a weekly level, the actual high might form on Wednesday while the market moves sideways within a narrow trading band until the Directional Change comes into play, perhaps the following week. During periods of high volatility it will be more common to find the Turning Point and Directional Change converge during the same time period. This normally occurs when a market is making a spike low or high.

Volatility Models

The Volatility portion of our Forecast Array provides an indication as to when a change in the current volatility trend will take place. Unlike timing, volatility is only concerned with percentage movement and NOT the actual point in time when highs and lows will be established in terms of intraday price objectives. Again, the targets reflect “turning points” in volatility and not the implicit level of volatility itself. Thus, the low in volatility might form on the highest bar while the high in volatility could unfold on the lowest bar.