Review of PWA Chart Construction

One of my goals is for our subscribers to have as much clarity as possible about how our charts are put together and how they work so I'm going to review that here. Greater clarity will mean greater success in the long run.

How the Charts are Built:

Our charts are constructed using a "centered moving average" (the dotted line) which is then replicated above and below the dotted line to create what I call the "range bands" that highlight buying and selling opportunities. What the term centered moving average means is that the latest data point is calculated and then graphed on the chart "one half span" in arrears. As an example, if we were using a 40 day centered moving average the last actual data point for the dotted line and the range bands would show on the graph 20 days back in time. The last 20 days are mathematical projection.

ITS PWA Chart - Half Span

Static vs Nonstatic Chart Sections:

Since the last "half span" (20 days in our example) are projection, the band levels are subject to slight change with each day as new price data comes in. A price increase will drag the bands up a little, and a price decrease will drag them down a bit. The magnitude of the change of band levels for today's date will decrease a little each day until the half span is reached (20 days in our example) at which point they become static and do not change further.

When you look at the static section of our charts (anything left of the half span of whatever MA setting you're using) the program seems to give very nearly perfect signals. In actual practice it isn't as perfect as the static chart would make it appear. The reason is that we don't make decisions with the static chart. We make decisions with the information on the far right (today's data) which is the NON-static section that is still subject to change.

When we are in a lateral market and the bands are flat the change from today to the half span point are minimal.

When we are in a rising market, the bands will be pulled up--and we like that a lot--so we're happy to let the price run as high as it will. In this instance what looked like an OK buy originally will look like a sensational buy when the chart becomes static because the bands will have been dragged up a bit each day driving our buy point further into the dark green buy band.

When the price falls after we make a buy, or more commonly starts up for a few days and then falls significantly, what looked like a good buy point (in the dark green band) might show later as being in the neutral zone and not a good buy at all.

The last point above is why I recommend that people always use the built in stop-loss line whenever they make a buy. The stop-loss doesn't come into play often, but when the market crashes, as it did in the autumn of 2008, it is invaluable. The problem, of course, is that a person never knows ahead of time when major corrections, or crashes, are going to happen so the stop-loss line is "insurance" against severe loss. Hence the recommendation.

Back Studies with PWA:

With PWA it is possible to do accurate back testing because of how the Master Program works. Each time any change at all is made to the chart the computer recalculates and redraws the chart. When a user resets the start date to go back in time, you are looking at the chart as it would have appeared originally. So, the best way to do a back study is this:

  • Using either the slider bar or the date dialog boxes, set the "last date on chart" (found in the lower right corner of the chart window) to a few days before the time frame you want to study.
  • Advance the chart a week or so at a time until the price line seems to be approaching a buy or sell band.
  • Then advance the chart a day or two at a time while reading the chart indicators.
  • Make buy or sell decisions and record them.
  • Then, calculate percentage difference between buys and sells and run totals to see if you would have done better than with simply a "buy and hold" strategy for the same time span.
  • In this process always set the stop-loss line when you make a buy. A good range to experiment with is 5% to 7.5% below your purchase price.

The same process is useful in comparing potential returns from different stocks or funds or for comparing different chart parameter settings.