Template Tools
The fast, slow and full stochastic PDF Print E-mail

The fast, slow and full stochastic


Overview (oscillators)

Oscillators work on the premise that markets tend to over extend themselves on both sides of the scale. Whether these extreme conditions are described as overbought or oversold (momentum), strength (trend) or something else, it is when these extremes are reached that it is assumed the market is in a position to begin a correction (minor or major). These extreme values in oscillators tell us when to be watchful of the ever dreaded and inevitable trend reversal.

Oscillators also give us information about the general trend. As trend analysis works with the price line, trend analysis is also a valuable part of analyzing an oscillator. This is where part of the "predictive" abilities of oscillators are seen. These events are also known as convergences and divergences. This is when the trend of the oscillator has broken before the price trend, and they are no longer moving in parallel.

Here are some general rules to use when applying your oscillator of choice

  1. Signals are stronger when they occur at the extremes
  2. Crossing of zero lines and equilibrium lines should be generally considered warning or directional signals. Strength of these crossings should also be watched and temporary penetrations should be ignored.
  3. Warnings/signals can be generated when the oscillator is diverging/converging with the price line.
  4. If the indicator generates a signal/line cross/extreme value, the signal/warning is stronger if combined with a complementing convergence/divergence. (ie: if they are still trending together (rising bottoms/falling tops), the signal is weaker)

By using all the information provided within an oscillator, these types of technical indicators can be a valuable tool in the technical analysis arsenal. In this newsletter we will be focusing on the Stochastic indicator.

The stochastic, the fast, the slow and the full.

The fast (original)

There are three types of generally accepted stochastic formulas. The original stochastic oscillator was developed by Dr. George Lane and is referred to as the fast stochastic. The stochastic's formula (%K) is based on two observations made by Dr. Lane.

Observations

  1. As an uptrend reaches it's end, closes tend to approach the daily highs more often.
  2. As a downtrend reaches it's end, closes tend to approach the daily low more often.



(fast stochastic (9,3) on Dow Jones Industrial Average, 1 year chart, captured July 17, 2005)

The stochastic has two lines, the %K and the %D. The %K is the plotted instrument (see observations) while the %D is the moving average of the %K. The %K is more sensitive and it is the %D line that triggers the trading signals.

The slow



(slow stochastic (9,3) on Dow Jones Industrial Average, 1 year chart, captured July 17, 2005)

Some traders felt that the %K line was overly sensitive. To correct this issue they smoothed the %K line. This was done by plotting a 3 day sma of the %K line rather than the original (fast) %K. The new %D line would be calculated by the new (slow) %K.

What does all that mean? Just that the new slow stochastic is a smoothed (averaged) fast stochastic. Have a look at the following graph,

The full

Then the full stochastic was created. Rather than being forced to use the 3 day SMA of the %K as in the slow stochastics, traders felt that this should be a variable. This created the third variable called the smoothing variable. This changes the amount of days to be used in the smoothing average of the %K.

You can also recreate the fast and the slow stochastic by the full stochastic.

To mimic the fast stochastic, use a 1 day smoothing number. To mimic the slow use a 3 day smoothing number.

Signals

Since the fast stochastic %K line is quite sensitive, it generates quite a few signal line crosses (%K crosses %D) while the slow stochastic is a little more frugal in it's dispensing of line cross signals. It is extremely important with any type of stochastic to evaluate the strength of signal generated. Here are some good tips to help decipher what is a good signal

  1. Signals are stronger when they occur at the extremes (crossed 80/20). Also a more conservative addition would be to wait till it crosses back out of the extreme value.
  2. If the indicator generates a signal/line cross/extreme value, the signal/warning is stronger if combined with a complementing convergence/divergence. (ie: if they are still trending together (rising bottoms/falling tops), the signal is weaker)
  3. Crossing of zero lines and equilibrium lines should be generally considered warning or directional signals. Strength of these crossings should also be watched and temporary penetrations should be ignored.

Example of a strong signal(s)

In this example, all the stages are met,

  1. There is a divergence between the price and the stochastics. This should warn the analyst of an impending change in the trend.
  2. This is followed by a signal line cross and a downwards penetration of the 80 line as well as my stochastics support line. While this is not enough for me, some traders might react. At this point the price has not yet breached my trend line, however the price has begun to fall.
  3. The %D crosses down through the 50 line. This is a more conservative confirmation signal. This rule also helps not reacting on false signals.
  4. Finally the %D has crossed down through the 20 line as well as the price has penetrated it's support line. This is a considered a very conservative signal for confirmation that the trend has reversed. This last rule is typically favoured by longer term traders.

TIPS & TECHNIQUES - Using the stochastic

When using stochastics be mindful of the effects the different parameters make on the signals that are returned when screening. The general rule of thumb is that the larger the parameter, the less sensitive the indicator becomes, the smaller the parameter the more sensitive the indicator is.

The first stochastic is a slow stochastic(9,3), the is a slow stochastic (21,9). As you can see the larger the parameter the smoother and less sensitive the stochastic becomes.