ABG Analytics logo
 

Two Technical Scoring Techniques

"Traditional" Scoring Technique

The traditional interpretations of the common indicators listed above suggest a simple scoring method for summarizing the information contained in them: give a point for an indicator in a bullish position, and subtract a point for a bearish level of an indicator. Clearly, that simple method assigns the same weight to all signals and in this sense is arbitrary. But, if the traditional interpretations of those indicators have any validity, we should buy stocks whose indicators are mostly bullish, and sell or short sell those stocks with most indicators in bearish positions.

The method described above is sufficient to produce an index score that might be used to generate buy and sell signals for a trading system. Starting with 0 points, add a point if:

1) current price is above its 50 day moving average
2) current price is above its 200 day moving average
3) the 50 day moving average is above the 200 day moving average
4) RSI is oversold (less than 30)
5) MACD is greater than 0
6) MACD is above its signal line

Subtract a point if:

7) current price is below its 50 day moving average
8) current price is below its 200 day moving average
9) the 50 day moving average is below the 200 day moving average
10) RSI is overbought (greater than 70)
11) MACD is less than 0
12) MACD is below its signal line

Clearly, the results of this system would change if other indicators were added to the ones above. The indicators we use here are only a small set of all the indicators we could use, but they are widely used, capture several different aspects of price movement and position, and involve short, medium, and long time frames. A more complicated system, including volume indicators, for example, could be similarly constructed.

This "raw" score will vary from -5 to 5, with 5 being the most bullish. For ease of use, that raw score is rescaled to a final score that ranges from 1 to 5 based on the following transformation (-5,-4=1; -3,-2=2; -1,0,1=3; 2,3=4; 4,5=5). This score is provided in the Weekly Technical Status Report. Also see the performance of trading systems using this summary indicator.

Statistical Scoring Technique

Do the bullish conditions really mean a higher likelihood that a stock will go up in the next 20 days? On average, what is the average change in price over the next 20 days if various combinations of the above conditions hold? Which conditions (if any) really matter, and how much do they matter? Are there better ways of combining these indicators? These are the questions that statistical analysis can address.

In our statistical analysis, we assign weights to various functions and combinations of those indicators listed above, beyond what was used in the traditional scoring method. Using many stocks and a long time period, we construct weights based on how well the indicators predict future percent changes in price over the next 20 days for 100 stocks over 1,000 days of data. This results in a measure of expected return over the next 20 days. That measure is then scored on a 1 to 5 scale, based on the five equally sized groups ranked by the prediction. This scoring system has the potential of predicting future price changes if past relationships between the technical indicators and future price changes continue to hold to some degree going forward. In the next section, we consider whether either of these technical scores has any predictive value.

Continue to part 4 - "Predictive Value of Technical Scores"

Contents: A View on Technical Indicators and Trading Systems