ABG Analytics logo
 

Systems using the "traditional" and "statistical" technical scores

We have already shown above that the "traditional" technical score based on the traditional interpretation of technical indicators has little ability to predict percent returns 20 days into the future, and to the extent it does, it works in the opposite of the implied direction. This is even more evident from the next two systems.

System 10: Buy if traditional Score=5, Sell if traditional Score=1 

Avg profit per trade (% gross)            0.4
Avg days held                            26.8
Profitable trades (%)*                   22.3

Avg drawdown (%)                          4.2
Max drawdown (%)                         79.7
Avg profit/avg drawdown                   0.1
Following the purported interpretation and buying stocks with a score of 5 and selling stocks with a score of 1, which amounts to buying stocks that have done well and selling those that have done poorly, results in an unprofitable system. Now consider doing exactly the opposite.

System 11: Buy if traditional score=1, Sell if traditional score=5 

Avg profit per trade (% gross)            3.4
Avg days held                            50.3
Profitable trades (%)*                   70.5

Avg drawdown (%)                         25.1
Max drawdown (%)                         95.7
Avg profit/avg drawdown                   0.1
Following this "reverse" strategy yields a profitable system, yet the system is intolerably risky.

Now we consider systems centered around our statistical score. The first simply buys when the statistical score is 5 and sells when the statistical score is 1.

System 12: Buy when statistical score=5, Sell when statistical score=1 

Avg profit per trade (% gross)            5.0
Avg days held                            59.7
Profitable trades (%)*                   61.2

Avg drawdown (%)                         12.2
Max drawdown (%)                         75.3
Avg profit/avg drawdown                   0.4
This results in an average profit per trade of 5% over 60 days on average. It is a risky system, but actually has a more favorable average profit to average drawdown ratio of 0.4. This is not good enough, but it is the highest of any of the systems discussed thus far. This is encouraging because it suggests that further refinements may be able to improve the system. The first thing to try is adding a stop-loss, since the risk must be reduced.

System 13: Buy when statistical score=5, Sell when statistical score=1, Stop-loss=10%

Avg profit per trade (% gross)            1.3
Avg days held                            33.0
Profitable trades (%)*                   40.8

Avg drawdown (%)                          4.1
Max drawdown (%)                         38.9
Avg profit/avg drawdown                   0.3
Adding a 10% stop-loss reduces the risk, but also reduces the average profit. The average profit to average drawdown ratio decreases to 0.3.

There is a further refinement which improves the system dramatically. It is to buy stocks with a statistical score of 5, but to wait until the momentum turns up as measured by MACD crossing over its signal line. The sell signal is similarly modified to require that MACD crosses below its signal line.

System 14: Buy when statistical score=5 and MACD crosses over signal line, Sell when statistical score=1 and MACD crosses below signal line, Stop-loss=10% 

Avg profit per trade (% gross)            7.6
Avg days held                            58.7
Profitable trades (%)*                   40.6

Avg drawdown (%)                          2.9
Max drawdown (%)                         25.5
Avg profit/avg drawdown                   2.6
This modification results in what we believe is a viable system. Applied to historical data, it has an average profit of 7.6 percent and a very favorable average profit to average drawdown ratio of 2.6! This system is even more profitable when the sell criterion is modified by selling when the score falls to 2 or below and MACD crosses below its signal line.

System 15: Buy when statistical score=5 and MACD crosses over signal line, Sell when statistical score<=2 and MACD crosses below signal line, Stop-loss=10% 

Avg profit per trade (% gross)            7.1
Avg days held                            45.0
Profitable trades (%)*                   48.4

Avg drawdown (%)                          3.4
Max drawdown (%)                         25.5
Avg profit/avg drawdown                   2.1
For some, this last system is preferable because it makes a similar profit per trade, but it is a faster trading system which allows more trades to be made in less time. Although these systems are presented mainly for illustrative purposes, the statistical score used with momentum confirmation as measured by MACD can yield valuable trading signals. These scores and all the indicators discussed above are provided in the Weekly Technical Score Report.

Continue to part 9 - "Conclusions"

Contents: A View on Technical Indicators and Trading Systems