Fig. 1. The Alpha and Beta of a stock, where Alpha is the vertical intercept and Beta is the slope of the best fit line.

Using the Beta to trade is quite common, but not so common is using the Alpha. Actually stock picking by Alpha is a much more rewarding and less dangerous task than stock picking by Beta.

In our method, we design Alpha for trading very short term i.e. 5 days, screening with volatility and volume condition indicators. The number of data points used to calculate Alpha is first aggregated in clusters to produce points for percentage changes over a certain number of days. This results in a very sparse data set, and the regression line is "forced" to fit these few number of points. The under-fitting is deliberate although unconventional by statistical theory standards.

To narrow down the choice, step-by-step filtering is next applied. You could first filter by volatility which can be represented by Average True Range or some other volatility indicator of your choice. The second filter could be some sort of Buy*Volume condition, for example:

H>HY1*V>VY1*L>LY1=1

which means: High of today>High of Yesterday AND Low of Today>Low of Yesterday AND Volume of Today>Volume of Yesterday. A third filter condition could be that today's Close should be greater than yesterday's Close, i.e.,

C>CY1

It is quite up to the individual to specify the conditions according to his risk profile and his trading style. A totally mechanical approach would not be successful. For example on days that the market was down, filtering with a Buy*Volume condition may not be appropriate. On days that the market was up, C>CY1 should be part of your filter, the stocks to be selected should have moved up with the market with the majority of stocks. And after filtering you could discard any stocks with negligible Volume from your list.

Some guidelines for the use of Alpha and Beta in trading are given below:

1 For very short-term trading, stocks with Beta >1.5 can be regarded as high Beta stocks.

2.Absolute values of Alpha depend on time span of data, and period over which the change is recorded. What is more relevant is the change in Alpha.

3. A stock with high Beta moves up fast when the Index goes up, but also moves down fast with the Index, unless it has a high Alpha value in which case, the Alpha value acts as a support.

4. A stock with high Alpha, but not necessarily high Beta, can move up fast when the Index moves up, if the circumstances for the high Alpha are still present or have increased in influence. This can be depicted as a moving up of the whole regression line, resulting in a higher point of intercept with the Y-axis.

5. Therefore the way to select stocks is to look for changes in Alpha or Beta rather than values of Alpha and Beta. The absolute Alpha and Beta values only show the status quo. To add an element of prediction, the change in Alpha would be more useful.

6. It is better to choose stocks with increasing Alpha rather than increasing Beta. High Beta stocks with low Alpha values require great alertness and usually intra-day trading strategies.

7. The most potentially rewarding stocks are those that have a high Beta as well as a high Alpha; with the added conditions that these values have not peaked,or are already on the way down. This can be confirmed by graphing the Alpha and Beta values.

8. When Alpha and Beta values are graphed, and put on a split screen together with the stock's price line chart, they are seen to be in waves each having a span of between 3-5 days. These waves reflect the inevitable profit taking. But trends and patterns in the waves can also be seen, and these can be analysed using traditional technical analysis concepts of trend, support, resistance and divergence. (See Fig. 2)

Fig. 2. The Alpha and Beta values with price charts of a stock and an index. The thicker line is the Alpha.

9. Generally, a stock is "in play" when the amplitude of its alpha waves are getting bigger while its alpha value is also trending up.

10. Trading short-term with Alpha assumes a trending and reasonably volatile market. In a sideway market, Alpha would not be useful. The determination of market direction and whether it is in a trending stage can be by means of indicators like the ADX and Moving Averages.

11. In a down-trending market, you could either buy stocks with consistently high Alphas for the market's rebound, or,if short-selling is allowed, choose stocks with high Beta and low Alpha.

12. By scanning several markets and seeing which have more stocks with Alpha values at the higher end of the market range, it is possible to select which market to participate in.(A frequency histogram of markets will show which side the values of Alpha are skewed towards).

Note: The number of data points used and the percentage change over number of days is proprietary information.