There are tens of thousands of stocks worldwide you could invest or trade in. How do you pick the best?
You need a screening system – a series of sieves, each of which excludes stocks that don’t match your criteria. Eventually you will be left with just a few stocks to examine in detail. If you want to design your own trading system, how can you do it?
Getting Started with a Stock-Trading System
At the most basic level, you need to do two things.
- You need to write down “buy” and “sell” rules. These rules will sieve stocks to tell you which stocks you should buy, when you should buy them, and when you should sell them.
- You need to test your rules to see whether your system is likely to make a profit. It’s easiest to do this if you have several years’ worth of historical share price data. Fortunately this sort of data is readily available.
Your First Trading System
I’ll assume you are new to trading systems and I’ll begin with a system using very basic of trading rules. These could be, for example:
Buy Rule: If yesterday’s closing price was greater than the previous day’s closing price, buy the stock today when the market opens.
Sell Rule: If yesterday’s closing price is less than or equal to the previous day’s closing price, sell the stock today when the market opens.
Here’s a very basic Excel spreadsheet to show you how you could implement these simple rules.
At this stage, you need to test whether your trading system is likely to be profitable.
To do this, you should test its performance with a wide variety of stocks. Test it over different time scales too. For example, you might want to check each year 2006, 2007…..2016 separately to ensure your system was profitable in every year. Now look at your results. Was your system profitable? Remember it has to cover your brokerage costs before you can count it as successful. You will almost certainly find the system we’ve used as an example above is not a profitable one. Your second trading system will need to be more sophisticated.
In the first circumstances, brokerage costs would be high. You can reduce the number of buy and sell decisions if you lower the volatility of daily prices.
If you study a large number of stocks, you will find that some stocks are naturally more volatile than others. The chart on the right shows the price history of two stocks, both rising by the same amount in the same time. The prices of the stock shown in red are more volatile than the stock shown in green.
The problem caused by high volatility is known as the “whipsaw effect” – frequent buy and sell signals are triggered by relatively large rises and falls in price within a short timeframe. In the chart above, you would buy and sell the red stock more frequently than the green stock because of whipsawing on the red stock. Over the whole period covered by the chart, both stocks rise by the same amount but you would make a lower profit on the red stock because of the costs of buying and selling more frequently.
To lessen volatility, you can use an average price – averaged over several days – rather than the actual price in your trading system.
You can reduce the number of “buys” and “sells” by employing smoothed / averaged data. Smoothed data reduce the impact of daily price fluctuations.
The (dashed) chart on the right shows the effect of smoothing the original (solid) data by taking the average of the previous 7 days data. Notice how, compared with the unsmoothed data, there are fewer peaks and troughs in the smoothed data. This naturally reduces the number of buy and sell signals.
As an alternative to using a simple average, you might want to consider use a weighted, moving average of closing prices in your system. You could use a weighted, moving average to give higher relevance to the most recent data and less relevance to older data.