The Value & Growth strategy selects smallish companies with a market cap of £25M – £250M from the prefiltered Trading Universe of sterling denominated LSE ordinary shares. These are selected, once every month, to pick liquid companies, without excessive debt, that get a reasonable return on either their capital employed or cash invested. This is accomplished using the following list selector definitions, which employ a layered selection strategy, where each sub-term selects from the list produced by the previous term. This has three main advantages:

1. It is more efficient, resulting in faster back-tests, as it cuts down the bulk of shares using simple selection criteria first, before employing more computationally intensive tests on the reduced remainder.
2. It allows Indicator Overlays to be defined on each sub-term, if required, so the number of shares that are selected at each stage can be seen.
3. As we’ll see, using a ranking to limit the results of the sub-selection further, in the final selection, is good practice. This is because trading list size determines, along with test period length, the time it takes to run a back-test, so applying a limit acts as a safe-guard to ensure the first back-test completes in a reasonable time. That test will then report the number of shares that were selected and the limit can be increased, if required. In this case the 80 share limit is never reached, as the V&G strategy list size is between 17 – 69 shares (average of 37) when back-testing over the last 10 years.

We’ll now build up the Lists Selectors and their Private Terms (called Selector Terms in list selectors) as follows.

The first Selector Term that is applied is the Small Cap term. This first selects trading universe shares with a market cap between £25M – £250M with a spread < 3%. The spread threshold is larger than ideal due to the fact that the shares are small caps and RuleTrader’s spread estimators, which are used in back-tests, lean towards worst-case estimates in order to ensure conservative test results.

Using the DEfT Wizard, which leads us through the creation of the definition with its ‘pick an option’ / ‘fill in the blanks’ interface, it is easy to create:

The resulting list from the ‘Small Cap’ selection is then further selected by the Manageable Debt list, which selects only companies that either:

1. Have a debt < 80% of the value of the company (measured by Shareholder Equity) and whose profits before interest and tax are at least twice their interest payments; or
2. Have no debt; or
3. Do not report a gearing figure

The last test (3) is often important to ensure that companies that do not report a particular figure are not inadvertently excluded. The definition for the Manageable Debt list is:

The next sub-selection produces the Adequate Performance list, by sub-selecting the shares from the Manageable Debt list. This selection requires each company to have a minimum return on either capital employed or cash invested i.e. that are doing a good job of turning investment into profits.

The minimum threshold is given by the strategy’s Minimum Annual Return term, which is defined as:

Normally List Selectors and their Selector Terms cannot access values from the Shared or Strategy Terms areas. The exception is terms that are defined as a constant, unchanging value. The Minimum Annual Return term meets this criteria, as it is defined as a specific percentage value.

The Adequate Performance list allows companies based on either ROCE or CROCI. ROCE focuses on operating profit relative to book value, so is fine for comparing mature companies with stable capital bases in the same industry, with similar accounting practices and limited write-offs.

CROCI, on the other hand, looks at cash-flow relative to the original cost of assets (by excluding accumulated Depreciation & Amortization) to reflect the total historical investment in the business. So it is better at evaluating companies from different industries, with different accounting practices, which may have cyclical capital expenditure patterns, or which have restructured or made acquisitions.

Secondly, the selection allows either the current year’s value or the value averaged over three years. Averaging helps ensure a degree of repeatability over the years and avoids excluding companies when macro economic factors might produce bad results in one year. However, it has the problem that: (a) it is slow to react to an improvement in the way the company is run and (b) back-tests that start from the beginning of the available ShareScope data (which is usually about 10 years ago) won’t have the historic data to satisfy this criteria in the first two years of the test.

There’s always a danger, with smaller cap shares, that they may be illiquid, making them difficult (costly) to buy or sell. To guard against this, the High Liquidity Selector Term picks shares whose average daily trading volume is at least 10x the number of shares we might like to buy or sell:

The High Liquidity list selector term calculates the nominal trade volume for each share it is considering, by multiplying the nominal trade-value by 100, to convert it to pence, then dividing by the share’s price, which is also in pence. Most values in ShareScope, other than price, are given in the major currency unit.

Now we know how many of the share (that is being tested for selection into the trading list) we might want to buy, all we need to do is ensure that the average trading volume over 60 days for that share, is some multiple larger than that potential purchase volume. This multiple is given by the Liquidity Cover term, which is defined in the spec’s Shared Term’s as:

Having whittled down the trading universe list to a viable list that we may wish to trade, the main Value & Growth Shares List Selector term picks the top 80 shares, with the most price momentum, from the High Liquidity list. The reason for limiting the number of shares was discussed earlier.

We use a 60 day ‘best fit’ trend-line to determine price momentum, rather than e.g. the medium AutoTrend price-channel, because AutoTrend channels are not available for list selection, as for performance reasons they are only calculated for the shares in the trading list, once it has been selected.

As you can see above, the main List Selector term is also where we define when the list will be selected. Selections can occur as the result of a condition, or based on some period. In this case we have chosen to re-select the list every month.

Because list selection has to test a large number of shares from, in this case, the trading universe, it can have an impact on test times. This can be addressed by using monthly or quarterly selection periods or by saving the selected list from the first test, then using that saved list for all future tests (see Test System).

This latter approach is good practice anyway, when you are comparing back-tests and particularly if you are using list sampling, because it ensures you are using the same trading list in each test. If the list is selected each time, then it may change due to the  inevitable and frequent changes in the underlying shares database.

Indicator Overlays and Chart Overlays are only updated while the share is in the trading list. As shares can be selected and deselected at any time, this means overlays on that share can be discontinuous. As it’s useful to have a continuous series of data, so you can track, for example, the number of open positions in each of the Example Trading System’s strategies, it’s a good idea to ensure that at least one non-tradeable security is always included in the trading list. This is achieved using the Index List Selector, which is defined in the Shared Lists area so it can linked in alongside each strategy’s list selectors (a strategy may have multiple list selectors selecting lists at different times. The selected lists are simply merged and de-duplicated to create the strategy’s overall list). As shown, this only needs to be selected once:

Note that to create a list with a single share, it is first necessary to create an empty list, to act as the container, then add the security to it.

Taken as a whole, the complete set of List Selectors and Selector Terms is: