Following on from a poor third quarter, the Nifty Thrifty has given up almost all the gains it made in the first half of the year.
It has risen just 3 per cent since December 2014, barely more than the benchmark investment in accumulation units of a FTSE 350 index-tracking fund, which has risen 2 per cent. The algorithm that selects the shares slipped up on a giant oil slick.
If you invested £30,000 in the Nifty Thrifty five and a half years ago it would be worth £47,210 today, a gain of 57 per cent. Over the same period, if you had invested £30,000 in the index-tracking fund it would be worth £45,125, a gain of 50 per cent.
To view the Nifty Thrifty's holdings and trading chronology, click here.
MORE SELECTIVE ALGORITHM
Both the Nifty Thrifty and the fund reinvest profits from dividends, and both select shares from the same index, the biggest 350 companies listed on the main market in London. The Nifty Thrifty's algorithm is more selective, though.
It invests in only 30 of the 350 companies, those non-financial companies that are financially strongest according to the F_Score - a nine-criteria rating designed by Joseph Piotroski, an accounting professor - and that have the highest combined profitability and valuation rankings according to a system devised by value investor Joel Greenblatt.
The F_Score was designed to weed out companies most likely to fail by identifying those that have become less profitable, less efficient, and more dependent on outside sources of finance (either by using more bank debt or issuing more shares).
Greenblatt called his system the Magic Formula, the idea being to select good companies, those that are highly profitable, at cheap prices. The measure of profitability is return on capital; the measure of value is the earnings yield.
These statistics are not foolproof, but they improve the odds of picking good companies at cheap prices, which is why the system only works reliably on large baskets of stocks over long periods of time.
Judging by the list of companies ejected this quarter, the algorithm has been comprehensively fooled. Only one, holiday firm Tui, made the portfolio any money.
Three oil and gas exploration and production companies proved to be disastrous investments, led by SOCO International which posted a total return, including dividends and trading costs, of -66 per cent.
Nostrum Oil and Gas returned -40 per cent and Ophir Energy -30 per cent. Weir, an engineering firm that supplies and services oil and gas producers, returned 34 per cent less than the portfolio invested a year ago.
The fifth company in our list of losers is Ferrexpo, which mines iron ore in the Ukraine. It returned -53 per cent.
It doesn't take a computer algorithm to work out the common factor. The profitability of natural resources companies can quickly come into question when the prices of the resources they mine fall.
China's slowdown has reduced demand for natural resources in general, and high levels of oil production in the US and Saudi Arabia have produced an oil glut in the short term, which the Nifty Thrifty algorithm failed to anticipate.
A year ago, the algorithm was looking backwards at the tail end of a boom in profits, while falling share prices indicated traders in the market were beginning to look forward to lower profits. According to the algorithm, the businesses were strong and the shares were cheap.
Now the algorithm is looking back at declining profitability, and it has rejected these companies. Instinct tells me that might be a mistake too, but instinct does not come into it when you are a slave to the machine.
When I started the portfolio, I staggered the share-buying as equally as possible over the four quarters of the first year, adding eight shares in June 2010, eight more the following September and then seven shares in December and seven more in March 2011.
Each share is replaced after one year unless it requalifies for the portfolio. To do this it must be in the FTSE 350 index, still have an F_Score of five or more out of nine, and still be one of the most highly ranked candidates according to Greenblatt's Magic Formula.
This time, all seven companies due for ejection were replaced. Dividends paid in the past six months and the proceeds from the ejections leave about £1,100 to invest in each of the seven new additions (see table above, click to enlarge).
They are newly listed low-cost airline Wizz; recruitment firm Hays; IMI, which manufactures valves and actuators for a wide range of industrial and commercial applications, London estate agency Foxtons, betting shop William Hill, Mitie, which provides services for property owners; and Burberry, the luxury fashion brand.
Although oil exploration and production company Afren went into administration in August, it languishes unceremoniously in the portfolio until the anniversary of its selection in March when I will probably write off what little remains of the Nifty Thrifty's investment.
Reportedly, the liquidator is selling off the company's art collection as well as its oilfield assets in Nigeria, but it won't be enough to pay off creditors.
The write-off will mark the end of a foray into oil that has exposed weaknesses in the algorithm.
Subscribe to Money Observer Magazine
Be the first to receive expert investment news and analysis of shares, funds, regions and strategies we expect to deliver top returns, plus free access to the digital issues on your desktop or via the Money Observer App.Subscribe now