We backtest stock screeners using the stock price history of thousands of stocks
and the historical fundamental financial data of each stock’s underlying company.
The only parameters used to make investment decisions are purely quantitative fundamental
performances of the companies (such as PER, Dividend Yield, Price to Book Ratio,
ROE, Equity to Debt ratio...). The only information used is the information available
prior to the investment decision (no analyst forecasts nor charting tools are used).
We investigate behavioral patterns of those screeners (IRR, volatility, Peak to Trough,
comparative performance relative to a given index).
We analyse the source of the returns (dividends, distribution of capital gains or
losses on all investments).
We optimize fundamental market screeners to come as close as possible to the investor’s
objectives
Once the algorythm of a screener is designed, we run our proprietary software to
finetune the parameters of the screeners to meet certain required criteria.
The optimization has to be performed carefully in order to avoid some of the pitfalls
of back testing such as « data mining » or « survival bias ».
We test the screeners on other periods of time or geographies than that that were
used for their optimization in order to check the longer term relevance of the screener.