Here's my idea. You save a ranking system, unweighted. You then specify a number of stocks (or a ranking level, e.g. 98), a universe, an investment time period (with starting and ending dates), and a holding period, with an option for variable slippage as well (if possible). You click on "optimize." P123 then tests the ranking system (perhaps using the rolling backtest tool) with different combinations of weights per factor (to make the number of combinations less daunting, the weights could be in iterations of 2.5%, i.e. 0, 2.5, 5, 7.5, 10, etc.) to come up with the best total return for the specified investment timeframe, number of stocks, and holding period. There would, of course, have to be a strict limit to the number of factors in the system.
Results: Total score: , # of Votes: , Average: 0.0
You will be able to view details once you have voted.
Requested by: yuvaltaylor
On date: 05/13/16
Category: Ranking

4 comments