Moving verification of «return-risk» model of US stock market

Abstract

This article addresses the development and testing of ‘Return-Risk’ model aimed at designing the investment strategies suitable for practical needs. The model rests on fundamental principles of portfolio analysis and incorporates the following properties: expected return and risk are derived from historical data while learning sample size and time span are set by practice requirements in such a way as to maintain efficiency and regular monitoring of investment portfolio. The model tested on independent material (data of 2017 were not used for the model) shows: the suggested method of moving verification results in higher forecast accuracy for return and risk of investment portfolio and, consequently, in higher quality of investment decisions. The model annual yeild in moving verification mode is 25%, whereas S&P500 index shows only 15% of the annual gain, i.e. the ‘Return-Risk’ model significantly beats the market. The win-loss ratio of deals (time spans) is 11:1.