A Measure of the Trading Model Performance with a Risk Component

Posted: 25 Dec 1998

Date Written: June 1994

Abstract

The purpose of this paper is to suggest a new measure of trading model performance which accounts for the following requirements: 1. a high total return; 2. a smooth behavior around a straight line; 3. a small clustering of losses; 4. no bias towards low-frequency trading models. It is important to define a value that describes the performance well to minimize the risk of over-fitting in the in-sample period and to be able to compare different trading models with each other. In the first section of the paper, we discuss the Sharpe index, a measure frequently used to evaluate portfolio models. We show that it does not fulfill all the above requirements. In the second section, we propose a new measure based on a risk averse trading profile and the utility function formalism of Keeney and Raiffa (1976). This measure is numerically more stable than the Sharpe index and exhibits fewer deficiencies. In the third section, this measure is extended to a multi-horizon measure in order to be able to account for the clustering of losses in the return curve. In the same section, some numerical aspects of the computation of this variable are discussed.

JEL Classification: G11

Suggested Citation

Dacorogna, Michel M. and Müller, Ulrich A. and Pictet, Olivier V., A Measure of the Trading Model Performance with a Risk Component (June 1994 ). Available at SSRN: https://ssrn.com/abstract=5418

Michel M. Dacorogna (Contact Author)

DEAR-Consulting ( email )

Scheuchzerstrasse 160
Zurich, 8057
Switzerland
+41795447327 (Phone)

Ulrich A. Müller

Olsen & Associates ( email )

Seefeldstrasse 233
CH-8008 Zurich
Switzerland
+41 (1) 386 48 16 (Phone)
+41 (1) 422 22 82 (Fax)

Olivier V. Pictet

Pictet Asset Management ( email )

Geneva
Switzerland

Here is the Coronavirus
related research on SSRN

Paper statistics

Abstract Views
910
PlumX Metrics