Purpose - The structure of the Malaysian fund market presents a unique setting in which to examine behavioural and cultural differences in the performance of fund managers. The purpose of this paper is to utilise Taylor s extension of the tournament model of Brown et al who argued that using an exogenous (endogenous) benchmark induces losing (winning) managers to gamble. This presents two competing testable hypotheses that are investigated in the current study. Design/methodology/approach - The authors use a sample of Malaysian unit trusts covering the period 1982 to 2010, applying the non-parametric cross-product ratio methodology to test all Malaysian funds and determine whether there is empirical evidence of tournament behaviour. The authors separate Malaysian funds into two main categories (conventional and Islamic) to find out whether different fund types affect the behaviour of the funds as a whole. Findings - Overall, Taylor s theory does not hold in the Malaysian fund market, as conventional funds display tournament behaviour regardless of the benchmark used. However, Islamic funds do not display any significant tournament behaviour. Originality/value - The current study uses a non-parametric approach to look for evidence of tournament (gaming) behaviour in the performance of fund managers in Malaysia. In doing so, the authors extend the tournaments literature by examining the performance of three data sets pertaining to the performance and evidence of tournament behaviour in: all managed funds in Malaysia; Islamic funds; and conventional funds. A major motivation for choosing the Malaysian data of unit trusts is to investigate and examine the behaviour of funds operating in an economy that is an emerging market in the rapidly expanding Asian economy; is a market that has a reporting period in line with the calendar year; and is an economy with a strong presence of Islamic funds (Shariah) and Muslim population.