The ranking of the selection proportions when performing truncated selection consequently favoured stronger selection, i.e. using fewer parents, but the effect of changing the selection proportion was small and non-significant (P > 0.05) in all cases. The increase in profit when performing selection, compared to Rand, was consistently significant (P < 0.05) and ranged from 1.2 to 1.4 σBV. Measured in relative increase, performing selection was much more beneficial when Rand gave a low profit, i.e. when there was a negative correlation between the traits and/or SIB had a relatively high economic weight.