The New Zealand forest industry produces a diverse range of log grades and sorts to meet domestic and export market demands and to maximise returns to the forest grower. An implication for the supply chain is the number of log grades and sorts a harvesting operation is expected to produce from one species, radiata pine (Pinus radiata). The number of log grades and sorts can impact on landing size and layout requirements, value recovery, log-making complexity, machine utilisation and quality control requirements. A study was conducted to investigate if the number of log sorts affects mechanised log processing productivity and value recovery. This would determine if any gross value gains derived from producing a higher number of sorts are offset by losses in log processing productivity. Two landing-based mechanised log processors at cable yarder harvesting operations were studied using different cutting scenarios producing five, nine, twelve and fifteen log sorts. The study collected data from over 26 hours of mechanised processing which included the processing of 578 stems at an average piece size of approximately 1.6 m³. Machine utilisation results showed processors spending 84% of total time on productive tasks and that 49% of total time was spent on the primary productive tasks of log processing. Quadratic regressions were used to model log processing productivity trends which showed piece size and cutting scenario as significant predictor variables (p-value <0.01). There was a significant difference between cutting scenario with five log sorts and the cutting scenarios with twelve and fifteen log sorts (p-values <0.05), as well as a significant difference between the nine and fifteen log sort cutting scenarios (p-value <0.01). There was not enough evidence to suggest productivity was different between cutting scenarios producing five and nine log sorts. Based on this analysis, it was likely that the null hypothesis that the number of log sorts does not affect log processing productivity should be rejected. At a piece size of 2 m³, the productivity model estimated processing productivity was 10% higher producing nine log sorts compared to producing fifteen log sorts. A linear regression model showed a strong relationship between gross value recovery, piece size and cutting scenario (p-value <0.01). Gross value recovery increased as the number of log sorts increased. A significant model suggested it is likely null hypothesis 2, that the number of log sorts does not affect gross value recovery, should be rejected. There were only some differences in variances between cutting scenarios which were statistically significant. Both the average results and regression estimates showed the five log sort cutting scenario recovering 94% of the value of the cutting scenario with fifteen log sorts. Incremental gains in value recovery as the number of log sorts increased were marginal, which appeared to be due to log prices for many major log grades trading in a close range in relation to historic price trends. Regression trends for productivity and gross value recovery indicated that the most optimal cutting scenario, in terms of processing value outturn per productive machine hour, was the cutting scenario producing nine log sorts. This suggests that declines in processor productivity offset gains in gross value recovery when producing twelve and fifteen log sorts. Market sensitivity analysis suggested that differentials in log prices impact on the number of log sorts which optimise the value outturn per productive machine hour from log processing.