Competition between parasite genotypes in genetically diverse infections is widespread. However, experimental evidence on how genetic diversity influences total parasite load is variable. Here we use an additive partition equation to quantify the negative effect of inter-genotypic competition on total parasite load in diverse infections. Our approach controls for extreme-genotype effects, a process that can potentially neutralise, or even reverse, the negative effect of competition on total parasite load. A single extreme-genotype can have a disproportionate effect on total parasite load if it causes the highest parasite load in its single-infection, while increasing its performance in diverse relative to single infections. We show that in theory such disproportionate effects of extreme-genotypes can lead to a higher total parasite load in diverse infections than expected, even if competition reduces individual parasite performance on average. Controlling for the extreme-genotype effect is only possible if the competition effect on total parasite load is measured appropriately as the average difference between the realised number of each parasite genotype in mixed infections and the expected number based on single infection parasite loads. We apply this approach to sticklebacks that were experimentally infected with different trematode genotypes. On average, genetically diverse infections had lower parasite loads than expected from single-infection results. For the first time we demonstrate that competition between co-infecting genotypes per se caused the parasite load reduction, while extreme-genotype effects were not significant. We thus suggest that to correctly quantify the effect of competition alone on total parasite load in genetically diverse infections, the extreme-genotype effect has to be controlled for.