We poorly understand the factors shaping variation in fitness among individuals, i.e. in their ability to make a contribution to the future gene pool. While short-term fitness proxies, e.g. lifetime reproductive success (LRS), are commonly used to measure fitness, how well do these proxies perform? Multigenerational human genealogical data allow the estimation of individual genetic contributions (IGC) - a fitness approximation closer to its theoretical definition - over many more years than is possible for other species. Here we use genealogical data from two local populations in Switzerland to estimate the IGC for 2,623 individuals on average 308 years after they were born. We find that the number of grandoffspring predicts IGC best explaining 28 percentage points more variation than LRS. Overall, LRS explains 29% of the variation in IGC, and 33% when accounting for offspring survival to adulthood. This suggests that offspring reproductive success is a key determinant of individual fitness. Nevertheless, we find that LRS only slightly underestimates the IGC of offspring as family sizes increase, and hence we find little evidence for an offspring quality-quantity trade-off. Together these findings suggest that, albeit relatively imprecise, LRS is a largely unbiased fitness proxy in this historic human population.