Abstract
We assess the accuracy with which future galaxy surveys can measure cosmological parameters. By breaking parameter degeneracies of the Planck cosmic microwave background satellite, the Sloan digital sky survey may be able to reduce the Planck error bars by about an order of magnitude on the large-scale power normalization and the reionization optical depth, down to percent levels. However, pinpointing attainable accuracies to within better than a factor of a few depends crucially on whether it will be possible to extract useful information from the mildly nonlinear regime.