We use a high-accuracy computational code to investigate the precision
with which cosmological parameters could be reconstructed by future c
osmic microwave background experiments. We focus on the two planned sa
tellite missions: MAP and Planck. We identify several parameter combin
ations that could be determined with a few percent accuracy with the t
wo missions, as well as some degeneracies among the parameters that ca
nnot be accurately resolved with the temperature data alone. These deg
eneracies can be broken by other astronomical measurements. Polarizati
on measurements can significantly enhance the science return of both m
issions by allowing a more accurate determination of some cosmological
parameters, by enabling the detection of gravity waves and by probing
the ionization history of the universe. We also address the question
of how Gaussian the likelihood function is around the maximum and whet
her gravitational lensing changes the constraints.