We have derived the uncertainties to be expected in the derivation of galaxy physical properties (star formation history, age, metallicity, reddening) when comparing broad-band photometry to the predictions of evolutionary synthesis models. We have obtained synthetic colors for a large sample (9000) of artificial galaxies assuming different star formation histories, ages, metallicities, reddening values, and redshifts. The colors derived have been perturbed by adopting different observing errors, and compared back to the evolutionary synthesis models grouped in different sets. The comparison has been performed using a combination of Monte Carlo simulations, a Maximum Likelihood Estimator and Principal Component Analysis. After comparing the input and derived output values we have been able to compute the uncertainties and covariant degeneracies between the galaxy physical properties as function of (1) the set of observables available, (2) the observing errors, and (3) the galaxy properties themselves. In this work we have considered different sets of observables, some of them including the standard Johnson/Cousins (UBVRI) and Sloan Digital Sky Survey (SDSS) bands in the optical, the 2 Micron All Sky Survey (2MASS) bands in the near-infrared, and the Galaxy Evolution Explorer (GALEX) bands in the UV, at three different redshifts, z=0.0, 0.7, and 1.4. This study is intended to represent a basic tool for the design of future projects on galaxy evolution, allowing an estimate of the optimal band-pass combinations and signal-to-noise ratios required for a given scientific objective.

20 pages, 9 postscript figures, 3 tables, accepted for publication in AJ