We considered a general model for sizing a stand-alone photovoltatic system, using as energy input data the information available in any irradiation atlas. The parameters of the model are estimated by multivariate linear regression. The results obtained from the numerical loss of load probability size method (LOLP) were used as initial input data to fit the mode. For this fit we have used daily global irradiation data taken from 222 US meteorological stations for the period 1961-1990. The expression proposed allows us to determine the photovoltatic array size, with a coefficient of determination to 0.96. This coefficient is independent of the used LOLP value. System parameters and mean monthly values for daily global irradiation on the modules surface are taken as independent variables in the model. It also shows that the proposed model can be used with the same accuracy for other locations not considered in the estimation of the model. We also propose a model which would allow us to calculate optimum tilts for the array surface taking the latitude into account as well as the variability of the incident irradiation.