The document describes a study comparing the L-Co-R co-evolutionary algorithm to other forecasting methods for medium-term time-series problems. The L-Co-R algorithm uses coevolving radial basis function neural networks (RBFNNs) and time lags to make predictions. The study uses several Spanish economic and transportation datasets and evaluates the accuracy of L-Co-R versus exponential smoothing, Croston's method, Theta, random walk, and ARIMA models based on mean absolute percentage error, mean absolute scaled error, and median absolute percentage error. Preliminary results show that L-Co-R performs well in predicting airline passenger numbers compared to other methods.