An important problem in statistical practice is the selection of a suitable statistical model. Several model selection strategies are available in the literature, having different asymptotic and small sample properties, depending on the characteristics of the data generating mechanism. These characteristics are difficult to check in practice and there is a need for a data-driven adaptive procedure to identify an appropriate model selection strategy for the data at hand. We call such an identification a model metaselection, and we base it on the analysis of recursive prediction residuals obtained from each strategy with increasing sample sizes. We propose graphical tools in order to enhance the study these prediction errors, and thereby the selection of a strategy. The methodology is illustrated on real and simulated data sets, including applications in multiple regression, polynomial regression and time series. When necessary, an automatic metaselection can be performed by simply accumulating prediction errors. We present an asymptotic result which can be interpreted as a consistency property of such an automatic metaselection. This is further studied in Monte Carlo experiments illustrating the finite sample properties of the metaselection procedure. We mainly focus in this study in the metaselection associated to the two popular model selection strategies: AIC and BIC (Akaike's and Bayesian information criteria respectively).