We consider the problem of dynamically rating sports teams based on the categorical outcome of paired comparisons such a s w i n , d r a w and loss in football. Our modelling framework is the cumulative link model for ordered response, where latent parameters represent the strength of each team. A dynamic extension of this model is proposed with close connections to nonparametric smoothing methods. As a consequence, recent results have more in uence for estimating current abilities than results

more »

... in the past. We highlight t h e importance of using a speci c constrained random walk prior for time{ changing abilities which guarantees an equal treatment of all teams. Estimation is done within an extended Kalman lter type approach. An additional hyperparameter which determines the temporal dynamic of the latent team abilities is chosen based on optimal one-step-ahead predictive p o wer. Alternative estimation methods are also considered. We apply our method to the results from the German football league \Bundesliga" 1996/97 and to the results from the American National Basketball Association (NBA) 1996/97.