This article reports on an attempt to forecast the outcome of the 2013 election to the German Bundestag. In contrast to the predominant academic approach to forecast incumbent vote shares from measures of government popularity, economic conditions and other fundamental variables, we entirely relied on data from published trial heat polls. Opposite to common practice in the news media, we did not take isolated polls as election forecasts in their own right. Instead, we used historical data to assess empirically the relationship between polls and election outcomes, and combined extrapolations from current polls in a Bayesian manner. The forecast was published one month ahead of the election. The retrospective evaluation of our method was added after the election. While our method is parsimonious and provides a large lead time, the performance at the 2013 election was underwhelming. We offer additional suggestions how the approach can be improved in future scenarios.
