Testing fixed probabilities

The pace of the electoral process is accelerating. Before we had the necessary time to include provincial parties to cover the case of Democracia Ourensana, a first survey arrived, urging us to push a first estimation. We decided to do so using a fixed probabilities model, where each party gets a probability which is independent of any other consideration, such as circumscription, prior vote, or demographics (hence the ‘fixed’ word).

Results are below. We played with Dou and Sumar, leaving one below and the other one over the electoral barrier (5% in Galicia). The model reacts correctly, leaving out the one below. However, seats allocation is plainly wrong for the case of provincial-only parties. While we find the way to do this, feel free to enjoy our first estimation of the measured data!

Fixed probabilities: