Calibrating - random model

Another naive model, useful mostly for calibration. We know that we want to add Democracia Ourensana and Sumar Galicia, but we still have not decided about voting likelihood for each case. An interesting option to calibrate the model is to see what happens if we assign the same likelihood for each party, drawing random samples from a normal distribution.

Results are shown below. As expected, each run shows slight differences, but all parties are there. Seats attribution are also random over ballot boxes, meaning that each ballot box has the same probability regardless of any feature, including circunscription. This is obviously not the case for Democracia Ourensana, which is expected to retrieve votes inside its own province only. That’s the next task to solve.