The Economist’s model of America’s congressional elections estimates both major parties’ chances of winning each House and Senate race, and of controlling each legislative chamber. The forecast combines national and race-level polls from FiftyPlusOne with fundamental data about past electoral results and the nominees in each contest, in order to estimate the probability of each possible outcome.

The model does this by first constructing 25,001 hypothetical scenarios for the national popular vote for the House of Representatives. In some simulations, Democrats will fare far better than expected across the country; in others, Republicans will. Next, it tacks on additional simulated uncertainty for how the outcomes of Senate races as a group may differ from those in the House.

Using these values, the model then estimates a different plausible range of outcomes in each race for each simulated national environment. So, for example, a strong Republican Senate candidate running in a historically Democratic state might have a 30% chance of victory amid a nationwide “red wave”, but just a 5% chance if the House popular vote is tied and Senate Republicans as a whole are underperforming their colleagues in the lower chamber. Finally, the forecast picks one outcome at random from these ranges for each race in each simulation.

For specific seats, the win probabilities presented here represent the share of these simulations won by each candidate. For chamber control, they represent the share of simulations in which the given party wins a majority of seats.