Our poll-based model has a perfect record projecting the winners in the 5 Presidential elections since 2000 and the 4 Senate projections since 2010. The general concensus is that the polls were wrong in 2016. But our model, which ignores all polls except the last poll in each state, proved to be accurate once again. Our model is designed to capture late voter trends that others generally miss by averaging previous polls.
We began our market-based analysis in 2016 and tracked the market projections for the Presidency and the Senate, based on the bid/ask pricing of market traded futures contracts on the PredictIt.org marketplace. The final projection going into election day was that there was a probability of 87.6% that Clinton would win and a probability of 65.8% that the Democrats would control the Senate. These market-based projections conflicted with our poll-based projections and proved to be inaccurate.