With the dust settling on the Democratic nomination, it's worth a retrospective look at how successfully our calculator was able to estimate delegate counts based on a state's popular vote. While most states allocate the majority of their pledged delegates based on district-by-district vote totals, the proportion of delegates won by each candidate has historically been very similar to the statewide popular vote. This was the fundamental assumption of our calculator.
Of the 16 contests the calculator adjudicated after launching in early March, predictions based on the popular vote were off by an average of only 1.9 percent. Six states and territories reported final delegate tallies that exactly matched the calculator's prediction: Vermont, Wyoming, Guam, Indiana, Oregon, and Montana. The calculator was least successful in predicting Texas, which might be expected due to the state's Byzantine primary system.
Slate would like to thank the dozens of readers who wrote in challenging our numbers and suggesting new features for the calculator over the course of this project.
- We estimate the number of delegates based on the overall state vote, even though delegates are awarded by congressional district as well. We felt comfortable making this approximation because in the primaries through Mississippi, there was only a 2.9 percent deviation between the percentage of the overall vote and the percentage of delegates awarded in primaries. The proportion of delegates awarded by congressional district, therefore, does not differ greatly from the statewide breakdown.
- The calculator does not incorporate superdelegates into its calculations. Superdelegates are unpledged and uncommitted and therefore can change their endorsements and convention votes at any time. As a result, we've simply noted at the bottom of the calculator how many superdelegates the leading candidate needs to win the nomination in a given scenario.
- All of the calculator's formulas and data come from Jason Furman, the director of the Hamilton Project at the Brookings Institution.