U.S. Presidential elections are often quite close, although the electoral college system tends to distort election results so that they appear to be more decisive than they are. The 2012 election is a good example. Although Obama won the electoral vote 332 to 206 – about 62% to 38% – the popular vote was split about 51% to 49%. It makes one wonder whether things might have easily gone the other way, given random factors nobody had control over.
Yahoo News has created a cool tool for investigating the role of chance in presidential elections. For any election from 1944 to 2008, you can specify a percentage of random fluctuation in the state-by-state vote to apply, and run a do-over once or many times to see how often that degree of chance might have changed the outcome.