Before the season started, LeBron James was the default MVP choice. In August, when ESPN Forecast pegged the likelihood of his victory being close to 80 percent, it was tough to argue. He was the best player in the world by almost every metric, a two-time reigning champion and the MVP four out of the past five years.
The rings, and the stats, suggested that while Durant might be a similarly effective offensive player, James on defense was one of the best in the world, and Durant was back in the pack.
The narrative changed, though, as the 2013-14 season unfolded.
The Pacers opened as the class of the league, making the Heat look sluggish by comparison. James entered the season fresh off a honeymoon, talking of back pain. Meanwhile, the best defensive metrics showed that the Heat played better defense with James on the bench.
That created an opening for Durant.
Durant's teammate Russell Westbrook missed a long stretch of the season with knee trouble. While he was out, Durant and the Thunder cruised. The team won 20 of 27 without Westbrook in one stretch, and Durant kicked off a torrid scoring stretch that continues to this day.
By January 2014, ESPN Forecast had switched camps and saw Durant as 59 percent likely to take the MVP title.
James responded somewhat -- for instance, with a 61-point scoring outburst in early March against the Bobcats. And as the Pacers swooned, the Heat -- who, like a lot of champions, showed signs of pacing themselves -- rose to the top of the East by stepping up play late in the season. But Durant's torrid play continued even with Westbrook back, and the Thunder played well enough as a team.
In March, the panel said Durant was 81 percent likely to win.
Durant scored 25 or more points in more consecutive games than Michael Jordan did, which raised a lot of eyebrows -- even as James ran out of games to wow MVP voters. By last weekend, the ESPN Forecast panel said Durant was 93 percent likely to win, which in the forecasting game is about as close to a lock as you can get.
Here are the results: