Musk certainly didn't explain the drop in mileage. As I understand it, that drop results from temperature management of the battery. When stopped in very cold weather, the car uses power to generate heat in order to maintain the battery within safe operating temperatures. In addition, when the battery is at the low end of that range, it under-reports the available mileage. That's what generated the Tesla tech's suggestion that it might regain some mileage while driven that morning. (You can see this in the graph. The slope of the graph in the little section from miles 400-411 is less than at any other point, and the section after the Norwich charge is also lower, although speeds were also lower here) However, after some time of driving or charging, the batteries will have risen to optimal operating temperature, and the range projection is correct. Mr. Broder may have assumed that he was going to recover all of the energy lost overnight, not just the little bit of error in projection. That may be why he convinced himself to leave with insufficient range to reach his destination. However, that argument fails Occam's razor in light of his earlier attempts to drive the car to failure and his departure expressly against the advice of the Tesla tech he talked to at that time.
As you can see, there is nothing special about his night stop. The night before, it predicted 79 miles. He ended up going 60 miles before it was out of juice. That's 25% shorter than the prediction, so the car must have lost charge during the night or due to cold weather, right? Well, if you look at the day before, the full charge predicted a range of 242 miles, but he only got 196 miles, which is 20% less than the prediction.
In other words: The predicted range is consistently higher than the actual range.
