Two samples, 900 trips each, drawn on one set of bins. Driving has the lower typical time, so on the averages it wins and the question looks settled. The shapes say otherwise: the car's distribution is wide with a tail that runs off to the right, while the train's is narrow and stops. You are choosing between a faster average and a bad day you can plan around. Overlaying the two is what makes that legible; side by side, you end up comparing bar heights instead of shapes.