Not a coincidence if you saw the recent monty hall problem post here (maybe a couple days ago). That gave me an itch to code up a side-by-side simulation with some controls to watch the results converge on the expected probabilities.