When Data Science Makes Us Sad: The Story of an Overbooked Flight

Airlines intentionally overbook flights as a calculated business decision based on data science and probability. Using the binomial distribution, it is possible to calculate the low probability of more passengers showing up than there are seats, based on historical no-show rates. The concept of expected value demonstrates that over many flights, the cost of compensating the very few bumped passengers is minuscule compared to the millions in revenue gained from selling otherwise empty seats. While there is a risk of negative publicity, airlines often mitigate this by offering vouchers to volunteers, ensuring the flight remains full and profitable without involuntary removals.