Two issues with your calculation: (1) You are considering contiguous 9-year periods across which the 14-person sets can be applied with the 48 years + 25. Because the span of time is recorded history, and with a brief look at the Forbes list, at least 1000 years are covered... or 966 sets of 9-year periods (1000-9-25), rather than 5.

(2) Life expectancy globally across this period of time gives a much wider range.

Overall, your calcs are too narrow to disprove Gladwell's assertion. We'd have to dig more deep into the actual numbers to get the real stats to show how likely/unlikely this is.