Actually, it is. The sample size (number of flights) is large enough to put an upper bound on the failure rate. Such reasoning is standard practice in reliability testing. And with 4 anomalies in 100 flights, the number of anomalies is also large enough to provide a reasonable lower bound to the...