An exponential waiting time has survival probability , with constant rate . Conditioning on already having waited divides the tail by , leaving the same distribution for the additional wait . This follows from the exponential model; it is not a rule for every real queue or device lifetime.指数等待时间的生存概率为 ,其中速率 恒定。已等待 而事件尚未发生时,把剩余尾部除以 ,得到的额外等待时间 仍服从原分布。这是指数模型的性质,不是所有现实队列或设备寿命都满足的规律。
The upper curve uses and marks elapsed times and . Their survival-probability ratio stays near 0.247 as moves. Below, the normalized green tail overlaps the gold original survival curve when both are plotted against additional waiting time .上图取 ,标出已等待时间 与 。随着 移动,两处生存概率的比值始终约为 0.247。下图改用额外等待时间 为横轴,归一化后的绿色尾部与金色的原生存曲线重合。