Downloaded from https://www.everyspec.com AMCP 706198 Examle No. 31 ѕ Same prior degree-of-belief as Example No. 30, but the test results are 3 fail- ures in 10 hr. What is your new degree-of-belief? Assume that the conjugate dis- . tribution applies. Steps 1 and 2 are the same as in Example No. 30. Procedure Example 3. Find the failuie rate and Its un- 3. Nt,tat - 3/10-hr certainty for the test results only. = 300/1000-hr. coeff. of variation - 1/v= 58%. 4. Combine the prior and test results, 4. Xnt,qfta = (3 + 4)/(10 hr + 1000 hr) A per conjugate distribution theory. 7/1000-hr coeff. of variittion = 1/4W7 38%.
Now, only a fool really believes that the failure rate is 7/1000-hr (about twice the prior belief). Anyone else is very worried about those test results. So, the prior degree-of-belief was most inappropriate. Perhaps r - 0.4 and T - 200 hr would have been better choices for the prior degree-of-belief. This conjugate method of choosing prior degree-of-belief does not have the flexi- bility that the discrete method dos, and it it easier to lead oneself astray. -- "A-.-.-