◆ SpookStack

probabilities using index-count filtering are larger than those for exponential filtering. For Titan, the results are mixed, further suggesting that Titan reliability has not improved in recent years. For comparison purposes, the same filtering techniques have been applied to all flight tests shown in the tables of Appendix D, regardless of configuration. The results are presented in Table 3. Table 3. Predicted Failure Probabilities for All Configurations Vehicle Atlas Delta Flight Phase 0 0-1 0-2 0-3 0-4 0-5 • 0 0-1. 0-2 0-3 Filter Technic ue Sample Failures /Total 0/7 56/532 91/532 111/532 114/532 137/532 0/196 4/232 6/232 10/232 Equal Weight 0 0.1053 0.1711 0.2086 0.2143 0.2575 0 0.0172 0.0259 0.0431 Index Expon. Expon Count F =0.99 F=0.98 0 0 0 0.0641 0.0422 0.0273 0.0990 0.0555 0.0311 0.1261 0.0802 0.0559 0.1330 0.0873 0.0627 0.1671 0.1150 0.0866 0 0 0 0.0164 0.0148 0.0110 0.0232 0.0201 0.0133 0.0279 0.0263 0.0150 Expon F =0.97 0 0.0190 0.0204 0.0455 0.0511 0.0725 0 0.0077 0.0085 0.0089 0-4 0-5* Titan 0 0-1 0-2 0-3 0-4 0-5· 0.0431 0.1078 0.0306 0.0534 0.1424 0.1632 0.1662 0.1958· 0.0279 0.0766 0.0137 0.0319 0.0771 0.0924 0.0942 0.1369 0.0263 0.0740 0.0187 0.0351 0.0719 0.0830 0.0840 0.1326 0.0150 0.0536 0.0281 0.0399 0.0662 0.0711 0.0712 0.1277 0.0089 10/232 0.0459 25/232 0.0349 3/98 0.0467 18/337 0.0750 48/337 0.0770 55/337 0.0771 56/337 0.1346 66/337 • Includes response mode 'NA' . A comparison of Table 2 and Table 3 shows that in most cases, but not all, exponential filtering produces failure probabilities for the representative configuration samples that are smaller than the corresponding probabilities for the all-configuration samples. The fact that most differences between corresponding samples are relatively small attests to the effectiveness of the exponential filter in down-weighting early launch failures. This is not the case for equal weighting of tests, where the predicted failure probabilities based on all configurations are up to 3.6 times as large. With respect to- the weighting of missile and space-vehicle performance data, RTI favors an exponential filter over either the equal-weight or index-count filters. Weighting percentages for the three filters are given in Table 4 for sample sizes of 4 to 1,000. Except for small samples, the percentages produced by equal weighting place too much emphasis on old data, thus failing to account for the learning process and 9/10/96 18 RTI