Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated: 1 1 00:00:01,350 --> 00:00:03,880 One of the reasons estimating a deliverable 2 2 00:00:03,880 --> 00:00:06,150 by rolling up the estimates of its activities 3 3 00:00:06,150 --> 00:00:08,840 provides the best accuracy is because of 4 4 00:00:08,840 --> 00:00:11,423 the statistical power of multiple estimates. 5 5 00:00:12,460 --> 00:00:15,370 That is, over many estimates some of the 6 6 00:00:15,370 --> 00:00:17,873 individual error cancels out. 7 7 00:00:19,470 --> 00:00:22,463 Any one estimate may have considerable error. 8 8 00:00:23,470 --> 00:00:26,590 However, the sum of multiple estimates 9 9 00:00:26,590 --> 00:00:29,293 converges close to the correct value. 10 10 00:00:31,320 --> 00:00:33,800 This helps the accuracy of the project level 11 11 00:00:33,800 --> 00:00:36,530 planning estimate twice. 12 12 00:00:36,530 --> 00:00:40,500 We estimate the sum of multiple deliverables and we estimate 13 13 00:00:40,500 --> 00:00:43,883 the deliverables as the sum of multiple activities. 14 14 00:00:45,470 --> 00:00:50,050 Doing both is the key reason the overall planning estimates 15 15 00:00:50,050 --> 00:00:52,940 will have an accuracy of plus or minus 10% 16 16 00:00:52,940 --> 00:00:55,063 experience shows again and again. 17 17 00:00:56,960 --> 00:00:57,860 Here's an example. 18 18 00:00:58,770 --> 00:01:01,713 Let's say the correct value of something is 100. 19 19 00:01:02,950 --> 00:01:06,130 Individual estimates from a normal statistical distribution 20 20 00:01:06,130 --> 00:01:08,970 will vary widely, but the sum 21 21 00:01:08,970 --> 00:01:11,663 of many estimates will be close. 22 22 00:01:13,520 --> 00:01:17,140 This table shows 20 estimates generated at random 23 23 00:01:17,140 --> 00:01:18,490 from a normal distribution. 24 24 00:01:19,530 --> 00:01:22,230 As you can see from columns two and three 25 25 00:01:22,230 --> 00:01:24,270 they jump all over the place, 26 26 00:01:24,270 --> 00:01:28,313 less than and greater than 100, with a wide error range. 27 27 00:01:29,400 --> 00:01:32,680 However, if we average them, as shown in the last two 28 28 00:01:32,680 --> 00:01:36,593 columns, we see that the error range is much less. 29 29 00:01:38,040 --> 00:01:40,490 Here are the estimates shown in a graph 30 30 00:01:40,490 --> 00:01:42,630 with the individual estimates in blue, 31 31 00:01:42,630 --> 00:01:45,220 with a wide error range, and the 32 32 00:01:45,220 --> 00:01:47,530 average of the estimates shown in green, 33 33 00:01:47,530 --> 00:01:49,633 with much more accuracy. 34 34 00:01:51,110 --> 00:01:53,990 Indeed, summing up 14 estimates or more 35 35 00:01:53,990 --> 00:01:57,630 the accuracy is well within 10%. 36 36 00:01:57,630 --> 00:01:59,970 This is a fairly typical result. 37 37 00:01:59,970 --> 00:02:03,130 As a rule of thumb if you have a project broken into 38 38 00:02:03,130 --> 00:02:06,980 15 deliverables or more and each deliverable is estimated 39 39 00:02:06,980 --> 00:02:08,960 by breaking it down into individual activities 40 40 00:02:08,960 --> 00:02:11,610 and rolling the estimate up, you should feel 41 41 00:02:11,610 --> 00:02:14,030 fairly confident the estimate for the whole project 42 42 00:02:14,030 --> 00:02:17,390 is pretty close, likely within plus or minus 10% 43 43 00:02:17,390 --> 00:02:20,140 if your core project team has experience in the domain. 44 44 00:02:21,970 --> 00:02:24,523 Every once in a while the universe gives you one. 45 45 00:02:25,580 --> 00:02:28,890 Simply break your project up into deliverables and then 46 46 00:02:28,890 --> 00:02:32,480 estimate the deliverables by breaking them into activities 47 47 00:02:32,480 --> 00:02:36,260 and that process alone will vastly increase the accuracy 48 48 00:02:36,260 --> 00:02:39,743 of your overall project plan budget and schedule estimate. 4324