WEBVTT 00:00:06.983 --> 00:00:10.684 As the story goes, the legendary marksman William Tell 00:00:10.684 --> 00:00:15.430 was forced into a cruel challenge by a corrupt lord. 00:00:15.430 --> 00:00:17.652 William's son was to be executed 00:00:17.652 --> 00:00:21.959 unless William could shoot an apple off his head. 00:00:21.959 --> 00:00:26.933 William succeeded, but let's imagine two variations on the tale. 00:00:26.933 --> 00:00:28.573 In the first variation, 00:00:28.573 --> 00:00:33.070 the lord hires a bandit to steal William's trusty crossbow, 00:00:33.070 --> 00:00:37.341 so he is forced to borrow an inferior one from a peasant. 00:00:37.341 --> 00:00:41.357 However, the borrowed crossbow isn't adjusted perfectly, 00:00:41.357 --> 00:00:43.446 and William finds that his practice shots 00:00:43.446 --> 00:00:47.752 cluster in a tight spread beneath the bullseye. 00:00:47.752 --> 00:00:52.608 Fortunately, he has time to correct for it before it's too late. 00:00:52.608 --> 00:00:54.372 Variation two: 00:00:54.372 --> 00:00:58.805 William begins to doubt his skills in the long hours before the challenge 00:00:58.805 --> 00:01:01.502 and his hand develops a tremor. 00:01:01.502 --> 00:01:04.619 His practice shots still cluster around the apple 00:01:04.619 --> 00:01:06.677 but in a random pattern. 00:01:06.677 --> 00:01:08.732 Occasionally, he hits the apple, 00:01:08.732 --> 00:01:12.619 but with the wobble, there is no guarantee of a bullseye. 00:01:12.619 --> 00:01:14.512 He must settle his nervous hand 00:01:14.512 --> 00:01:19.201 and restore the certainty in his aim to save his son. 00:01:19.201 --> 00:01:23.639 At the heart of these variations are two terms often used interchangeably: 00:01:23.639 --> 00:01:26.369 accuracy and precision. 00:01:26.369 --> 00:01:27.942 The distinction between the two 00:01:27.942 --> 00:01:31.517 is actually critical for many scientific endeavours. 00:01:31.517 --> 00:01:35.501 Accuracy involves how close you come to the correct result. 00:01:35.501 --> 00:01:39.636 Your accuracy improves with tools that are calibrated correctly 00:01:39.636 --> 00:01:42.013 and that you're well-trained on. 00:01:42.013 --> 00:01:43.714 Precision, on the other hand, 00:01:43.714 --> 00:01:48.212 is how consistently you can get that result using the same method. 00:01:48.212 --> 00:01:52.034 Your precision improves with more finely incremented tools 00:01:52.034 --> 00:01:54.511 that require less estimation. 00:01:54.511 --> 00:01:59.327 The story of the stolen crossbow was one of precision without accuracy. 00:01:59.327 --> 00:02:02.888 William got the same wrong result each time he fired. 00:02:02.888 --> 00:02:08.065 The variation with the shaky hand was one of accuracy without precision. 00:02:08.065 --> 00:02:11.241 William's bolts clustered around the correct result, 00:02:11.241 --> 00:02:15.449 but without certainty of a bullseye for any given shot. 00:02:15.449 --> 00:02:18.179 You can probably get away with low accuracy 00:02:18.179 --> 00:02:21.076 or low precision in everyday tasks. 00:02:21.076 --> 00:02:24.580 But engineers and researchers often require accuracy 00:02:24.580 --> 00:02:30.262 on microscopic levels with a high certainty of being right every time. 00:02:30.262 --> 00:02:32.772 Factories and labs increase precision 00:02:32.772 --> 00:02:36.333 through better equipment and more detailed procedures. 00:02:36.333 --> 00:02:39.170 These improvements can be expensive, so managers must decide 00:02:39.170 --> 00:02:44.013 what the acceptable uncertainty for each project is. 00:02:44.013 --> 00:02:46.098 However, investments in precision 00:02:46.098 --> 00:02:49.317 can take us beyond what was previously possible, 00:02:49.317 --> 00:02:51.532 even as far as Mars. 00:02:51.532 --> 00:02:54.551 It may surprise you that NASA does not know exactly where 00:02:54.551 --> 00:02:58.535 their probes are going to touch down on another planet. 00:02:58.535 --> 00:03:02.484 Predicting where they will land requires extensive calculations 00:03:02.484 --> 00:03:06.247 fed by measurements that don't always have a precise answer. 00:03:06.247 --> 00:03:11.254 How does the Martian atmosphere's density change at different elevations? 00:03:11.254 --> 00:03:14.049 What angle will the probe hit the atmosphere at? 00:03:14.049 --> 00:03:17.227 What will be the speed of the probe upon entry? 00:03:17.227 --> 00:03:20.764 Computer simulators run thousands of different landing scenarios, 00:03:20.764 --> 00:03:24.391 mixing and matching values for all of the variables. 00:03:24.391 --> 00:03:26.058 Weighing all the possibilities, 00:03:26.058 --> 00:03:29.439 the computer spits out the potential area of impact 00:03:29.439 --> 00:03:32.840 in the form of a landing ellipse. 00:03:32.840 --> 00:03:37.528 In 1976, the landing ellipse for the Mars Viking Lander 00:03:37.528 --> 00:03:44.336 was 62 x 174 miles, nearly the area of New Jersey. 00:03:44.336 --> 00:03:45.918 With such a limitation, 00:03:45.918 --> 00:03:50.608 NASA had to ignore many interesting but risky landing areas. 00:03:50.608 --> 00:03:53.975 Since then, new information about the Martian atmosphere, 00:03:53.975 --> 00:03:56.451 improved spacecraft technology, 00:03:56.451 --> 00:04:02.333 and more powerful computer simulations have drastically reduced uncertainty. 00:04:02.333 --> 00:04:06.186 In 2012, the landing ellipse for the Curiosity Lander 00:04:06.186 --> 00:04:10.046 was only 4 miles wide by 12 miles long, 00:04:10.046 --> 00:04:14.251 an area more than 200 times smaller than Viking's. 00:04:14.251 --> 00:04:18.492 This allowed NASA to target a specific spot in Gale Crater, 00:04:18.492 --> 00:04:23.341 a previously un-landable area of high scientific interest. 00:04:23.341 --> 00:04:26.199 While we ultimately strive for accuracy, 00:04:26.199 --> 00:04:30.480 precision reflects our certainty of reliably achieving it. 00:04:30.480 --> 00:04:32.501 With these two principles in mind, 00:04:32.501 --> 00:04:34.202 we can shoot for the stars 00:04:34.202 --> 00:04:37.121 and be confident of hitting them every time.