0:00:06.983,0:00:10.684
As the story goes, [br]the legendary marksman William Tell

0:00:10.684,0:00:15.430
was forced into a cruel challenge[br]by a corrupt lord.

0:00:15.430,0:00:17.652
William's son was to be executed[br]

0:00:17.652,0:00:21.959
unless William could shoot [br]an apple off his head.

0:00:21.959,0:00:26.933
William succeeded, but let's imagine[br]two variations on the tale.

0:00:26.933,0:00:28.573
In the first variation,

0:00:28.573,0:00:33.070
the lord hires a bandit to steal[br]William's trusty crossbow,

0:00:33.070,0:00:37.341
so he is forced to borrow[br]an inferior one from a peasant.

0:00:37.341,0:00:41.357
However, the borrowed crossbow[br]isn't adjusted perfectly,

0:00:41.357,0:00:43.446
and William finds that his practice shots

0:00:43.446,0:00:47.752
cluster in a tight spread [br]beneath the bullseye.

0:00:47.752,0:00:52.608
Fortunately, he has time [br]to correct for it before it's too late.

0:00:52.608,0:00:54.372
Variation two:

0:00:54.372,0:00:58.805
William begins to doubt his skills[br]in the long hours before the challenge

0:00:58.805,0:01:01.502
and his hand develops a tremor.

0:01:01.502,0:01:04.619
His practice shots still cluster[br]around the apple

0:01:04.619,0:01:06.677
but in a random pattern.

0:01:06.677,0:01:08.732
Occasionally, he hits the apple,

0:01:08.732,0:01:12.619
but with the wobble,[br]there is no guarantee of a bullseye.

0:01:12.619,0:01:14.512
He must settle his nervous hand[br]

0:01:14.512,0:01:19.201
and restore the certainty in his aim[br]to save his son.

0:01:19.201,0:01:23.639
At the heart of these variations[br]are two terms often used interchangeably:

0:01:23.639,0:01:26.369
accuracy and precision.

0:01:26.369,0:01:27.942
The distinction between the two

0:01:27.942,0:01:31.517
is actually critical [br]for many scientific endeavours.

0:01:31.517,0:01:35.501
Accuracy involves how close you come[br]to the correct result.

0:01:35.501,0:01:39.636
Your accuracy improves with tools[br]that are calibrated correctly

0:01:39.636,0:01:42.013
and that you're well-trained on.

0:01:42.013,0:01:43.714
Precision, on the other hand,

0:01:43.714,0:01:48.212
is how consistently you can get [br]that result using the same method.

0:01:48.212,0:01:52.034
Your precision improves [br]with more finely incremented tools

0:01:52.034,0:01:54.511
that require less estimation.

0:01:54.511,0:01:59.327
The story of the stolen crossbow[br]was one of precision without accuracy.

0:01:59.327,0:02:02.888
William got the same wrong result[br]each time he fired.

0:02:02.888,0:02:08.065
The variation with the shaky hand[br]was one of accuracy without precision.

0:02:08.065,0:02:11.241
William's bolts clustered [br]around the correct result,

0:02:11.241,0:02:15.449
but without certainty of a bullseye[br]for any given shot.

0:02:15.449,0:02:18.179
You can probably get away [br]with low accuracy

0:02:18.179,0:02:21.076
or low precision in everyday tasks.

0:02:21.076,0:02:24.580
But engineers and researchers[br]often require accuracy

0:02:24.580,0:02:30.262
on microscopic levels with [br]a high certainty of being right every time.

0:02:30.262,0:02:32.772
Factories and labs increase precision

0:02:32.772,0:02:36.333
through better equipment [br]and more detailed procedures.

0:02:36.333,0:02:39.170
These improvements can be expensive,[br]so managers must decide

0:02:39.170,0:02:44.013
what the acceptable uncertainty [br]for each project is.

0:02:44.013,0:02:46.098
However, investments in precision

0:02:46.098,0:02:49.317
can take us beyond [br]what was previously possible,

0:02:49.317,0:02:51.532
even as far as Mars.

0:02:51.532,0:02:54.551
It may surprise you that NASA[br]does not know exactly where

0:02:54.551,0:02:58.535
their probes are going to touch down[br]on another planet.

0:02:58.535,0:03:02.484
Predicting where they will land[br]requires extensive calculations

0:03:02.484,0:03:06.247
fed by measurements [br]that don't always have a precise answer.

0:03:06.247,0:03:11.254
How does the Martian atmosphere's density[br]change at different elevations?

0:03:11.254,0:03:14.049
What angle will the probe[br]hit the atmosphere at?

0:03:14.049,0:03:17.227
What will be the speed [br]of the probe upon entry?

0:03:17.227,0:03:20.764
Computer simulators run thousands[br]of different landing scenarios,

0:03:20.764,0:03:24.391
mixing and matching values[br]for all of the variables.

0:03:24.391,0:03:26.058
Weighing all the possibilities,

0:03:26.058,0:03:29.439
the computer spits out [br]the potential area of impact

0:03:29.439,0:03:32.840
in the form of a landing ellipse.

0:03:32.840,0:03:37.528
In 1976, the landing ellipse[br]for the Mars Viking Lander

0:03:37.528,0:03:44.336
was 62 x 174 miles,[br]nearly the area of New Jersey.

0:03:44.336,0:03:45.918
With such a limitation,

0:03:45.918,0:03:50.608
NASA had to ignore many interesting[br]but risky landing areas.

0:03:50.608,0:03:53.975
Since then, new information [br]about the Martian atmosphere,

0:03:53.975,0:03:56.451
improved spacecraft technology,

0:03:56.451,0:04:02.333
and more powerful computer simulations[br]have drastically reduced uncertainty.

0:04:02.333,0:04:06.186
In 2012, the landing ellipse [br]for the Curiosity Lander

0:04:06.186,0:04:10.046
was only 4 miles wide by 12 miles long,

0:04:10.046,0:04:14.251
an area more than 200 times [br]smaller than Viking's.

0:04:14.251,0:04:18.492
This allowed NASA to target [br]a specific spot in Gale Crater,

0:04:18.492,0:04:23.341
a previously un-landable area [br]of high scientific interest.

0:04:23.341,0:04:26.199
While we ultimately strive for accuracy,

0:04:26.199,0:04:30.480
precision reflects our certainty[br]of reliably achieving it.

0:04:30.480,0:04:32.501
With these two principles in mind,

0:04:32.501,0:04:34.202
we can shoot for the stars

0:04:34.202,0:04:37.121
and be confident [br]of hitting them every time.