Author Topic: Accuracy vs. Uncertainty  (Read 2824 times)

0 Members and 1 Guest are viewing this topic.

Offline nikkoderic

  • Contributor
  • Posts: 22
  • Country: ph
  • We're all human, we're just like you man.
Accuracy vs. Uncertainty
« on: May 20, 2015, 01:14:59 pm »
Hello! I just want to ask whats the difference between accuracy and uncertainty. I already did try to search it on google but I still dont get it. All I understood is that accuracy is qualitative and uncertainty is quantitative. I want to know the difference between the two in practical terms. Apologies for my crude english. :D If you didnt get my question then maybe share your insight about these two. Thanks!
 

Offline nfmax

  • Frequent Contributor
  • **
  • Posts: 858
  • Country: gb
Re: Accuracy vs. Uncertainty
« Reply #1 on: May 20, 2015, 02:02:26 pm »
The official definitions are available from the JCGM (Joint Committee for Guides in Metrology) on-line vocabulary here: http://www.iso.org/sites/JCGM/VIM/JCGM_200e_FILES/MAIN_JCGM_200e/Start_e.html. Accuracy is defined at 2.13 and uncertainty at 2.26. There is an equivalent French language guide on the same suite, if that is easier for you.

And congratulations on asking this question on World Metrology Day! http://www.worldmetrologyday.org/
 

Offline Asmyldof

  • Regular Contributor
  • *
  • Posts: 148
  • Country: nl
  • Freelancer - Persnicketist - Do'er of stuff
    • Asmyldof's Home. It's old, not quite impressive, but it's there.
Re: Accuracy vs. Uncertainty
« Reply #2 on: May 20, 2015, 03:01:22 pm »
Simply put it comes down to this;
The accuracy is how much a single measurement expresses the actual value. In other words: Off by how much can your measurement be.

The uncertainty describes how much the variation in the accuracy can be of a set of measurements. So, if you do 10 measurements you get 10 different results. The "distance" between those results can be considered a term of uncertainty.

For example, if you shoot at a target with a single gun:
The accuracy of your gun is the definition of, assuming your aim is perfect at the bullseye of your target, how much off each shot can be at maximum.
The uncertainty can be a description of the variation of that accuracy.

In most metrology accuracy can be compensated in part, but uncertainty is an absolute statistical and cannot be "tuned out".
In um movement systems people often are more scared of uncertainty than inaccuracy.

EDIT: It can be more complex in situations and of course in both there's a lot of maths which can link them, but for understanding's sake I left out some of the commas and caveats.
If it's a puzzle, I want to solve it.
If it's a problem, I need to solve it.
If it's an equation... mjeh, I've got Matlab
...
...
(not really though, Matlab annoys me).
 

Offline suicidaleggroll

  • Super Contributor
  • ***
  • Posts: 1455
  • Country: us
Re: Accuracy vs. Uncertainty
« Reply #3 on: May 20, 2015, 03:07:43 pm »
Simply put it comes down to this;
The accuracy is how much a single measurement expresses the actual value. In other words: Off by how much can your measurement be.

The uncertainty describes how much the variation in the accuracy can be of a set of measurements. So, if you do 10 measurements you get 10 different results. The "distance" between those results can be considered a term of uncertainty.

That's precision, not uncertainty.

Accuracy is how far "off target" your average measurement is.
Precision is the spread of your measurements.
Uncertainty is the combination of accuracy and precision, it's the "catch all" if you will.

You can have high accuracy but low precision, you can have high precision but low accuracy, but in either case you have poor uncertainty.  The only way to have good uncertainty is with both high accuracy and high precision.

At least that's the way I've always understood it.
« Last Edit: May 20, 2015, 03:09:21 pm by suicidaleggroll »
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 10109
  • Country: gb
    • Having fun doing more, with less
Re: Accuracy vs. Uncertainty
« Reply #4 on: May 20, 2015, 03:39:46 pm »
Simply put it comes down to this;
The accuracy is how much a single measurement expresses the actual value. In other words: Off by how much can your measurement be.

The uncertainty describes how much the variation in the accuracy can be of a set of measurements. So, if you do 10 measurements you get 10 different results. The "distance" between those results can be considered a term of uncertainty.

I know of one instrument which had a stability and repeatability of 0.001dB, and an accuracy of 0.1dB.

Since it was designed to measure changes in a cable's impedance as a function of temperature, nobody gave a damn about the accuracy. When it takes a week to do a temperature test, you need an instrument that is stable/repeatable over at least that timescale.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline tggzzz

  • Super Contributor
  • ***
  • Posts: 10109
  • Country: gb
    • Having fun doing more, with less
Re: Accuracy vs. Uncertainty
« Reply #5 on: May 20, 2015, 03:44:03 pm »
The official definitions are available from the JCGM (Joint Committee for Guides in Metrology) on-line vocabulary here: http://www.iso.org/sites/JCGM/VIM/JCGM_200e_FILES/MAIN_JCGM_200e/Start_e.html. Accuracy is defined at 2.13 and uncertainty at 2.26. There is an equivalent French language guide on the same suite, if that is easier for you.

And congratulations on asking this question on World Metrology Day! http://www.worldmetrologyday.org/

Useful; thanks.
There are lies, damned lies, statistics - and ADC/DAC specs.
Glider pilot's aphorism: "there is no substitute for span". Retort: "There is a substitute: skill+imagination. But you can buy span".
Having fun doing more, with less
 

Offline Asmyldof

  • Regular Contributor
  • *
  • Posts: 148
  • Country: nl
  • Freelancer - Persnicketist - Do'er of stuff
    • Asmyldof's Home. It's old, not quite impressive, but it's there.
Re: Accuracy vs. Uncertainty
« Reply #6 on: May 20, 2015, 04:09:48 pm »
That's precision, not uncertainty.

Accuracy is how far "off target" your average measurement is.
Precision is the spread of your measurements.
Uncertainty is the combination of accuracy and precision, it's the "catch all" if you will.

(...)

At least that's the way I've always understood it.

You are right, of course. Corrected, I stand it.
If it's a puzzle, I want to solve it.
If it's a problem, I need to solve it.
If it's an equation... mjeh, I've got Matlab
...
...
(not really though, Matlab annoys me).
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf