What is more important for process capability? An accurate process or a precise process...you can only choose one.

I was about to say both, but you specified to choose only one. Then i say its precision. My stand is "Variation is the enemy". If my process is predictible/consistent, then being off(in-accurate) is easier to remedy than being centered but unpredictible/inconsistent. :-)

I was about to say both, but you specified to choose only one. Then i say its precision. My stand is "Variation is the enemy". If my process is predictible/consistent, then being off(in-accurate) is easier to remedy than being centered but unpredictible/inconsistent. :-)

I tend to agree..but I am finding more and more clients that are working on centering their processes versus reducing variability. I say they are taking the easy way out.

Give me precision any day. I can sleep at night. It may not be the best, but I won't be surprised.

I tend to agree..but I am finding more and more clients that are working on centering their processes versus reducing variability. I say they are taking the easy way out.

I think it's the lack of knowledge or rather insufficient knowledge. Over here, most general textbooks on statistics, SPC or management (at the introductory or lower levels) drum into your mind about control limits and the need to keep within limits but is silent on issues like accuracy versus precision.

What is more important for process capability? An accurate process or a precise process...you can only choose one.
Covers,

I am not sure I agree with you.

Let’s take an example. Results from a Gage R&R

Gage R&R

%Contribution
Source VarComp (of VarComp)
Total Gage R&R 0.09143 7.76
Repeatability 0.03997 3.39
Reproducibility 0.05146 4.37
Operator 0.05146 4.37
Part-To-Part 1.08645 92.24
Total Variation 1.17788 100.00


Now let’s double the gage R&R

Gage R&R

%Contribution
Source VarComp (of VarComp)
Total Gage R&R 0.18286 7.76
Repeatability 0.03997 3.39
Reproducibility 0.05146 4.37
Operator 0.05146 4.37
Part-To-Part 1.08645 92.24
Total Variation 1.26984 100.00

This demonstrates that when we double the contribution from the gage, total variation increases by only 0.09143 units increasing the 6 sigma spread of the process by 0.25 units.

On the other hand, if we double the bias (accuracy) of the gage, the mean moves the entire distance of the change in the bias. Assume bias was 2 units and we double it to 4 units. Then one side of the capability curve moves an additional 2 units toward the specification on that side.

For my money, give me an accurate gage. Precision will take care of itself in the R&R process.

Best regards,

Jim Shelor
PMP, CSSBB

Great description, Jim:

Accuracy has it for me too. If your measurement isn't accurate, then the precision is useless, anyway......

The two go hand in hand of course, but I had to vote for accuracy. What good to have a repeatable low variability process if it is not accurate. Makes me think of the term "consistently wrong".

Accuracy is also more insidious than precision. At least a precision problem make itself well known by poor repeatability. If your accuracy changes, you just change the process to compensate without questioning the gage. At least until the next calibration cycle comes around, and you have to do a risk assessment.

This question reminds me of a story in a Quality book, I read a looong time ago. The story concerned two snipers trying to shoot each other. One sniper got off five rounds, all in the other snipers shoulder, grouped so closely you could cover them with a playing card. The wounded sniper drew a bead on the first and fired, hitting the first sniper in the head - killing him!

The moral of the story? Precision is 'pretty' - accuracy is deadly.....

First precision, then accuracy. It's no good trying to hit the target if you can't do it more than once.

First tighten control so as to bunch together the results. Then the whole group can be shifted with a carefully designed intervention. Eventually the ideal is precise accuracy.

Jim, excellent as always.

Question: could the answer depend (as Jim's post would suggest) on which tools (and expertise) in process improvement you have in place?

What is more important for process capability? An accurate process or a precise process...you can only choose one.

If we can only choose one, then everything is going to go to he11 in a hand basket in short order. The problem comes in when we think of the two as separable, and they're not. They aren't different goals; they're inseparable components of the same goal. Even if we do allow for separate consideration of precision and accuracy, the values aren't absolute; what's reasonably precise is still a function of what's reasonably accurate, and vice-versa. The purpose of a process capability study is to understand variation in the process, and "variation" includes proximity to the target (accuracy) and dispersion around it (precision). How can you consider one and not the other?

I voted precision as a inaccurate precise process is easier to correct as you know the process is consistent.

If it is accurate but imprecise you never are sure how it will react to a change.

Another example,

I used to competitively 10 pin bowl. For years and years I could be counted on for good consistent scoring, but could never get the big game (i.e. win $$ in pots). Since I knew I was consistent, when I made changes I was still precise but at a different target (resulted in 2 perfect games in 6 months).

If I wasn't precise, I may get the big game but I could just as easily cost my team a game (or series) if I was on the wrong side of my range of scores.

Instead, I could still be counted on but my worst game was much higher than prior to the changes.

If we can only choose one, then everything is going to go to he11 in a hand basket in short order. The problem comes in when we think of the two as separable, and they're not. They aren't different goals; they're inseparable components of the same goal. Even if we do allow for separate consideration of precision and accuracy, the values aren't absolute; what's reasonably precise is still a function of what's reasonably accurate, and vice-versa. The purpose of a process capability study is to understand variation in the process, and "variation" includes proximity to the target (accuracy) and dispersion around it (precision). How can you consider one and not the other?

Hello,

I agree with Jim.

If you have, let's say "0" variation in a manufacturing operation (it's impossible, but just to provide an example), you're darn "precise"; and if at the same time you're completely off target (let's say, outside the specifications); you're not "accurate" and have produced 100% crap.....

I believe you want both...Just a thought...

Stijloor.

Just to be precise and accurate in answering:notme:, are we talking about the process itself, or the instruments used to measure the process? Jim S addressed the gage specifcally, but others seem to be thinking more about the process itself.


The first question is how well you can make the actual parts. Some processes are very consistent once they are set up, but hard to adjust - precise but not accurate. Some show large but consistent random variation, but you can set the long-term average just where you want - accurate but not precise.

Once the parts are made and you want to measure them, then the gage matters. A gage that is not precise may not be able to capture small variation in parts, but that may not really matter if the parts are well within spec. A gage that is not accurate, on the other hand, may tell you that a part is in spec when it is really out of spec. (and people tend to assume that lots of digits on a readout implies lots of accuracy, which is not always the case.

Perhaps what is needed is a precise process and an accurate gage.


Tim F

Great description, Jim:

Accuracy has it for me too. If your measurement isn't accurate, then the precision is useless, anyway......

Dear Me,
How can you get precision if the measurement is inaccurate??Beats me
hollow,as to how you concluded about the precision to be useless.Would you be kind enough to explain your point of view,from your angle?!?

Umang:nope:

This question reminds me of a story in a Quality book, I read a looong time ago. The story concerned two snipers trying to shoot each other. One sniper got off five rounds, all in the other snipers shoulder, grouped so closely you could cover them with a playing card. The wounded sniper drew a bead on the first and fired, hitting the first sniper in the head - killing him!

The moral of the story? Precision is 'pretty' - accuracy is deadly.....

The dictionary meaning for 'Precision' is EXACT,meaning no variation,and
'Accuracy' means CORRECT,permitting variation.Therefore my interpretation
of your enlightening story is,that the five rounds fall in the catagory of
ACCURACy,and the single killing shot was fired with PRECISION;and hence
from my perception the moral of story is "Accuracy looks pretty-Precision
is deadly"

Comments invited from one and all:bigwave:

Umang

If we can only choose one, then everything is going to go to he11 in a hand basket in short order. The problem comes in when we think of the two as separable, and they're not. They aren't different goals; they're inseparable components of the same goal. Even if we do allow for separate consideration of precision and accuracy, the values aren't absolute; what's reasonably precise is still a function of what's reasonably accurate, and vice-versa. The purpose of a process capability study is to understand variation in the process, and "variation" includes proximity to the target (accuracy) and dispersion around it (precision). How can you consider one and not the other?

Hi Jim
Well said!! I am in full agreement with you.:yes:
One can be ACCURATELY PRECISE or PRECISELY ACCURATE.Both amounts to be one and the same thing.Only hair splitting the meaning brings Precision above the Accuracy.:agree:

Precision-The ability of a measurement to be consistently reproduced.

Accuracy- The ability of a measurement to match the actual value of the quantity being measured.

This is the dictionary definition that is appropriate in this context.


In all honesty the two go hand in hand as we have all pretty much agreed in this thread. However, I still would suggest that inaccurate readings which are highly repeatable are useless ;)

In all honesty the two go hand in hand as we have all pretty much agreed in this thread. However, I still would suggest that inaccurate readings which are highly repeatable are useless ;)

However, inaccurate parts (i.e. not centered) that are precise (i.e. repeatable) are valuable - for two reasons.
1) they may well be in spec even if they are not exactly centered.
2) It is often easier to center a process than to reduce variation.

Tim F

Precision-The ability of a measurement to be consistently reproduced.

Accuracy- The ability of a measurement to match the actual value of the quantity being measured.

This is the dictionary definition that is appropriate in this context.


Those aren't dictionary definitions; at least I don't think you'll find the differentiation we're familiar with in any standard dictionary. This is another of those instances where standard definitions have been augmented in order to convey meaning in a specialized sense, and the specialized usage isn't widespread enough for the lexicographers to bless it yet. Nothing wrong with that--it happens all the time in English--but the fact is that "accuracy" and "precision" are interchangeable in some senses of the words, in some contexts, and it's further reason that we need to make sure that transmitter and receiver are tuned to the same frequency when we're trying to communicate.

I don't put any credence into any versions of any dictionary...after all they now include slang and other garbage words in frequent use. However, those definitions are word for word from Websters directly in front of me, so that would suggest to me that they are 'dictionary definitions'. Further the definition suits the context of this question and reflects my original interpretation of the terms used in this context. If you have a better definition, go ahead and use it, your previous post however parallels those definitions in different wording (proximity/dispersion).

As I said before, and you reiterated, the two are synonyms, hand in hand, often mean the same. As you've pointed out neither seems more important than the other because they are both key to having a capable process...if your accuracy is insufficient the process is not good, if your precision is insufficient the process is not good.

Those aren't dictionary definitions; at least I don't think you'll find the differentiation we're familiar with in any standard dictionary.

Actually, a quick check at Dictionary.com shows that many of the dictionaries do make this distinction. Granted, this is only one of several definitions, and usually there is also a definition that equates precision and accuracy, but reading through the whole list of definitions clarifies the distinction.


Tim F

Actually, a quick check at Dictionary.com shows that many of the dictionaries do make this distinction. Granted, this is only one of several definitions, and usually there is also a definition that equates precision and accuracy, but reading through the whole list of definitions clarifies the distinction.
Tim F

You're right. I should have checked first:o. I'm glad to see that the dictionaries have caught up. I retract my earlier statement, and promise I'll never do it again. :cool:

Dear Me,
How can you get precision if the measurement is inaccurate??Beats me
hollow,as to how you concluded about the precision to be useless.Would you be kind enough to explain your point of view,from your angle?!?

Umang:nope:

Well, yes. When I was an apprentice, I ran a jig borer. It had very high accuracy and was also very precise. It had a set of optical 'sights' to make alignment of the machine quill over the center-line of the feature to be machined. It was often that when using this optic device, you could be very precisely, within a few 10,000th's of an inch, but at the wrong dimension, therefore not very accurate!

I am very excited by the debate that has ensued. Though those that voted for accuracy has their merits, I still content that precision in either a process or a measurement instrument is critical. If you know you are inaccurate (you have a bias), you can shift the process or measurement device to compensate for your inaccuracy. For example if I am consistently (highly precise) a dimension that is +0.2" off of truth, then I know I just take my reading and substract 0.2 to get the truth. But if my process/measurement is not precise, then I have no confidence in my value.

I used the story of the two snipers in my training. I also say if you have only 1 bullet than accuracy is important, because you cannot measure precision with only 1 value. But if both people shot 3 bullets, the accurate sniper would on average hit the head, but one bullet might hit the head, another the arm another miss the person entirely. On the other hand, the precise sniper would have shot the first bullet, hit the shoulder, moved their process and the next two bullets would hit the head!

Just another view on this topic.

The two go hand in hand of course, but I had to vote for accuracy. What good to have a repeatable low variability process if it is not accurate. Makes me think of the term "consistently wrong".

...but would it be better if you are consistently wrong rather than not knowing when you are wrong and when you are right?:2cents:

"This question reminds me of a story in a Quality book, I read a looong time ago. The story concerned two snipers trying to shoot each other. One sniper got off five rounds, all in the other snipers shoulder, grouped so closely you could cover them with a playing card. The wounded sniper drew a bead on the first and fired, hitting the first sniper in the head - killing him!

The moral of the story? Precision is 'pretty' - accuracy is deadly.....
"
accuracy is deadly ...if you are lucky...:2cents: :cool:

If you know you are inaccurate (you have a bias), you can shift the process or measurement device to compensate for your inaccuracy. For example if I am consistently (highly precise) a dimension that is +0.2" off of truth, then I know I just take my reading and substract 0.2 to get the truth.

This is the crux of the matter. If your accuracy changes between calibration cycles, you will often not know it. Instead, you think the process has changed when it has not. Most would then adjust the process back to nominal "as defined by the biased gage". In reality they have shifted the process off target.

This is the crux of the matter. If your accuracy changes between calibration cycles, you will often not know it. Instead, you think the process has changed when it has not. Most would then adjust the process back to nominal "as defined by the biased gage". In reality they have shifted the process off target.

Changes in accuracy...means you are not precise on some level. I am not talking about looking a processes over time..I am saying at any given time.

Have attached a small file showing my understanding of Precision and Accuracy.

Making a decision I would side with “Accuracy”, precision could always be achieved at a cost:

Why “grind” when “turning” meets the required precision.
Why “polish” a surface when “shaping” meets the required precision.
Why use a “1%” resistor when a “5%” meets the required precision.

In the above examples “accuracy” required is the guidance’s to the requirements.

Have attached a small file showing my understanding of Precision and Accuracy.

Making a decision I would side with “Accuracy”, precision could always be achieved at a cost:
In the above examples “accuracy” required is the guidance’s to the requirements.

Ahhh, now i understand your(plural) point. i think the difference in our point of view is on how we picture the problem.

pls look at the attached and compare it with the previous 1. would you still side with accuracy?

Have attached a small file showing my understanding of Precision and Accuracy.

Making a decision I would side with “Accuracy”, precision could always be achieved at a cost:

Why “grind” when “turning” meets the required precision.
Why “polish” a surface when “shaping” meets the required precision.
Why use a “1%” resistor when a “5%” meets the required precision.

In the above examples “accuracy” required is the guidance’s to the requirements.

Interesting, you would take the process in the upper right that uses the entire "specification" versus a process in the lower left that with a lower cost could be centered?

Both precision and accuracy come at a cost. It is cheaper to fix accuracy issues than it is to fix precision issues.

Well, yes. When I was an apprentice, I ran a jig borer. It had very high accuracy and was also very precise. It had a set of optical 'sights' to make alignment of the machine quill over the center-line of the feature to be machined. It was often that when using this optic device, you could be very precisely, within a few 10,000th's of an inch, but at the wrong dimension, therefore not very accurate!

Despite being precise to a mili inch,if you are directing it at a wrong place;
then user is to be blamed not the precision.A surgeon with precision,wants
to perform surgery on the Kidney;in stead he operates Liver with precision.
The fault here lies with the surgeon,and not with precision.
2.In a machine shop all the production machines viz. Lathe,Milling,shaping
etcetra have to be ACCURATE;but the similar machines for a tool room
have to be,and are called 'PRECISION MACHINES';needed to achieve very
high degree of accuracy.
There are many more examples like Caliper and Vernier Caliper,the later
being a 'Precision instrument'.All the instruments have got to be accurate,
but Precision instruments are many notches up and a class by themselves.
In the nutshell,Precision is at a respectfully higher level than normal accuracy.

Your comments

Interesting, you would take the process in the upper right that uses the entire "specification" versus a process in the lower left that with a lower cost could be centered?

Both precision and accuracy come at a cost. It is cheaper to fix accuracy issues than it is to fix precision issues.

Perfect and befitting reply:applause::agree1::applause:

Both precision and accuracy come at a cost. It is cheaper to fix accuracy issues than it is to fix precision issues.

In many cases, the costs are negligible in view of the benefits. And I say again, the two values are relative, not absolute, and not separable in practical terms. The question was about process control, and if you have accuracy without some measure of precision, you have no control. It's easy to get into circular thinking about this if you're not careful, because while accuracy should be considered a universal requirement, the degree of precision is dependent upon what you're willing to put up with in terms of variation. But no matter where you set the limits on variation, you're still talking about the required degree of precision, not whether precision is important or not.

Ahhh, now i understand your(plural) point. i think the difference in our point of view is on how we picture the problem.

pls look at the attached and compare it with the previous 1. would you still side with accuracy?
Yes

Thanks for the file; I’m a great believer in a visual format.

I have a problem with “2” I consider accuracy must always cluster within the “tolerance” range to be correct.

An Example:

If I was required to manufacture an assembly that included a “bar” that was available as off the shelf “bar stock” meeting the “accuracy, albeit a wide tolerance band from nominal, then I would not set up a lathe / grinder to achieve the stated nominal size, to me it would be an unacceptable cost to achieve “precision”.

Do other consider this acceptable in terms of “accuracy” in the world of manufacturing?

Thanks

Yes

Thanks for the file; I’m a great believer in a visual format.

I have a problem with “2” I consider accuracy must always cluster within the “tolerance” range to be correct.

An Example:

If I was required to manufacture an assembly that included a “bar” that was available as off the shelf “bar stock” meeting the “accuracy, albeit a wide tolerance band from nominal, then I would not set up a lathe / grinder to achieve the stated nominal size, to me it would be an unacceptable cost to achieve “precision”.

Do other consider this acceptable in terms of “accuracy” in the world of manufacturing?

Thanks

It depends entirely on your customer's requirement.If you are given a sample to be duplicated,then you need high degree of accuracy to manufacture
'precisely' the same component.You can not compromise on dimensional tolerances without approval.
Second scenario,you are given a drawing,then you have to use high degree of
'precision' for critical dimensions(indicated therein) and +/-0.2 where
tolerance is not mentioned.
You do not have the authority to change the 'Tolerance' at your whim;unless
of course if you are manufacturing for your own consumption.:cool:

Thanks Umang, as long as you achieve "accuracy" within the stated tolerance band, as stated in my comments, then it acceptable to your customer requirements, as the tolerances become smaller and smaller there will be a point when both "accuracy" and "precision" converge

Thanks Umang, as long as you achieve "accuracy" within the stated tolerance band, as stated in my comments, then it acceptable to your customer requirements, as the tolerances become smaller and smaller there will be a point when both "accuracy" and "precision" converge

Now you got the bull by horns!:tg:

The dictionary meaning for 'Precision' is EXACT,meaning no variation,and
'Accuracy' means CORRECT,permitting variation.Therefore my interpretation
of your enlightening story is,that the five rounds fall in the catagory of
ACCURACy,and the single killing shot was fired with PRECISION;and hence
from my perception the moral of story is "Accuracy looks pretty-Precision
is deadly"

Comments invited from one and all:bigwave:

Umang

To add fuel to the definition debate, if you look up ISO's definition of the term 'accuracy' you will notice that when applied to a set of results, it encompasses BOTH trueness (lack of bias) and precision concepts. What you call here 'accuracy' is called 'trueness' in ISO's terminology.

Here I cite some definitions taken directly from ISO 3534-1:1993 (Statistics - Vocabulary and symbols - Part 1: Probability and general statistical terms)

3.11 accuracy: The closeness of agreement between a test result and the accepted reference value.
NOTE - The term accuracy, when applied to a set of test results, involves a combination of random components and a common systematic error or bias component.

3.12 trueness: The closeness of agreement between the average value obtained from a large series of test results and an accepted reference value.
NOTES
1 The measure of trueness is usually expressed in terms of bias.
2 Trueness has been referred to as “accuracy of the mean”. This usage is not recommended.

3.14 precision: The closeness of agreement between independent test results obtained under stipulated conditions.
NOTES
1 Precision depends only on the distribution of random errors and does not relate to the true value or the specified value.
(the other 2 notes are omitted here)

There is an ongoing debate around this issue and it has been addressed recently by other bodies such as the Royal Society of Chemistry:

Terminology - the key to understanding analytical science: Part 1: Accuracy, precision and uncertainty (http://www.rsc.org/images/brief13_tcm18-25955.pdf)

among other people:

Understanding the meaning of accuracy, trueness and precision (http://www.springerlink.com/content/k867505421705034/?p=9c6c5a58915b4250b7566b0746139fd3&pi=0)

Just my :2cents:

(BTW, I don't know if it's OK to partially cite copyrighted material such as an ISO standard, feel free to edit the post if necessary)

Good if you get both accuracy & precision.
There is no second thought that the main object of any measurement is to be accurate.

Good if you get both accuracy & precision.
There is no second thought that the main object of any measurement is to be accurate.

Would you think the same thing if the measurement was your cholesterol? Let's say you went to the doctor every month to have your cholesterol checked. Here is the data you got over the 1 year period of time. The true cholesterol reading (if it was known and stable) would be 195. So on average the test is accurate, but month to month you would be changing your diet.

190
207
195
177
232
202
187
197
177
192
197
187

Mean = 195
CV = 7.5%

Now, what if the true reading was 195 but you got the following results

217
211
220
208
216
219
217
210
207
218
222
215

Mean = 215
CV = 2.3%

In this case, you have a bias of 20 units, but you are not changing your diet month to month.

I would take the 2nd set of results....that is the question...I take a precise method over an accurate one...

Just my 2 cents:

Isn't it all relative (or what came first, the chicken or the egg ?):

If I am the consumer of the ouput from a measurement system, I might place more value on accuracy.

If I am responsible for the quality and maintenance of the measurement system over time, I may place more emphasis on the precision.

If I am building a new measurement system, I would like to have precision built in, then I could bias for accuracy later.

MJC

I agree with michael

what is accuracy without precision, its almost the same as seperating Cpk from Cp. I attatched a gile to explain to what I mean


Hope this makes you understand

FYI, my contribution. ;)

From the "Handbook of Dimensional Measurement" (Francis T. Farago and Mark A. Curtis)

Accuracy:

The concept of accuracy is of common knowledge: it designates the degree of agreement of the measured size with its true magnitude as expressed in standard units of measure.


Precision:

The term precision is used quite often, and is sometimes incorrectly substituted for the word accuracy. Precision expresses the degree of repeatability of the measuring process. Precision designates how closely identical values are obtained when repeating the same measurement at various intervals, or duplicating them by means of different instruments.


Enjoy!

Excuse me if these hav been said before but precison when applied to a process should mean that you are on target, nominal ?, and accurate would thus also mean the same.

As in on target and grouped closely.

More importantly do you want accuracy or repeatability?

Process capability will show if you are both accurate / precise Vs repeatable. If the process is both accurate and repeatable then your Cp and Cpk are both close to each otehr or the same. If you have a repeatable but not accurate process the Cp will be good but not Cpk

Think of this all as a target.

Excuse me if these hav been said before but precison when applied to a process should mean that you are on target, nominal ?, and accurate would thus also mean the same.

As in on target and grouped closely.

More importantly do you want accuracy or repeatability?

Process capability will show if you are both accurate / precise Vs repeatable. If the process is both accurate and repeatable then your Cp and Cpk are both close to each otehr or the same. If you have a repeatable but not accurate process the Cp will be good but not Cpk

Think of this all as a target.
I am not sure I agree with you.

Precision provides the ability to accurately determine the process variation, while accuracy provides the capability to accurately determine the mean with respect to the target.

Best regards,

Jim Shelor
PMP, CSSBB

From the discussion so far, here is my take on the matter

A process as a whole can be centered (or not) and repeatable (or not).
A measurement of a part can be accurate (or not) and precise (or not).


Suppose you have a dimension that should be 100 (and you also have a perfect instrument for measuring that you borrowed from NIST). You make a bunch of parts and measure them. IF you get
x-bar = 100; sigma = 0.2: you are centered and repeatable.
x-bar = 105; sigma = 0.2: you are not centered, but repeatable.
x-bar = 100; sigma = 10: you are centered but not repeatable.
x-bar = 105; sigma = 10: you are not centered and not repeatable.

OK, now you take a specific part and measure it with your ideal instrument from NIST and you get 100.40 . Then you take your normal instrument used to measure the parts during production. Depending on the instrument, your results might be:
accurate and precise: You repeated measure 100.40 for Part 1
not accurate but precise: You repeatedly measure 100.47 for Part 1
accurate but not precise: you get different values when you remeasure the part - 100.42; 100.37; 100.40; 100.44; 100.37. The results are precise only to 100.4. I.e. you should round to the nearest 0.1 even though the instrument gives one more digit on the display.
not accurate and not precise: you get a series a readings 100.2 +/- 0.1Note that precision can be determined with the instrument itself. Accuracy can only be determined by comparing with a trusted instrument.

Tim F

Excuse me if these hav been said before but precison when applied to a process should mean that you are on target, nominal ?, and accurate would thus also mean the same.

As in on target and grouped closely.

More importantly do you want accuracy or repeatability?

Process capability will show if you are both accurate / precise Vs repeatable. If the process is both accurate and repeatable then your Cp and Cpk are both close to each otehr or the same. If you have a repeatable but not accurate process the Cp will be good but not Cpk

Think of this all as a target.

Accurate = On target
Precise = Grouped closely

Example:
A shootist hitting a target at 100yds and maintaining a 8 inch pattern is accurate, however not precise. A second shootist hitting a target at 100yds and maintaining a 2 inch pattern is not only accurate but also precise.

Accurate = On target
Precise = Grouped closely

Example:
A shootist hitting a target at 100yds and maintaining a 8 inch pattern is accurate, however not precise. A second shootist hitting a target at 100yds and maintaining a 2 inch pattern is not only accurate but also precise.

Accuracy has a wider tolerance band,while precision is in smaller tolerance band.Anything falling in 'PRECISION'category has got to be accurate but vice-versa is not true;id est,every thing 'ACCURATE'does not fall in the category of precision.
Let us take example of measuring instruments:
1.The measuring scale has 12" or 30cms on each edge.The least count is
1/8"and 1mm respectively,which means that the scale can reliably measure up to the 'accuracy' of 1/8"or 1mm.Measurement taken by this scale is 'accurate' despite low degree of accuracy(wide tolerance band).
Now consider a 'Precision instrument' viz a Vernier caliper or a Micrometer
having a least count of 0.02mm(smaller tolerance band).Measurement taken by 'precision instrument' is also accurate,with higher degree of accuracy,and is called 'Precision'.
Next,take a Contactless Micrometer,having a least count of 1 milli inch (smallest tolerance band).Measurement taken by this is also accurate,with extremely high degree of accuracy.This falls in the category of 'High Precision Instruments'.
2.Similarly consider machine tools.The ones used in production have low degree of accuracy,while similar machines for tool room have high degree of accuracy and are called 'Precision Machines'.Then there are Copying Lathe,
Copying Milling etcetra,having much higher degree of accuracy.These are called 'High Precision Machines'.
Thus we arrive at the logical conclusion that any reading,however afar from the 'True Value',but falling within the requisite parameter,is considered
'ACCURATE'.In contrast 'PRECISION' is close to the 'True Value'.Closer the reading higher the precision.That is why 'Precision Bombing' means,every bomb
must hit the target-not a single miss allowed!A sniper/sharp shooter has to be 'Precise' while killing one in the crowd!!A surgeon has to operate with
'Precision' because a single mistake makes the difference between Life and Death!!!
So you see,Accurate=Varying degree of accuracy around the target.
Precise=Grouped closely around the target(Bulls eye):cool:

I don't say what I mean,but I mean what I say