0

I wasn't that sure what kind of mean value the following is, I hope you can help me out:

What it's about:

I have this testseries of about 30 elements, that delay the boot sequence of an embedded System. I checked them separately to say, element #3 adds e.g. 0.4 s to the boot sequence (lets call it tadd).

Now when I want to know, in what time the system boots when all 30 elements take influcence, I can assume this by adding each t.add of every package to a tbase (the time the system boots without any of these elements). Let's say I get a time t.assume = 30 s. Now I'm measuring the system with all the elements enabled and I measure the time with 29.5 s.

so subtracting each other lets me know how good my assumptions were:

t.accuracy = t.assume - t.measured = 30 s - 29.5 s = 0.5 s

Now to the question:

What kind of a mean value is this, if I divide t.accuracy 0.5 s by the number of elements used ?

--> 0.5 s / 30.

Is it the mean failure value of each element, or does it even exist? And if so what's the name of it.

Thanks for your support!

gt6989b
  • 54,422
user3085931
  • 113

1 Answers1

1

This would be average error per element. In other words, on average, how big a mistake did I make in assumptions for each element of the sequence.

gt6989b
  • 54,422
  • thanks. Is "average error per element" the actual terminology for it ? – user3085931 Mar 26 '14 at 13:50
  • @user3085931 I am not familiar with your process's terms, but if I had to explain to someone in English what that quantity was, this is the term I would use. – gt6989b Mar 26 '14 at 13:52
  • I'm not so deep into these average methods that's why I got to ask again: If it would be an average error/element, wouldn't that mean that it's possible that e.g. 15 out of the 30 elements have the exact value and the others have then in turn a higher offset? Couldn't it be, that "my" value, is actually just saying each element element puts the result 0.5 s/30 away from the right value ? So that with 15 selected elements there's an offset of 0.5 s/15 and for 30 selected elments there's an offset of 0.5 s ? – user3085931 Mar 26 '14 at 14:07
  • @user3085931 That it is an average means that some actual values could (and likely will) be higher and some could (and likely will) be lower. Average must be in the middle, so it cannot be some will match exactly and others will be all higher or all lower. You should have a mix. – gt6989b Mar 26 '14 at 14:14