Suppose, you have


Then y is 25% more than x, if x is used as the "base" value. If y ist the base, then x is 20% less than y.

So if I want to say, that the difference between x and y is ... percent, what value could I use? I can think of using 22.5% being in the middle between 20% and 25%, but that does not feel well-founded.

Should I not use percentages at all?

  • $\begingroup$ "the difference between x and y is ... percent" - of what? $\endgroup$ Jul 16, 2011 at 13:13
  • $\begingroup$ @J.M., so you are asking for a "base" (do you call it that?) as well. I thought, I could do without. But @Sylverdrag 's answer suggests otherwise. $\endgroup$ Jul 16, 2011 at 13:37
  • $\begingroup$ If you want to avoid a base, then you could take the logarithm of the quotient (i.e. the difference of the logarithms), since $\log_e \frac{10}{8} \approx 0.22314\ldots$ and $\log_e \frac{8}{10} \approx -0.22314\ldots$, but this is not a percentage. $\endgroup$
    – Henry
    Jul 16, 2011 at 22:58
  • $\begingroup$ This is an interesting idea, @Henry! Thank you. $\endgroup$ Jul 19, 2011 at 0:43

1 Answer 1


A percentage can not stand alone, it's always a percentage of something.

You have to specify which number you are using as a base. No way around it...unless of course you work in marketing.

  • $\begingroup$ I think, I understand. So I could probably do the following: Take the mean between x and y and then calculate the ratio between y-x and the mean? What I mean is $\left| \frac{y-x}{(x+y)/2} \right|$ . $\endgroup$ Jul 16, 2011 at 13:40
  • $\begingroup$ @Sebastian: You could do that. But then you would be using a convention that as far as I know no one else uses. $\endgroup$ Jul 16, 2011 at 13:48
  • $\begingroup$ @André, ok, I hoped there was even a name for that formula. I should probably think twice, before using percentages if I want to described positive and negative "movements" (as in -20%, but +25%) at the same time. $\endgroup$ Jul 16, 2011 at 14:13
  • $\begingroup$ @Sebastian: We can speak of percentage deviation from the mean. However, the context there is not a mere two observations. Using it on say two prices would be misleading. $\endgroup$ Jul 16, 2011 at 14:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.