software

If I divide two negative numeric cells and put the result into a
percentage cell, it will positive, even if the change is negative. For
example:

A B
1 -6249 -5810

A1/B1 = 107.56%

The change is moving from B to A. That means I should have a negative
change. How can I show that?

Thanks,
BrettHarlan Grove is pretty smart.

Biff

quot;brettquot; gt; wrote in message ups.com...
gt; If I divide two negative numeric cells and put the result into a
gt; percentage cell, it will positive, even if the change is negative. For
gt; example:
gt;
gt; A B
gt; 1 -6249 -5810
gt;
gt; A1/B1 = 107.56%
gt;
gt; The change is moving from B to A. That means I should have a negative
gt; change. How can I show that?
gt;
gt; Thanks,
gt; Brett
gt;
On 8 Jan 2006 18:19:12 -0800, quot;brettquot; gt; wrote:

gt;If I divide two negative numeric cells and put the result into a
gt;percentage cell, it will positive, even if the change is negative. For
gt;example:
gt;
gt; A B
gt;1 -6249 -5810
gt;
gt;A1/B1 = 107.56%
gt;
gt;The change is moving from B to A. That means I should have a negative
gt;change. How can I show that?
gt;
gt;Thanks,
gt;Brett

You could always multiply your answer by -1. However, a quot;negativequot; percentage
is not generally a meaningful concept.

What real world problem are you trying to solve?--ron

quot;brettquot; wrote:
gt; If I divide two negative numeric cells and put the
gt; result into a percentage cell, it will positive, even
gt; if the change is negative. For example:
gt; A B
gt; 1 -6249 -5810
gt; A1/B1 = 107.56%
gt; The change is moving from B to A. That means I
gt; should have a negative change.

What makes you think so? -5810*107.56% is -6249,
the correct answer. If you used -107.56% arbitrarily,
you would get 6249, which is wrong.

gt; How can I show that?

Since your choice of negative quot;changequot; (factor)
seems arbitrary, I don't know what you would
want in all cases -- for example, B is 2 and A is
-4, and B is -2 and A is 4. Note that when B is
4 and A is 2, the quot;changequot; (factor) should not
be negative.

Your request makes more sense to me when we
are talking about actual change, not growth factor.
Whenever A is less than B, we might reasonably
want to express the percentage change as negative.
But we must use that notion of change carefully.
For example:

To compute percentage change (C1):

=IF(B1=0, A1, SIGN(A1-B1)*ABS((A1-B1)/B1))

To apply (use) percentage change to B1 (D1),
which should equal A1:

=IF(B1=0, C1, B1 SIGN(C1)*ABS(C1*B1))

Examples (B1=before, A1=after):

A1B1C1D1 (should = A1)
24-50%2
42100%4
-2-450%-2
-4-2-100%-4
2-4150%2
-42-300%-4

Some people will quibble with my choice when B1=0.
It is arbitrary.