Enter the formula =52.8-8.8*6. The answer is, of course, zero. Now enter it
as =(52.8-8.8*6). Excel will return a non zero number. Also try
=70.4-8.8+61.6 with and without parentheses. If you happen to divide by any
of these you will get a very large number.

http://www.cpearson.com/excel/rounding.htm

--


Regards,

Peo Sjoblom


"CRatt" wrote in message
...
Enter the formula =52.8-8.8*6. The answer is, of course, zero. Now enter
it
as =(52.8-8.8*6). Excel will return a non zero number. Also try
=70.4-8.8+61.6 with and without parentheses. If you happen to divide by
any
of these you will get a very large number.

Most decimal fractions, including .8 and .6 have no exact binary
representation (much as 1/3 has no exact decimal representation). Excel
and almost all other software follows the IEEE standard for double
precision storage to approximate your input numbers. The nonzero values
are the result of exact calculation from the approximate inputs. The
exact results for =(52.8-8.8*6) and =(70.4-8.8-61.6) are
-/+ 1/140737488355328
which in decimal is exactly
-/+ 7.10542735760100185871124267578125E-15
Excel correctly reports this to its documented limit of 15 decimal digits as
+/- 7.105427357601E-15

Without the parentheses, you get zero because the final operation
calculates the difference between two numbers that agree to 15 decimal
digits. Under that circumstance, Excel arbitrarily zeroes the result on
the assumption that any possible difference from zero is the result of
binary approximations.

Jerry

CRatt wrote:

Enter the formula =52.8-8.8*6. The answer is, of course, zero. Now enter it
as =(52.8-8.8*6). Excel will return a non zero number. Also try
=70.4-8.8+61.6 with and without parentheses. If you happen to divide by any
of these you will get a very large number.