True / False problem

It's probably a precision issue....

try:

=round(b21,2)=round(C17-C15,2)

It is most likely some sort of precision problem... if B21, C17 and/or C15 are results of formulas, then it is more likely the case...

What are the numbers like if you format them as General? Is B21 exactly the same as C17-C15?

Perhaps you can attach your worksheet, pointing out the problem.

In your sample sheet...

B21=676.59

C17=39,825.69

C15=<blank>

so how is B21 supposed to equal C17-C15?

Perhaps you meant C17-B15?

so then formula would be =ROUND(B21,2)=ROUND(C17-B15,2) which results in TRUE.

Originally Posted by NBVC

In your sample sheet...

B21=676.59

C17=39,825.69

C15=<blank>

so how is B21 supposed to equal C17-C15?

Perhaps you meant C17-B15?

so then formula would be =ROUND(B21,2)=ROUND(C17-B15,2) which results in TRUE.

yes...it was B15...I had moved data and forgot. I still don't get why it would require ROUND. 2 of the 3 are entered data and the 3rd one is addition.

Thanks for the help!

For some strange reason Excel has issues with precision in some operations, especially when it is using the result of another formula within the affected formula...

If you try subtracting B15 from C17, you actually get result of 676.520000000004

Nothing strange, and it is NOT an Excel thing. Your computer (and therefore Excel) uses a floating point representation of numbers, which intrinsically has finite precision.

For example, you know that 1/3 = 0.333 ..., with those 3s going on until your fingers get tired. If they get tired after 6 digits, the closest number you can store to 1/3 is 0.333333. No matter how many digits you had available, you couldn't write the number exactly. The resulting error propagates to every downstream calculation you perform.

IEEE double precision gives you 15 digits, which is adequate for most purposes -- but even it it had more, lots more, you would still have the same problem.

Further, the floating-point format is not decimal, it's binary. (Actually, there is an IEEE decimal floating point format, but it seems to have gone nowhere). The number 1/10 = 0.1 is actually a repeating decimal in binary; in Hex, it's 0.19999999... . So you can't store the value of a dime exactly in Excel (or on your computer).

If you're dealing in dollars, you should generally round to two decimals after every calculation. When you make a loan payment, you pay some exact number of cents, and the principal is reduced after some exact number of cents in interest is subtracted from the payment. All the little remainders are gone, and you start fresh for the next month -- which is why the last loan payment may differ by a penny or two from all the other payments.

It's just one of those things you need to be aware of and compensate for when necessary.