enter image description here

It's supposed to be zero but for some reason the result produces a peculiar series of digits at the end.

Also, here's the .xlsx of it:

  • I found a workaround with round(), and I heard it's a problem with floating arithmetic. I do not know if it can be fixed.
    – j riv
    Apr 6, 2013 at 11:00

2 Answers 2


This happens because of the limited precision of floating point numbers representation on computers and cannot be fixed because it is an intrinsic problem of the way number are represented on a computer. There might be workarounds, such as round() or a clever implementation of the formulas to reduce error propagations.

Excel uses 8-byte (64 bit) floating point representation, meaning that it can achieve a maximum accuracy of about 10^-15. The inaccuracy can then spread across the steps of a calculation.

I suggest that you read this article on Wikipedia that investigates this topic.

  • They were using floating point on the windows calculator too to aquire new and different math :-) Simple rounding could have solutioned it. Instead I get to read 20 pages of stuff about it to find out that FP has a specific accuracy as used, and within that accuracy used, it should just be rounded back to within that accuracy. Rounded proper not chopped.
    – Psycogeek
    Apr 6, 2013 at 12:40
  • Advanced users may decide if they need a rounded or truncated result based on their needs. Unnecessary rounding slows down the calculations. Apr 8, 2013 at 14:04

Self post: It's a documented limitation for floating point arithmetic: http://support.microsoft.com/kb/78113

IMO it should be warning more visibly when the results can be so blatantly erroneous to not be considered a bug, but anyway.

A workaround is to use round().

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.