Report Numbers and Rounding I have a report that I have generated on some data. The requirements of this report in particular is that the values for every row of data should be rounded to the nearest dollar. My boss is checking this report against another similar report to see if the numbers match up, but the requirements of that report was that it round to the nearest penny. Everyone understands that the numbers will be different to a degree. Is there a way to mathematically calculate the range that we can expect the numbers to differ (+/-). I dont want to leave it up to just what "feels about right" numbers can

As Ilmari Karonen says, if you round to the nearest penny, and then to the nearest dollar, the result should be the same as if you rounded directly to the nearest dollar.

If on the other hand you are only checking the sums of each rows, then rounding differences may become apparent, and the more terms there are in each row, the more likely they are to occur. I once wrote a note _May not sum to total due to rounding: the probability of rounding errors_