The problem is with your data. You have 635 input points, but only 5 distinct X values. This means that for large amounts of your data, there is no X variation. The moving regression cannot calculate a slope, because the data are aligned vertically and the sum of squares variation of X is zero. Dividing by zero then gives you an overflow (infinity). The formulation of the weighting factor exaggerates the problem.

If you use 100% of the data points, you can use the program, (50% still had an overflow). The LOESS result is slightly less interesting than simply looking at the averages.

Another thing: using the same X inputs and outputs means the program will calculate duplicate output values each time a duplicate X appears in the input range. Use a second table for the outputs, without the duplicate X values, to minimize calculation times.

I tried to use the loess utility and received a message similar to one of your other users: runtime error ’6′ overflow. The solution to this problem wasn’t on this listserv, can you please help me to figure out reasons for getting this error message.

Thanking you in advance,

Shirley

Thanks again for putting the time into this utility!

I use this utility frequently, at least once per day, mostly in 2003, and I’ve never encountered any instability. Do you get an error message.? Does it happen on sheets on which you’ve already used the utility, or on sheets on which you haven’t used the utility, or both? Does it happen right after Excel opens, or after it’s been running for a while? Is the data completely numeric?

Let me just say thanks for putting together a great tool. It’s perfect for analyzing my data and bringing datasets to a common scale for comparison. However, after some use of the add-in it’s started to exhibit some instability. It launches fine and I can click in the “X Input” box, but if I click anywhere else on the interface it closes immediately. I’m currently on Excel 2010, but had a similar issue on 2003.

Have you ever seen that before or have any recommendations for a fix?

Thanks,
Jason