Created attachment 94945 [details] Sample showing Excel vs LO results When using the exponential smoothing option under the new statistics menu it exhibits the following problems; 1. The first result in the series should be #N/A as it has no previous data. 2. The cell references are incorrect on several counts 3. The results are the same as the input Also, would it be possible to supress the column headers with a option in the dialog box? When compared to Excel, putting the alpha value into a separate cell is a good idea.
Hi, 1. The first "result" is the initial forecast, which can be calculated in many ways and it is usually the first value. See [1] and related links. 2 and 3. Your data is row oriented so you have to set "Group By" to row. The output is always column oriented tough - maybe it should depend on the "Group By" selection. Also note that in Excel's "Damping factor" = (1 - Alpha). When you set "Damping factor" to 0.7 this means Alpha is 0.3; Who knows why they did it this way as it is clearly misleading. Suppressing of column headers is a to-do. [1]: http://en.wikipedia.org/wiki/Exponential_smoothing
Tomaz: reading your comment about this one, should we put "NOTABUG"?
Yes, I have set it to NOTABUG.