Looking math info about tssmooth exponential

[Eva Smith]

I found the formulas used to calculate "Simple Exponential Smoothing". Manual calculation of formulas via excel showed different results from STATA. Where can I read more info about how STATA do math with "tssmooth exponential"? If there are any examples, it will help me a lot

[Nick Cox]

Did you look at the manual entry? Here is a direct link: https://www.stata.com/manuals/tstssmoothexponential.pdf

See also https://www.statalist.org/forums/help#spelling

Perhaps what you need is more documentation about what Excel does, or to show us what you mean.

[Eva Smith]

For example, i take this article

Forecast values calculation is correct. RMSE = 2140.74 for a=0.29. Then i try to do exponential smoothing.

In the stata i got:

. tsset var1
        time variable:  var1, 1 to 9
                delta:  1 unit

. tssmooth exponential new1 = var2, parms(0.29)

exponential coefficient  =       0.2900
sum-of-squared residuals =     34653715
root mean squared error  =       1962.2

var1 var2 new1
1 17496 15944.25
2 18965 16394.26
3 12535 17139.77
4 14781 15804.39
5 16722 15507.61
6 15858 15859.78
7 16053 15859.26
8 14565 15915.45
9 15747 15523.82

I don`t understand how "new1" was calculated.

[Nick Cox]

You're showing us what I presume is an Excel screenshot, but you need to tell people here who are competent in Excel (not me; sorry I won't be answering anything about Excel, and am just trying to help you pose a question that can be answered) how column G was calculated, namely using which Excel function or other code.

Otherwise #2 is already the answer to what Stata is doing as the manual entry gives the formulas.

Why 0.29?