# Thread: F-test : What happens if your dgf is too big ?

1. ## F-test : What happens if your dgf is too big ?

Hi guys ,

I have a question about F-tables.
What value should I put for the F 20000,17000 , would it be correct to use the same value for the last value in my F- table which is 1.32 for F infinty, 60 ?

2. I use online calculators...
The one I usually use didn't work well with 1000 and 2000 dfs
BUT I just found this
http://davidmlane.com/hyperstat/F_table.html
I just put used 100,000 and 200,000 dfs.
AND I checked with the percentile point of 1.
I know that as the df goes to infinity that the mean converges to 1.
And it 'worked'.
There's also these tables....
http://www.danielsoper.com/statcalc/calc08.aspx
Note that there are F tables if you look on top.

3. so is 1 the tail probability and also is this only true for f table or all kind of tables ?

oh btw thx for the links eagle

4. $\displaystyle E(F_{a,b})\to 1$ as b goes to infinity
The mean of a chi-square is it's dfs, which is why the mean of an F is near 1.