>Currently the fillfactor of the fragmented table is 0.
>What I'm looking for is a standard formula to calculate the most reasonable fillfactor that I should use.

The fillfactor should represent the volatility of the data and the impact of the implied growth that you're willing to deal with. If you have a table whose size is 100 pages and you recreate the clustered index with a fillfactor of 50, you've created enough space to allow twice the number of rows before node spliting begins but you've also doubled the number of pages (200) and the amount of work SQL Server must do.

I guess the short answer is that I don't know of any formula that will provide you with the optimal fillfactor.

-Mike