According to this[1] list, 100 would be vastly over estimating. Even including 9/11, and disregarding all incidents with no fatalities, the average based on the GTD data is about 12.5. If you are willing to concede that 9/11 was a once-in-ever event that is unlikely to be matched, then the number drops below 2.

And that's probably still a vast overstatement of the likely fatalities for any given plot, given that 2103 of the 2381 recorded by the GTD have 0 fatalities.

[1] http://www.start.umd.edu/gtd/search/Results.aspx?chart=fatal...