Robert Dewey
EF5
It doesn't quite work that way imo at least. The period of study needs to be as long as the frequency of events being considered to get reliable results. Since the super outbreak is at best a 120 year anomaly, you shouldn't consider it in a study that only spans 75 years. Also, the use of the terminology "tornado alley" implies a region of favored frequency of occurence of tornadoes, doesn't it? Really, even the signal across the deep south is low frequency - but more on a 5-10 year period than 150 years.
Glen
I would agree, it's tornado days per century, yet the data doesn't span that far - So, does that mean they are extrapolating data for the remaining years, or simply not using that data? Either way, it skews the study. I also agree with Mike G., that most reports prior to 1950 may be rather inaccurate - I would tend to move that date up to 1970-ish, or up until the Fujita scale was widely used - Since this study uses the Fujita scale (F2 or greater). In my opinion, the main 'alley' is located from northern TX into NE and western IA... But that doesn't mean there aren't other hotspots, as some have noted.