That makes sense. I did not have the data for total miles driven by only trains with hazmat. I used the assumption that the proportion of trains carrying hazmat stayed constant from year to year (maybe the not the best idea looking back). The trend for total derailments is very similar to the adjusted.