[Blood of Tyrants]

Yes, it sounds farfetched. They probably get theeir numbers by equating every death of every person who ever smoked as a "smoking related death" whether they smoked for one month and quit 60 years ago or died from an accident, or died from an illness where the link between smoking and that illness is specious at best.

[GraceCoolidge]

I believe I've read something similar to what you say. In some of these materials, a death is "smoking related" if the deceased was a smoker... even if he or she dies in an automobile accident (maybe because they were driving to the 7-11 for cigarettes???). Reminds me of that "statistic" about spousal abuse on Super Bowl Sunday.