Dear Cecil,
in the article you mention the increases in cancer risk these behaviors cause. However one needs to know what the occurrence of the cancer in the general population is before one can conclude which behavior is more lethal.
Example (with bogus numbers) if you double your risk of lung cancer and the prevalence of lung cancer in the population is 2 out of 1000, you have a chance of 4 in 1000 of getting lung cancer (an increase of 2 in 1000). If the prevalence of melanoma is 10 out of 1000 and you have a 50% higher chance of getting melanoma you have 15 out of 1000 chance of getting it (an increase of 5 in 1000). So you’d have a higher chance of melanoma than lung cancer.