>an extensive study "Mathematics and Smoking" that evaluated 1000 randomly chosen mathematical smokers and 1000 non-smokers, and concluded, that the correlation between smoking and mathematical ability is 0.765 (plus-and-minus 0.003). This interesting study was rejected by the Notices of the American Mathematical Society, not because they found anything wrong with it, but because of the PC mafia that was worried that this study would encourage young graduate students to take up smoking in order to improve their math. Please give me a break! Graduate students are smart enough to know that correlation is not causation, and just if they did not smoke until now, adopting this bad habit would not increase, by epsilon, their chance of proving RH.
https://sites.math.rutgers.edu/~zeilberg/Opinion155.html

Zeilberger thinks it would be stupid to conclude from this correlation that there's causation, but I disagree. Think about it. Mathematical ability should be correlated with intelligence, and more intelligent people are typically more aware of the dangers of smoking. Being smarter should mean you're LESS likely to smoke, so the fact that the smokers turn out to be better mathematicians should mean that increases in mathematical ability provided by smoking are even GREATER than the correlation itself suggests.
If you want to be a good mathematician, why don't you smoke?