That would mean 10-20% do develop lung cancer which sounds high and really bad.

Why the downvote? I don't think this was asked in bad faith.

While out of scope for the question, it should be mentioned with any tobacco use conversation that smoking has a whole litany of other serious health concerns associated with it as well, including but not limited to heart disease, diabetes, stroke, COPD, and emphysema..

@DanK: And even sticking to cancer, there's a whole host of non-lung cancers that smoking raises your risk factors for dramatically (larynx, oral, esophagus, bladder, etc.). If you "successfully" die of any of these problems (cancer or otherwise) before you get lung cancer, you'd be counted as not developing lung cancer, even if lung cancer would have occurred in the extra ten years (on average) that non-smokers get beyond smokers.

Most drunk-driving trips also do not cause accidents. But this is a very purposely misleading way of showing statistics.

@JeromeViveiros Downvotes happen. Some people don't understand this site and downvote questions that they find irritable. Those irritable questions based on dubious claims made in notable source are the kinds of questions this site aims to address. I've learned that upvoting the most irritable questions is oftentimes the best thing to do at this site.

@JeromeViveiros Nitpick: do you mean irritating questions? A question can't be irritable. (You shouldn't anthropomorphise questions — they hate that…) Or do you mean something else?

Here are the papers the Nature Genetics paper cites to establish this rate: scholar.google.com/… scholar.google.com/…

83% of Russian roulette players survive each round.

@kutschkem On the other hand, I'm sure there are some people out there who say, "80 to 90% of life long smokers never develop lung cancer? I should have taken up smoking sooner!" (ugh!)

I'm a fan of @AnderBiguri's comment. The can believe the article's claim - but its use at the beginning of the article has more to do with search engines and clickbait than it does science and health. If you think about it, you have a magnificently high probability of living your entire life without being struck by lightning ... but that's a really bad reason to walk around holding a lightning rod.

@Ander, on the other hand, arguably, it's less misleading for each one's assessment. (Let's isolate just the lung cancer case for simplicity). If as a smoker you feel "I smoke, so I will die from cancer!" - well, that's wrong, in all likelihood (80-90%) you won't. It's the constructs like "factor X doubles your risk of <blah>!" (much liked by the media) that are often misleading. While technically true, doubling your risk from (say) 1E-6 to 2E-6 is insignificant for you personally on the background of other risks; even 100x may be insignificant. You can only observe it on a big population.

A better question is perhaps how many percent of lung-cancer victims were or had been smokers - perhaps followed by other causes; like asbestos, living in a city, ...

@ShadowRanger I think your comment should be an answer. Especially the fact that smokers may die prematurely before even getting that cancer is quite relevant.

@BaardKopperud You can google that, and IIRC it turns out to be about 90% of lung cancer patients are smokers, or something like that.

@SebastiaanvandenBroek: bta's answer seems to have it covered well enough.

@Zeus not that different though. 20% of chance of getting lung cancer is A LOT. As humans we don't do ANYTHING that is anywhere close to have us 20% likely dead (lung cancer will kill you, recovery rate is negligible). As others have said, it's like if billions of people would play a round of Russian roulette. That's the fair comparison. Also its not double the risk, before smoking and industrialization, lung cancer was rare.

"tobacco products being the cause of 90 percent of deaths" -- what's the denominator? It's not 90 percent of all deaths. Maybe 90 percent of deaths of smokers?

The time unit "lifelong" may have a lower value for smokers to begin with ...

@Ander, my point is not the actual numbers. Whether something is "a lot" or not, everyone can decide for themselves. I'm saying that presenting the data as "risk increase" (which is sort of a counter-argument to the original question: "yes, but...") is probably even more misleading when taken in isolation than the survival rate. In other words, yes, 83% survival rate of a round of Russian roulette is subjectively a fairer indication of affairs (for the person taking the chance) than the 17% death rate (which is an infinite jump in risk, by the way!)

@Ander, by the way, if you read LangLаngС's answer, recovery rate of lung cancer is far from negligible (around 50% or better). But generally for advanced cancer the indicative metrics is AFAIK different: survival 5, 10, etc. years after the diagnosis.

@AnderBiguri Talking about the rate of lung cancer before industrialization is a bit moot because before industrialization, life expectancy was 40 years old, and age happens to be a pretty big risk factor for cancer. Not that I'm saying it's not a factor...

@HagenvonEitzen Makes a great point: what does lifelong mean? I mean of course it means their entire (adult?) life, but that can vary a LOT. And the ones that don't get cancer will live longer, and thus that pool will appear to be very healthy because they survived possibly by chance, not necessarily because they have good genes, etc. They might, but statistically P(people who didn't die from risky activity having been harmed by risky activity) is necessarily going to be smaller than P(dying from risky activity) because it's a self-selecting group. So that may be a factor here too.

In a similar vein to how sometimes people who live to 100+ when interviewed say they smoked, ate fatty foods, etc. and how it would be a mistake to assume that's why they lived to 100 and conclude that their habits were healthy after all.

What really makes the statistic disingenuous is that the 10% risk is an unnecessary risk. There is simply no good reason to smoke cigarettes, and that, coupled with it's increased risk of cancers (of all types), is enough to establish that smoking cigarettes is inexcusable folly. Contrast this with the risk of drowning when swimming for recreation, or crashing on a bike ride, and risks associated with other activities; there are benefits to these activities which render the risks acceptable.