[QSIITurbo]

Yeah, it has its weaknesses and also from my comment below: "So I read the study for fun and the non-adjusted rates are 3.84% death rate (cohort study) in no exercise vs. 2.35% in 1-30 minutes and 2.08% in more. And that's not adjusted for pre-existing conditions if I understood correctly. So meh. Even though 20% and 20% sounds really great!"

I didn't find anything there for longer periods of time but odds ratio is not a great indicator if your base chance of dying is already really low. Or maybe spend the whole week exercising to squeeze that last 0.1% out of your chance of dying.

[ramblerman]

3.84% to 2.08% is a 45% reduction in the amount of people that died. I don't understand your point, there is no obfuscation here, that's just how percentages work.

[QSIITurbo]

Um, no. 3.84% to 2.08% being an approx 40% reduction is exactly what is being avoided currently in research since its reduction of less than 2 percentage units. Imagine a medication that has a "20% and 40% reduction in mortality": You'd have to have more than 150 people treated to the max to save one person vs other cases (more than 75 to save vs no medication). Not 5 as your percentage implies. Not very efficient way to spend money if you ask me.

[ramblerman]

(3.84 - 2.08) / 3.84 = 45% less people dead

It's not 'my' percentage, nor subjective.

I mean you could argue the study is biased or the sample size too small, but your understanding of percentages seems flawed.

[QSIITurbo]

I recommend you take basic university statistics. Especially study efficacy vs effectiveness so you might have an idea how you're being mislead here.