Absolute vs relative riskRelative risk reduction · Absolute risk reduction
"Halves your risk" is meaningless until you know what the risk was to start with.
Relative risk expresses a change as a proportion of the starting risk. Absolute risk expresses it as the difference itself. Both describe the same finding and they read completely differently, which is why the choice between them is the most common way statistics mislead honestly.
The worked example is short. A fall from two cases per thousand to one case per thousand is a 50% relative reduction and a one in a thousand absolute reduction. Both statements are true. Only one of them tells you what it is worth to you.
Headlines almost always use the relative form, because it is the larger number, and because it does not require knowing the baseline. That works both ways: doubling a very small risk is still a very small risk, and a modest relative reduction in a common outcome can matter enormously.
The number needed to treat converts absolute risk into something more intuitive: how many people have to take a treatment for one of them to benefit. It follows directly from the absolute difference and it is rarely quoted in press coverage.
The same asymmetry applies to harms. A drug side effect reported as a 60% increase sounds alarming until you learn the baseline rate was three in a hundred thousand.
This is one of the most practically useful ideas in reading health news, which is why it has its own entry here and why our evidence pages try to give absolute figures where the source provides them.
In practice, whenever you see a percentage change, ask: percentage of what, and how common was it to begin with?