There has been renewed interest in whether we should understand standards of decision-making capacity (DMC) to be risk-relative. Critics of risk-relative standards often highlight a puzzling asymmetry that they imply; a patient may have the requisite DMC to consent to a treatment that is in their best interests, whilst lacking the requisite DMC to refuse that same treatment, given the much higher risk that this would entail. Whilst some have argued that this asymmetry suggests that risk-relative standards are nonsensical, in this paper I defend a ‘quality of evidence’ view of such standards. I begin by outlining DMC’s purported gate-keeping role in medical ethics, and identifying three key normative claims that undergird this role. I then explain how two competing theories of risk-relative standards are incompatible with at least one of these claims. Drawing on Douglas’ distinction between standards of ‘true capacity’ and standards invoked in the ‘test’ for capacity, I then outline my ‘quality of evidence’ view. I explain how the view is compatible with the aforementioned normative claims, and outline the nature of the asymmetry it implies. I conclude by responding to the objection that there is no meaningful distinction between ‘true capacity’ and the ‘test’ for capacity.