Yet it is the subjective perception of the risk rather than the actual statistical probability of risk itself that people base their choices on."Risk is subjective?" No, it's not. One's perception of the risk may be, but the actual risk is not subject to one's perception of it.
So in application, the 'actual risk' is less relevant as a discussion point, since that is not what people think about in their assessment of risks anyway.
Although I know the risk is low, it's still not zero, which is why I choose ATGATT anyway. Other people obviously think their risk is considerably lower to the point of being nearly impossible (improbable ?) and therefore chose to ride in t-shirts, shorts and sandals.