I mean, depending on your calculations and scale, you might go a little more precise with it. At a diameter of, say, 10m for a semicircular bridge arc, that’s a difference of 0.7m.
(For mathematicians, the difference will be 0.00796m and then some I can’t be arsed to write out, but compared to the total arc of 15.7m, that’d be a deviation of 0.05% which is basically zero anyway)
Wouldn’t that be within tolerance of 3‰?