The standard error doesn't appear to be factoring into the calculation. Only variance seems to be driving the circle radius.