In 2001 there was a mistake with the scaling for 4U, which lead to abnormally high results. One of these values is based off the mistake, one is based off the correction. I have alas forgotten which is which, but try both and see which one gives a lower result.