Right - and so the more appropriate thing to do is not look at the raw likelihood of any one particular value but instead look at relative likelihoods to understand what values are more likely than other values.