While I certainly agree with the outcome of Jan's assignment of oxidation numbers, I find it very necessary to point out that this is not the only possible assignment, not even the most common one.
I have discussed that in an answer to Electronegativity Considerations in Assigning Oxidation States. The official rules of the IUPAC indeed assign +1 to fluorine. This is counter-intuitive, as you stated, but that is the way the strict application of the rules give it to us. For me one of the most disturbing flaws of these rules. What you called an anomaly is indeed due to an incomplete set of rules.
I have laid out the difference of the assignment in my other answer to the topic in Oxidation state of oxygen in HOF.
It is absolutely necessary to know that there are two opposing systems out there and one should be aware of the flaws of each of them.
Furthermore, there is not really a correct way of assigning oxidation numbers, not even when you start considering fractional oxidation numbers. Hence there is also no way of correctly predicting oxidation numbers. They were introduced as a tool to better understand and keep track of the changes in the charge density. They usually do not come close to measured or calculated partial charges and they often oversimplify bonding situations of molecules.
Does that mean we should abandon the concept? No. It is a very helpful tool, especially when considering redox reactions. It gives you a formal framework to track a reaction and to make simple predictions.