My personal belief is that MOST of the time, open source will still win out - but that isn't some sort of fundamental rule of the universe. Any discussion of TCO that doesn't bring up support is missing the mark.

In this particular example I agree 7zip is a better choice, but many other open source projects either have a paid support plan or lackluster support (meaning you'll be supporting it yourself). Both of those scenarios need to be calculated into the TCO.