but then you run into the old saw about damn lies and statistics. the argument needs to stand on its own and be taken in good faith, otherwise discussion is futile.
If your argument is based on some fact (e.g. some number being above some threshold), then you absolutely need to back up that number to make the argument resolve into "it's most likely true". If you don't back the number up, then the argument resolves only to "if the data is indeed as argued, then it's most likely true".
Keeping beliefs with free variables in them is fine, IMO, as long as you're careful about not accidentally collapsing them into an outcome without having first checked the dependencies.
sure, i'd go one step further and note that there are no absolute certainties, so those free variables always exist, even implicitly. in that vein, most threshold numbers like that should at least come with uncertainty bounds (damn lies and statistics).
said another way, outside of math and formal logic, real-world arguments resolve at best to "if the data is indeed as argued, then it's most likely true" (perhaps gödel is relevant here, as @marcinzm referenced downthread[1], another citation!).
(edit: that's not to say provide no links, just that it won't make the argument, per my other comment, https://news.ycombinator.com/item?id=23899088, ironically, a citation.)