As with most things in life, and here on WikiTree in particular, we evaluate and decide what works best in a given situation. So any hard and fast rule is often not a good choice when you have humans involved, especially large groups of humans.
When I'm evaluating a relationship in genealogy, (not necessarily one under the umbrella of a project) I try to take into consideration how much research may have actually been done on a person, and how much more is possible before I remove any relationships. If it is a random coal miner in 1780, there are very likely only a couple of documents extant that would touch on his life. So his birth record and marriage record, if those even exist, may be all we could ever possibly know about him. Since there may never be proof that John Campbell baptised in 1760 in Cumnock is the same John Campbell who married in 1781 in Cumnock, but who had children with a correct "naming pattern," do we disconnect him? We look for other possible candidates, and say, finding none, we make a suppostion that he's perhaps/probably the same person and leave it hoping that further evidence might someday be uncovered. He's not important to thousands of people, probably only to the direct descendants who might someday prove more by DNA.
In the case of a well-known hero/villain of his time, even though it was in 1300, there may have been a little more known, available and studied about the subject. IF there was a marriage and there were children, and IF there was evidence of them, we would probably know by some contemporary documents IF those docs were extant at anytime in the last few centuries. That there is no concrete evidence doesn't prove that Wallace had no family, but it does tell us that they were never mentioned in contemporary documents and are likely to never be proven now. Could he have had family? Certainly. Do we have a single concrete document that says he did? No.
If that situation changes (DNA or some other magical thing we don't yet know about), we could always add newly-documented relationships. In the interim, on Wallace in particular, I think the Scotland Project's decision to leave any descendents mentioned only in the biography, along with their limited source info, is the safest route. It can always be revisited.
Is it important to try to add "supposed" family to an important historical figure like Wallace? I think it's more important to maintain a higher level of integrity for a profile that many people will be looking at for many different reasons. If all trustworthy sources found to date tell us we really don't know if he had progeny, why would we want to wander into the fictional side? It makes WikiTree, the managing project, and any project managers look incautious.