Late response to your question.
First, of course the landscape of trails is constantly changing and even dynamic / crowdsourced maps are always at least a little behind the latest state of things. I'm not looking for the nth degree of accuracy, just fantasizing about enhancements to the 80%.
Second, the phenomenon I'm bitching about is much mitigated by good trail network design, dedicated to riding. (Crossing spaghetti strands are frowned upon, and uninterrupted loops are a goal.) So if you're lucky enough to have such trails local to where you are, the topic isn't pressing.
Third, I have a bit of a singular perspective on this because my S.O.
makes MTB maps. She actually does go out and ride 99% of the trails she depicts. One of her original design goals was to show short stubs of the "wrong" trails, which are legion in our suburban area. Unfortunately the many land trusts and other open space stewards almost uniformly insisted that she NOT do that. They are deathly afraid of landowners squawking about "encouraging people to use illicit trails," even when DIScouraging accidental use of said trails is precisely the point. In fact in certain cases they don't want her including trails that the trusts' own literature indicates are open for riding. The reason? "We don't wasn't to advertise that too much." Personally I find this attitude unfathomable in an age when the world desperately needs more people experiencing the outdoors first hand, while it still exists. But again it's all about walking on eggshells with landowners and abutters.
Anyway, bottom line, my simple point is that when you're out there on the ground at an intersection, having a map that shows a two-tined fork when you're looking at a three-tined one is, yes, better than useless, but nowhere near as good as one that shows all three trails.