A good point is a good point. Does not matter who or what the source is. Titles are irrelevant. They don't make things more or less true. Just like numbers of people or popularity don't make things more true. Seems to me that you all are more concerned about the "source" rather than what actually works. Usually this is solved by simply doing (conducting experiments in real time with all concerned bodies present) rather than arguing but that's not an option. Truth is you can argue till your ego's go blue in the face, but projecting results in a vaccum without accounting all variables is pointless. The most you can do is provide a point and hope Larian rolls for insight.
You realize a 'good point' is a point that's been qualified right? Which means it's inherently tied to a context which includes the source. While it doesn't necessarily subtract away from the truthiness of a thing, it does mean it may fit less into your context to which you're trying to apply the 'good point' to.
And the truth is that we do have in place formal rules of logic and other systems that make arguing/debating/discussing these things meaningful/fruitful and depending on the subject matter doesn't necessarily require you to run an experiment (usually because it's already been validated by the processes outlined previously.
Hmm I should define "good" to be consistant? That's fair. My definition of good is a hypothesis that based on a contextual setting with the malleable point"can it work" can be tested. It's loose. You are correct on this. I promote real time testing because of this in fact. The meaning comes from the testing in real time not the past results. In order to develop a creative break through you have to do things yourself. Using templates is fine, but they never define what you do. To discover something you must be open to the fact that nothing is concrete. You cannot do that if the data is thrown out before at least the screening phase of testing because the source is something you don't like or worse something that is beneath you. Which is exactly what most ego's demand. I prefer a more humble hands on approach to prevent loss of creativity and stop lazy "he said she said" arguments that always pop up when reliance on old data takes precedence over live data testing. Things are easier now so it's easy to swing past little things, but those little things could be what makes the difference. This of course is strictly for the discussion of Baldurs gate 3.
So as you already know this cannot happen which by default means the limit is post your point. Its all Larian after that.