Quote Originally Posted by MaxWilson View Post
I don't quite follow, maybe because I never played 3E and very little 4E. In AD&D2 terms this is like a nonweapon proficiency in, say, Mathematics. Two trained mathematicians may overlap a lot, but if they are close in skill each of them is still likely to know at least SOME things the other does not, and the roll of the die is to represent that variation in exactly which treatises have been read and remembered and which conversations have occurred with other scholars. The one with a higher score will know slightly more on average and be right more often, but not always.

In 5E terms I do like feeding my players info, especially about monsters, both via sages and via character background knowledge, e.g. I roll everyone's Arcana and based on the results hand everyone an index card summarizing what they've heard is true about vampires in this world. I secretly love it when the most ignorant guy (lowest bonuses) gets the best result and the most accurate info but nobody believes him because he's always wrong. ("Of course you don't need wooden weapons to harm them, Creed! We have magic weapons and those are better!")
I think what Tanarii is getting at is in 3/.5 knowledge checks had a clause that you could not retry them unless you gained an additional rank in the skill. Essentially, if you rolled a 1, that was your knowledge for that topic at least in the short term. It made it possible that more generally knowledgeable characters could still be beaten on specific topics sometimes, coupled with approximately 80 bazillion knowledge(subskills) it could make for a very diverse team of scholars.