A number of posts in other threads mention randomness in D&D, and how that might be off-putting to some, and in some situations it just plain goes against the story.
For example, someone used the example of the cleric that has a 10 intelligence. For that cleric, even at mid level, his trained religion check would be only a 13 or 14 while even an untrained wizard of the same level would most likely be equal to or better (depending on Intelligence score of course). This is only one example, but many others exist.
So, should each class have an expertise mechanic (like the rogue or bard) so that key checks can double the proficiency bonus? (Like Fighters can pick from Athletics, Acrobatics, Intimidate; Rangers pick from Animal Handling, Perception, Survival, Nature, etc.) Perhaps choosing 1 every 4 or 5 levels?
Or, should there be a feat that grants a few so that anyone can gain expertise if the table uses feats?
Or, should a DM just let PCs that seem to have key requisites succeed without rolling at all? (i.e Cleric with training in Religion should be able to automatically succeed on a DC 10 Religion check -
What are some other options people have to address these situations where a PCs should be better in a key proficiency?
I'd suggest a different approach: different types of tests.
- General knowledge/trivia: standard roll anyone can try.
- Something anyone trained was exposed to but may not remember: standard roll, trained gets a +n bonus.
- Something anyone trained "should" know: Anyone who is trained is told the answer, no roll needed. Others may roll.
- Something only the initiated are told: Player was told answer and wll hopefully remember. Assuming the player forgets, only that player may roll to remember.
The first case covers the majority of the typical in-the-field player initiated cases. I find the last two cover the majority of "story" cases where there is a narrative expectation someone in the party knows a fact.