The Three Most Misunderstood Rules in D&D and Pathfinder

Hey, you. Yes, you — the one who paused combat for twenty-three minutes last Tuesday to debate whether the rogue could hide behind a barrel that was definitely too small to hide behind. Or maybe you’re the DM sitting across from that person, slowly gripping your screen in both hands and wondering whether game night was a mistake. Either way, hi. Welcome. You are in exactly the right place.

Here’s the uncomfortable truth that no rulebook will tell you: the most dangerous monster at your table isn’t the lich, the dragon, or the overpowered paladin/warlock multiclass that someone “just wanted to try.” It’s a rules argument that nobody wins and everyone remembers.

D&D and Pathfinder are massive, beloved games played by an estimated 50 million people worldwide. D&D Beyond alone has seen more than 30 million characters created on its platform. That is a staggering number of people who, at some point this week, paused a fight to argue about stealth. And the wild part? Most of those arguments are happening over rules that have perfectly good answers — answers that are just buried, badly worded, or actively contradicted by other parts of the same book.

So here are the three rules that derail more sessions than any other, why they go wrong, and — most importantly — the actual fix you can run at your table tonight.

Rule 1: Schrödinger’s Rogue — How Hiding Actually Works

Let’s start with the one that has caused more mid-combat courtroom dramas than anything else in the history of this hobby: hiding.

You know the scene. The rogue uses their bonus action to hide, rolls a 24 on Stealth, and then announces they’re going to walk out from behind the pillar and stab someone. Another player immediately objects. The DM pulls up three different YouTube videos on their phone. A Wikipedia-style citation war breaks out over the exact wording of the Player’s Handbook. Forty-five minutes later, everyone is tired, and nobody feels good.

This happens because the hiding rules in D&D are genuinely, sincerely bad. And the 2024 Player’s Handbook, which was meant to fix this, arguably made it worse.

What the 2014 rules actually say: To hide, you need to be “heavily obscured” or out of direct sight of enemies. Once hidden, you become “unseen and unheard.” Jeremy Crawford, the lead rules designer for 5th edition, clarified on this publicly: “Invisible means unseen. Hidden means unseen and unheard.” That distinction matters — because when you attack from hiding, you break the hidden condition after the attack resolves. That means the rogue gets the advantage on the attack. They just can’t hide again until they’re back in cover.

What the 2024 rules changed: The new Hide action grants the Invisible condition rather than a standalone “hidden” state. That sounds cleaner on paper. In practice, it created a new problem — the Hide action’s own text contains a line-of-sight requirement that conflicts with how Invisibility normally works elsewhere in the same book. Multiple rules analysts, including those at Merric’s Musings and Tabletop Builds, have flagged this as a direct contradiction within a single paragraph. So yes — the update managed to generate a brand new rules debate inside the fix for the old one. Incredible achievement.

What Pathfinder 2e does: PF2e actually has the clearest stealth framework of the three. Enemies track you across three states: Observed (they see you), Hidden (they know roughly where you are but can’t see you), and Unnoticed (they have no idea you exist). You only need cover or concealment at the end of your movement to Sneak successfully. The system has friction, but it has logic.

The table fix: Stop rolling Stealth every single round. Make one Stealth roll at the start of a “scene” — when the rogue first enters a dangerous area or tries to re-hide mid-combat. That roll stands until circumstances change dramatically, like a torch being lit or the rogue running into the open. Use the passive Perception of enemies as the floor; if nobody in the room has a passive Perception above 12, a rogue who rolled an 8 can still melt into the shadows because, narratively, nobody in that room is particularly observant. Set the expectation before the session that the DM’s ruling is final during combat, and then you can debate after the session if you want. The fight continues. The rogue gets to be cool. Everyone wins.

Rule 2: The Reaction Tax — What Opportunity Attacks Are Actually Doing to Your Combat

Here is a thing that happens at almost every table, in almost every campaign, in almost every fight: the monsters don’t move.

Not because they’re tactically brilliant. Not because they’ve been told to hold the line. They don’t move because moving costs them free damage from every player in reach. Damage they can’t afford, triggered automatically, just for having the audacity to take a single step. This is the Opportunity Attack system, and it is quietly making your combat worse every single session.

But the rules debate isn’t just about whether OAs are good design. It’s about what actually triggers one.

The most common mistake: Reactions are not actions. This sounds obvious, but it almost never plays out that way. Opportunity attacks are a reaction, which is a separate resource from your action, bonus action, and movement. Players regularly forget to use their reaction at all, burning through entire combat rounds with a perfectly good Counterspell or Shield spell they never remembered to fire. Conversely, DMs sometimes forget that monsters with reactions can use them for more than just OAs.

The second most common mistake: Standing up from prone does not provoke an opportunity attack. Crawford confirmed this clearly and directly: standing up uses half your movement speed, and movement of that kind does not count as “leaving a creature’s reach.” Neither does teleporting, being shoved, or being carried by forced movement. Only voluntary movement out of reach triggers the OA. Your wizard can Misty Step right out of a grapple without gifting the grappler a free hit.

The Pathfinder 2e angle: In PF2e, Attacks of Opportunity are a class feature — not a universal ability that every creature possesses. Your Fighter has it. The random goblin you’re fighting almost certainly does not. This is a feature, not an oversight. It means movement is genuinely meaningful for most of the game, and the Fighter’s ability to punish that movement is part of what makes them feel distinct. If you’re running a PF2e game and treating every enemy like they have AoO access, you are making the game significantly harder than intended and accidentally removing one of the system’s most elegant design pillars.

The table fix: Print a “Reaction Cheat Card.” Seriously. It’s a piece of paper that lists the four most common things a reaction can do at your table: take an opportunity attack, cast Shield, cast Counterspell, use a class-specific reaction ability. Put one in front of every player. The number of times a player has said “wait, I have Shield?” at the end of a fight they barely survived is heartbreaking. Once everyone knows what their reaction does, they start using it — and suddenly the fighter thinks twice before running past the enemy wizard, because she definitely has Counterspell loaded and is definitely looking at you.

Rule 3: Critical Theater — The Nat-20 Myth and the Knowledge Heist

This is the one that lives rent-free in the heads of approximately half the D&D-playing world, and it is a lie. A beautiful, intuitive, completely understandable lie, but still a lie.

The myth: Rolling a natural 20 on a skill check is an automatic success, no matter how high the Difficulty Class.

The truth: Critical successes and failures — the automatic hits and misses — only apply to attack rolls and death saving throws. That’s it. A natural 20 on an Arcana check with a +0 modifier against a DC 25 is still a 20. Which is still a failure. Your character is not suddenly the world’s greatest scholar because the dice gods briefly smiled.

The reason this myth persists is that it feels correct. Dramatically, narratively, it makes sense that a perfect roll should mean something spectacular. And here’s the thing — it should mean something. It just doesn’t mean an automatic success in the rules as written.

This exact misunderstanding has a twin in Pathfinder 2e, and it might be even more damaging because it involves the Game Master actively withholding information from their players.

Pathfinder 2e’s Recall Knowledge problem: Recall Knowledge is the action players use to learn useful facts about enemies — immunities, special abilities, weaknesses. It is, in theory, one of the most interesting information-economy decisions in PF2e. In practice, many GMs treat a successful Recall Knowledge roll as an excuse to hand over one deliberately vague, uselessly abstract piece of lore. “You recall that this creature is known to haunt cold places.” Great. That’s the whole answer. Thanks.

This is wrong. Michael Sayre, Paizo’s Design Manager, addressed this directly: “Useful information, in the context of this game, means actionable. If succeeding at the check does not offer a reason to change your course of action based on the information revealed, it is not useful. If the end result would be the same as if you had failed the check, it is not useful. If your GM is opting to give you useless information, that is a failing on their part.”

Read that last sentence again. Sayre — the person whose job it is to defend this system — called it a failing. That is a remarkable thing to put in writing, and it happened specifically because the community was getting this rule so catastrophically wrong.

The table fix for D&D: Adopt what I’ll call the “Passive Floor” rule. A natural 20 on a skill check doesn’t automatically succeed — but it does grant the best possible version of a partial success narratively. The DC 25 Arcana check that your character fails with a 20 still fails, but the DM describes the failure as “you know you’re on the edge of remembering this — the answer feels close.” You’re not magically competent. You’re almost competent. That’s a dramatic difference, and it keeps the fun of the big roll without dismantling the DC system entirely.

The table fix for Pathfinder 2e: Before a player makes a Recall Knowledge check, the GM quietly decides in their head what a success will reveal — and commits to something actionable. If you can’t think of a fact that would cause the players to change their behavior, the monster probably shouldn’t have a Recall Knowledge payoff for this particular roll. Try again with a different skill, or describe a visual detail that implies the same information. The roll should mean something. Every time.

The Actual Problem Underneath All Three

If you read these three rules and noticed a pattern, congratulations — you’re right. The common thread isn’t bad dice luck or difficult players. It’s that all three of these rules share the same design problem: they’re written to be technically correct and practically ambiguous at the same time.

Jeremy Crawford put it plainly in the Sage Advice Compendium: “Many unexpected things can happen in a D&D campaign, and no set of rules could reasonably account for every contingency.” That’s true. It’s also, if we’re being honest, slightly convenient. Fifth edition was designed around the principle of “rulings, not rules” — the idea that the DM’s judgment fills the gaps. That philosophy is great when the gaps are small. It is less great when the gap is the entire hiding system.

The fix isn’t to learn every rule to perfection. The fix is to agree, before a session starts, how your table handles the three or four rules that always cause trouble. Write them down. Stick them to the DM screen. Settle it once, in advance, when nobody is mid-combat and emotionally invested in the outcome.

The rogue gets to be sneaky. The fighter gets to punish movement. The bard gets to feel smart for knowing a monster’s weakness. And you get to spend your session actually playing the game instead of litigating it.

What to Actually Do Right Now

If you’re a Game Master reading this: before your next session, pick the one rule from this list that causes the most friction at your table. Write one sentence — literally one sentence — that states your ruling on it. Put it somewhere visible. Tell your players at the start of the session. That’s it. That’s the whole fix.

If you’re a player: stop Googling mid-combat. I say this with full compassion — it never helps. Bring the question up after the session, in a calm moment, with tea and no initiative order to worry about. You’ll get a much better answer, and your DM will love you for it.

The three most misunderstood rules in tabletop RPGs are not mysteries. They have answers. The trick is making sure those answers live at your table instead of on page 147 of a book nobody opened during the argument.

Now go run a session. And please, for the love of all that is holy, let the rogue hide.

Found this helpful? Share it with your DM. Or your player who argues about stealth. Actually, definitely share it with that person. They need it more than the DM does.