I'm a new DM of D&D 4e and I'm looking for some advice for handling failed and fumbled information rolls.

I am using a house-rule that allows for misinformation when a 1 is rolled on the knowledge check.

I think the game system is probably less relevant, i am more interested in player knowledge vs character knowledge.

Let's say the player does a nature check, to find out if a monster (let's say a Troll) is weak against fire: he rolls the dice and fumbles, so since he fumbled I would tell him that he is very certain that this Troll is immune to fire.

For the character this might make sense, but the player knows that he fumbled, so he can assume the opposite, right?

I am not sure what is the right way to handle this situation. What should I tell the player, who obviously knows that he fumbled for his character?

  • 8
    \$\begingroup\$ @Ernir: This question doesn't seem to be about the specific rules of fumbling, but more the general problem of player knowledge vs. character knowledge. OP should probably revise the question to clarify this. \$\endgroup\$ – Grubermensch Apr 16 '14 at 15:33
  • \$\begingroup\$ @SevenSidedDie I'm bringing up this edit conflict on meta: meta.rpg.stackexchange.com/q/3498/1204 \$\endgroup\$ – doppelgreener Apr 20 '14 at 14:19
  • 1
    \$\begingroup\$ Related: How do you discourage "player knowledge" as a GM. \$\endgroup\$ – AceCalhoon Apr 20 '14 at 16:04
  • \$\begingroup\$ @AceCalhoon Actually, I think this is a duplicate of that. The answers are even substantially the same. \$\endgroup\$ – SevenSidedDie Apr 20 '14 at 18:09
  • 1
    \$\begingroup\$ Maybe make the false knowledge not an automatic opposite? instead of weak to fire falsifying to immune to fire, maybe the false knowledge is immune to bright lights, or weak to ice, or something. Or even adding partly true knowledge to false knowledge ("weak to sunlight and maybe a bit to fire, but really vulnerable to water"), so it will be harder to sort false from true - after all, even true knowledge may be disregarded if it isn't believed. \$\endgroup\$ – Megha May 25 '17 at 2:27

Be unpredictable

The reason that the player can do this is that the rules are well-known. In order to avoid this problem, introduce some mystery. When the player rolls a skill check where the quality of the result shouldn't be known by the character, you should also roll (in secret of course). If your roll is in the high half of the die range, then treat his roll as normal (1 is failure, 20 is success, etc.). If you roll in the low half, then reverse the number on the player's roll.

Your player is no longer able to predict success or failure based on the die in front of them, because there's now a 50% chance that it is exactly the opposite. If you do this consistently, your players will simply stop trying to guess what's really going on.

Your players should try to separate themselves from their characters

In general, part of playing tabletops is being able to know things as a player and not know them as a character. You should encourage your players to do this, possibly by rewarding bonus XP (or even roll modifiers) when your players nobly ignore their own knowledge in favor of playing the character's unfortunate ignorance.

  • 5
    \$\begingroup\$ +1 for separating from one's character. I'd be careful with giving bonuses for this though - I'm not against the idea, but I think it should be reserved for major actions. We've had it happen in a game I'm in, when a player took a course that he knew may kill his character purely because the character didn't know that, and it was in their personality to want to do it. The reverse of the "roleplay bonus XP" idea is to penalise players who meta-game too much. As a DM, you can always be more cruel. If players abuse the system, the DM can threaten to do likewise to make things harder for them. \$\endgroup\$ – anaximander Apr 16 '14 at 17:41
  • 5
    \$\begingroup\$ Or... just roll yourself and hide the roll. There's absolutely no point in letting the player roll, if he doesn't know how to interpret the outcome. \$\endgroup\$ – o0'. Apr 21 '14 at 8:05
  • 1
    \$\begingroup\$ Furthermore, sure, it's easy to say "keep your knowledge away from your character", but it's not easy and it's not fun: whenever possible it's much better to sincerly not know better. \$\endgroup\$ – o0'. Apr 21 '14 at 8:06
  • 1
    \$\begingroup\$ @Lohoris There's no point in it from the DM's perspective, or from a cold statistics perspective, but for whatever reason a lot of players don't feel like they're playing a game unless they get to roll dice, regardless of whether the die roll actually matters. \$\endgroup\$ – DCShannon Aug 28 '14 at 18:16
  • 1
    \$\begingroup\$ @Lohoris It's the DM's job to help the players have the sort of fun that they enjoy, not explain to them what is or is not fun. Rolling dice is a form of Sensory Pleasure. \$\endgroup\$ – DCShannon Aug 28 '14 at 21:54

In D&D 4e, "fumbling" a skill check on a natural 1 is a house rule only. By the rules as written, a natural 1 on a skill check is not even an automatic failure, much less a fumble — it's just 1 less than a roll of 2.

The critical hits and automatic misses introduced in D&D 4e are only in the context of combat, and nowhere else. (This is why you won't find a general rule saying there aren't automatic misses on skills — because nothing ever says there is.)

Also, in general, failing a Knowledge skill check means "you don't know the answer" or "you're not sure", not "you think you know something that's actually false".

  • 5
    \$\begingroup\$ You could also take this to mean that the quality of the roll reflects the quality of your answer. Knowledge check of 5? "You hypothesize Trolls are actually a type of ghost." Check of 15? "You've heard trolls are vulnerable to fire, but the man you heard it from was really drunk." Check of 35? "You know of scholars who are trying to figure out why fire is so dang good against trolls." \$\endgroup\$ – PipperChip Apr 16 '14 at 17:24
  • 9
    \$\begingroup\$ Fumbles and auto-successes on skill checks is strongly disadvised. Aspects of the system works very poorly with those in play. \$\endgroup\$ – KRyan Apr 16 '14 at 18:05
  • \$\begingroup\$ @Glenn I've updated your answer with 4e's actual content on the topic, and eliminated the 3.5e stuff. (I cannot stress enough: you should avoid answering 4e with 3.5e material like you should avoid answering a chess matter based on checkers.) Could you check it to see if it still conveys what you'd like it to convey, and edit it if necessary? \$\endgroup\$ – doppelgreener Apr 20 '14 at 12:54
  • 1
    \$\begingroup\$ This answer would be much better if it also answered how to handle player knowledge with fumbles as a house-rule. If they choose to continue using fumbles, this answer doesn't provide any help for them. \$\endgroup\$ – SevenSidedDie Apr 20 '14 at 17:34

Here's a range of options, suited to different playstyles. Also you can mix-and-match.

  • "You Don't Know": If the player fails a roll, the GM says they don't know. This is probably the simplest approach.

    What about "botching?" Shouldn't you make it more than just "You don't know." Enh, maybe they still don't know. Even in games that feature special "critical" success and failure, sometimes it's better to just make a skill check a binary thing rather than try to shoehorn extra outcomes.

  • Secret Rolls: Make the roll for the player, hidden in some way (you can use a GM screen, if you have one — I don't recommend getting one just for this purpose, though). Tell them the result of the check without ever telling them a value.

    This is a popular method because it's easy to implement within the framework of a traditional-ish RPG.

  • Social Contract to Avoid "Metagaming": Let the player roll the check. But then, regardless of what the player knows, their character acts based on the information provided by the result.

    I don't recommend this approach for games with a strong challenge-focused tactical element, since it's very frustrating to know the "right" answer but not be able to do it. Some players may also complain that having this separate knowledge hurts their sense of immersion.1

  • Unreliable Information: Provide a lot of information, some of which is legit and some of which is erroneous. A lot. Like, a deluge. Now the player has to deal with sorting it out. Acting on it is a huge risk, but so is not acting on it.

  • Collateral Consequences: The "critical failure" costs you something extra. For example, maybe the skill check is normally a freebie, but if you fail, it wastes your next action since you're standing there kinda dazed.

    This works smoothest in games with relatively flexible resolution systems. But it's also a reasonable way to add a dash of weirdness and whimsy to a grids-and-powers kind of game — if that's what the whole group is looking for. If you use this option, you should also probably also add extra results to "critical successes." (And don't make the "critical failures" ridiculous or punishing. A dash of comedy or a minor setback will get you a lot further than nonsense like "Your pants fall off" or "You chop your own head off with your axe.")

For a general discussion on how to come to agreement about cool consequences for die rolls, right there at the table, I recommend this answer.

1 - What "immersion" actually is can be a very, very slippery thing. In this case you can pretty much take people at their word if they say they feel a bit of cognitive dissonance playing this way, though.

  • \$\begingroup\$ partial-truths work wonders as well in my experience. The player knows it's wrong, but doesn't know what is right still. \$\endgroup\$ – Mooing Duck Apr 16 '14 at 19:53
  • 1
    \$\begingroup\$ Very nice compilation, +1. I'll add that I've enjoyed a very special-case technique lately: "Oh, you failed the roll. Okay then, the answer Calesta dredges from her memory is that the troll is vulnerable to cold. It's so, so very wrong, but if you have Calesta act on it, you get XP... So whaddya do?" Very niche though; it needs the right creative agenda, and the right balance of value between correct action and bribe currency. \$\endgroup\$ – SevenSidedDie Apr 16 '14 at 20:16

The typical way to do this is to roll for the player where only you can see the result (such as behind a screen or your hand). Any time you are rolling for hidden information, you're justified in making the roll yourself.

Looking for secret doors? You roll, and tell what they do or don't find. Racking their brains to remember something useful about trolls? You roll, and tell them what they do or don't remember.

Note that the issue of hidden rolls versus rolling openly is a topic that people are quick to heatedly debate online. Neither is more correct, as it is merely a preferences of play style. Hiding knowledge/search–type rolls solves the problem of players using "metagame" information that their characters wouldn't know, is all. If you don't have that problem, then instituting hidden information rolls isn't solving anything.


This can be difficult in many games, since while it's all very well to tell people they have some sort of moral obligation to separate player and character knowledge, feeling obliged to compromise your character's safety or the party's goals in the name of "good play" can be an unfun catch-22 for some players. And of course acquiring all that player knowledge is a big part of the fandom for many people.

Personally I find informed decisions the most fun in 4th Edition, so I conceal very little mechanically-relevant information from the players (most of the value in skills comes from earning XP in Skill Challenges IMO, so I don't see this as sidelining the niche of knowledge skills to any meaningful extent).

If you don't feel that this can be resolved satisfactorily with simple social contract against metagaming at your table (as AlexP suggests), I have two suggestions in a 4E context:

  • Reskin like crazy. I find this helpful anyway, but if a player can't rely on a troll having the same mechanics as they're used to, they need that knowledge check to be able to match up their player knowledge ("Oh, it's a reskinned Aboleth, cool, I know what to do.")

  • Give a standing quest to not use player knowledge, rewarded every session/level/whatever. The key difference here is that it's not a bonus you have to grant, but something the players can choose to throw away if they feel they really need the player knowledge - an interesting decision in its own right.


The core of the problem is that when this happens, the player has two choices: a) make their character act particularly ignorant about this thing b) connive some way for their character to figure it out

Neither of which is fun.

Another option, is to use your GM powers, and modify a small aspect about the monster. "Yes, you remember trolls are generally weak to fire."
"Great, I cast firey-fire attack at it!" "It's engulfed in flames, then the flames die out as you see it has coated itself in a protective frost layer... oh, no, you were wrong. This is an ICE Troll, a somewhat uncommon variant species. They aren't vulnerable to fire like their cousins..."

Another alternate option would be to say: "Ok, you're going to have to roll a D6 - this is how many rounds before your remember that they ARE weak to fire." or similar. This means their character isn't totally ignorant, they're just caught off guard and it takes a little bit for the information to come up.

  • 1
    \$\begingroup\$ I prefer this approach. It makes the game much more resistant to system-level metagaming (e.g. "oh my character is fighting a troll; D&D trolls are vulnerable to fire so my character should use fire-based attacks") without players having to contrive behaviors they know aren't in their character's best interest or scheme up a way to "accidentally" learn this. \$\endgroup\$ – Doktor J Apr 17 '14 at 13:36

I don't like the "You don't remember anything" way. I use this method with my players (the game is Pathfinder, but I think this advice is quite general):

  • I don't let know them the DC of the roll
  • If they have success, I say them clearly useful informations ("trolls are weak to fire")
  • If they fail, I say them some silly lore or wrong informations ("trolls stinks so much that people get sick").

Usually they understand when the fail and when they have success, but they cannot infer correct informations.


I am going to share my ideas on gaming, from various systems over many years. While it does not directly address the OP's question per se. I think it might be helpful.

Meta gaming is a reality of the process, how you handle it should be part of your gaming philosophy and you players should have a basic understanding of your philosophy. Here are some of the ways I approach gaming.

  1. The cardinal rule that can never be broken is have fun. That does not mean no disagreements or unpleasant things happen. At the end of the session everyone should leave smiling and wanting more.

  2. Almost as important as the first. You are telling a story (really that is all you are doing) with lots of input from your players. All of the elements of a good story of book need to be there.

  3. I customize everything to some degree (usually a lot) to fit the feel of the story being told. Trolls, don't have to be susceptible to to fire, or regenerate, etc. This mitigates player knowledge and meta gaming based on it. In my experience players generally find this to be a good thing after they get over the initial surprise.

  4. Meta gaming that is small in scale, reasonable for the situation, and does not break the plot should be overlooked unless it gets out of control.

  5. No character should ever die because of good role playing. Its fantasy, if a character chooses sure death to save the party, or achieve some gravely important goal throw them a bone. Heroic fiction, daring doo and all that.

  6. The characters of Stupid, selfish, greedy, etc, players suffer and can die, even if they never take any risks at all. This applies to the excesses that disrupt the story or group's enjoyment, not good role playing.

  7. Mechanics wise, for every action, decision, and often die roll the players make I roll a private modification. (I roll a lot of dice when I play) This allows me to account for external or unknown (to the players) forces that might affect the outcome of an action. This keeps the players on their toes, because their die roll may not always be the final answer. They could have performed an action perfectly but because of an unknown it had no chance of success. And most importantly it allows me to guide the story in the correct direction when it needs a nudge.

I hope this has been helpful.


To add a bit more for the OP. First and foremost the role player should role play. They should do their best to live the character, even if it means doing something that the player knows will not work. They then have the opportunity to learn in character from their mistakes and grow as a person (character). See point 5.

I touched on player knowledge and customizing aspects of your story in point 3. I noted some other answers kind of said the same thing. There was a reference to an ice troll in one answer. Take this concept up a notch. Consider the monster Manual or rule book of your choice to be a book of lore rather than law. What the player (thinks he) knows is actually the lore that the character knows from fairy tales, stories, and legends. This will make the meta gaming to role playing transition easier for the player to live with and eventually enjoyable. This will also teach players to not meta game without you always having to spank them on the hand with a ruler as a GM. They will also come to enjoy your creative license.

  • \$\begingroup\$ This looks like it could be a really good answer if you were to, at the bottom of your post, indicate how you would apply it to the OP's problem. Other than that, welcome to the site and if you have questions about how it works check out or FAQ. Once you have 20+ rep, you can also come join us in chat. \$\endgroup\$ – Dakeyras Apr 17 '14 at 19:38
  • \$\begingroup\$ I did touch on it the OP's question in number 3 a bit. Some other answers had similar suggestions. I will try to come up with a decent example and add to the post. \$\endgroup\$ – Tony Heflin Apr 17 '14 at 20:12

Not the answer you're looking for? Browse other questions tagged or ask your own question.