I had this issue with a knowledge check in our last session. Such checks do not require proficiency, everyone can try them. The paladin wanted to know if he could recall anything about a god the group had learned about. After I asked for an Intelligence (Religion) check and he failed (his Intelligence 8 did not help, even though he had picked Religion in paladin school), next the wizard tried, and after the wizard failed, the monk, and that failing, the fighter.

The game has group checks that one can use if the group tries to achieve something as a group. But in this case, it originally is not a group undertaking; and it also would not make sense that two people out of four need to know something, for the group to know something. And because the best character's skill is always more likely to succeed than a group on average, where everyone else is worse, mandating group checks here means the players would be better of for just one of them trying to remember something. But that seems to make no sense.

Most characters do not have that many skills to pick, so my players tend to focus on those that have most impact on gameplay, like stealth, perception, investigation or athletics. Knowledge skills other than Arcana for wizards are picked mostly for roleplaying reasons. For a classical party of four like ours here, against a Medium Difficulty (DC 15) check, the one character that has the knowledge skill has a 35% chance to succeed. The wizard has a 45% chance. The other two characters are not proficient and at 10 Int have 30%.

Even with those measly values, the chance that at least one of them will get there for a Medium difficulty check is about 83% (thankfully, they don't have guidance, or it would be well over 90%). That is too high for my taste -- for example, combat is calibrated around a 65% to hit chance: enough to often have the satisfaction of success, but still with a large enough risk of failure to keep the rolls interesting. With 83%, it feels more like you expect to succeed, and not making it is a let-down after all this effort.

Does anyone have good, practical ways, to avoid this issue? I've thought about only allowing rolls for those that have proficiency, but that effectively would limit it to a single character, and the players would probably deem it unfair. Or I could just accept it (my current approach).

  • \$\begingroup\$ This question is a potential system agnostic dupe target. \$\endgroup\$ Commented Dec 16, 2023 at 21:30
  • \$\begingroup\$ Let me read it, and if it helps we can close as dupe, as far as I am concerned. This one has some things that are specific to 5e re chances, available methods under the rules etc., but what I am really looking for may well be agnostic. \$\endgroup\$ Commented Dec 16, 2023 at 21:32
  • \$\begingroup\$ My answer here would support the idea that running it as a group check probably isn’t the way to go. \$\endgroup\$ Commented Dec 16, 2023 at 21:33
  • \$\begingroup\$ @ThomasMarkov Looking at the agnostic one, the question is this issue, the agnostic answers however suggest things that are contrary to the 5e rules - for example "Let it Ride", i.e. only one check may be made, or help by the all other players, where the 5e help action only benefits form 1 helper. Or not rolling if there is no chance of success/failure that makes a difference - but there is a difference in outcome. Zachiels at least has the interesting idea to ask all to roll right away, to speed things up. \$\endgroup\$ Commented Dec 16, 2023 at 21:48
  • \$\begingroup\$ "the chance that at least one of them will get there for a Medium difficulty check is about 83% [...] for example, combat is calibrated around a 65% to hit chance:" You're comparing completely different things, though. That' 65% chance to succeed on one roll for one character only, Which you compare to the chance four characters doing a roll each to succeed. Moreover, combat is not the same - you don't just roll once and stop. Each character is expected to attacks more than once in a combat. Keep trying. While with a knowledge roll most of the time it's one-and-done. \$\endgroup\$
    – VLAZ
    Commented Dec 16, 2023 at 22:16

4 Answers 4


A character must conceivably have that knowledge in the first place to be eligible to roll.

This is related to the "only roll when there is a chance for success" guidance. To be eligible to roll for a knowledge type skill check (often as an Intelligence (History) check), a character must conceivably have at one time in their life acquired that knowledge. In the example you give, it makes perfect sense that a paladin, as a student of the divine, would have knowledge of the divine. But why would the wizard, monk, and fighter have that knowledge in the first place?

This is a world building and character development opportunity.

This question, “why would you know anything about this?”, is actually a question I ask players in the moment, to allow them to do some in vivo world building and back-story building. If the fighter wants to make a skill check related to obscure knowledge of some god, then they need to come up with a good back-story reason explaining why they would have ever known about that god at all. Doing stuff like this regularly leads to the players slowly building up their backstories. So this might play out for your fighter like this:

Paladin fails

Fighter: I'd like to try to roll that too.

DM: Why would your character know anything at all about this?

Fighter: thinks for a moment . . . Because my father was actually very religious during my childhood, but later left faith behind after my mother passed away.

DM: Roll with disadvantage. (or straight roll/advantage depending how you think their answer should influence the roll.)

Scenes like this have played out in my games many times, and they lead to plot hooks and plot points that you can keep in your back pocket. Maybe later on we find out that the fighter's father was excommunicated from his church after being falsely accused by a evil, deceitful priest. Boom, you just got a character developing side quest where you find out that the priest actually murdered the fighter's mother, just because the fighter wanted a shot at the roll. The idea here is not to prevent repeated checks entirely, but rather to use this kind of off-the-cuff character development as a carrot-on-a-stick for getting another shot at the roll. Taking this further, the information given can also be tailored to the backstory of the character making the check. The paladin, if successful, would have been given a detailed summary of whatever divine knowledge was at stake, but the fighter, with their limited exposure to the faith as a child, perhaps would not recall all of the details available, even with a successful check.

Doing this turns a one-dimensional procedural exercise (vulnerable to being gamed as described in your question) into a multi-dimensional, engaging roleplaying opportunity with real stakes.

  • \$\begingroup\$ I think what I like about this answer is that it is entirely within the rules, and does not just outlaw multi-rolling, but it instead enriches the game and hands it to the players to work for a better outcome with a little creativity, instead of just mechanically rolling. \$\endgroup\$ Commented Dec 17, 2023 at 7:28
  • 2
    \$\begingroup\$ Your last sentence should be a test the writers use every time they finish a rule, it is exactly what a DM should think about all the time. \$\endgroup\$
    – SeriousBri
    Commented Dec 17, 2023 at 10:01

Tell people stuff, don't make them roll at all

This is more of a frame challenge, because I don't think knowledge rolls should be used, and I am not 100% sure they are even a thing in 5e.

This isn't a RAW answer, but I do think it works so much better. My first character was a lore bard and after a bunch of failed rolls my immersion was broken with just how little my character actually knew, so I needed to change it.

In my games knowledge is proficiency based, if you aren't history proficient then you can't roll (unless you have some specific reason to know in your backstory), and if you are history proficient then you automatically know.

I don't understand why a DM would ever hide something about their world behind a roll. You went to effort creating that lore, it means something in your world and then you create a situation where you don't tell anyone about it.

DMs see their encounters subverted and avoided all the time, but knowledge doesn't have to be that way. Just tell people about your world, and use the players active interest in a subject to do it in a way that isn't just a forgettable lore dump.

Your players want to know about the amazing world they adventure in: tell them!

So how do I do it:

Sarah the barbarian poses a question about what they know. I ask if they are proficient in a relevant skill, they say no, I tell them they don't know much, they heard a rumour once about how the castle may be haunted, but that's about it.

This piques the interest of the rest of the group, I ask who is proficient in the skill and narrate the discovery of all the lore as a conversation between the players. This was everyone who took the proficiency feels they contributed.

Nobody has proficiency? Well I give them another way to learn it, ask a local or find a book etc.

Question is about impossible to know knowledge? Well nobody knows that so neither do they and a roll wouldn't change it, but maybe they know enough to set them on a quest to find out!


Maybe a bit of a frame challenge, based on playing different RPGs.

Some skills checks should not be rolled. If the barely literate barbarian has a lucky roll, but the smart wizard has a bad roll, it might not make much sense to for instance read an inscription.

So the way I do it is a bit like passive perception. PCs have a "passive intelligence", the checks have a DCs, and a character might or might not succeed without rolling.

But this is only for no pressure situations, where there is time.

If there is pressure (need to understand this scribble on the wall to open the door while monsters are closing in) then the checks and randomness become relevant to simulate stress.

This does not completely answer the initial question, but it limits a lot the cases where it happens and works well for me, with I find a more natural and satisfying role play.


Different Skills
Give the options to roll different skills depending on what type of knowledge they want. For the example of a deity, players could roll Religion for the deity's role in the pantheon, Arcana for the abilities the deity grants their followers, or History for the cultural impact of the deity and how they have acted in the past. The trick to really making this work is telling the players they all have to declare what skill they will role before anyone roles.

This creates an interesting tactical situation. If everyone picks the same skill, the chance of getting that information is extremely high, but the chance of getting the other information is 0%. If everyone picks different skills, the party has a chance to get all the information, but any specific piece has a lower chance.

In my experience, players tend to spread out and pick different skills, because they will have different relevant skills they are proficient in (e.g. Wizards rolls Arcana, Paladin rolls Religion, Fighter rolls History), and players with no good skills end up using the Help action to give advantage to someone with a higher modifier.

Bad Rolls Give Bad Information
This option is dangerous, but I have done it before and it has been hilarious. Lets say two players roll to determine if a certain plant is poisonous and the elf rolls fantastic and the dwarf rolled horrible. I would tell the elf that the plant is fine and the dwarf that it is a deadly toxin that kills anyone who even smells it. In my group, the results would be ~10 minutes of funny RP where the party tries to convince the dwarf that tomatoes are not deadly and elves are not demi-gods for eating them.

This works because the players know the results of the rolls, so they know which information is true and which is false. In addition, the false information I gave was so extreme, some type of conflict would have to result. Since my players are fine acting out character conflicts in funny ways (it is honestly at least 25% of a session), me giving out false information like this facilitates the type of play they want rather than impeding it. If the information was more plot-relevant or my players didn't feel comfortable with that same type of RP, then I would not give false information in this way.

Homebrew: Sum for Group Skill Checks
This is a homebrew rule I use, but I think it would help. Normally for group skill checks everyone rolls and if half the party passes, everyone passes. Instead, I have everyone roll and compare the sum of the rolls to the DC. Usually I set the DC as:

DC = What the DMG Recommends for a Check of this Difficulty * Number of Players

I usually also take a few points off the DC just to give the players a little bit of an edge, but if you don't the results of this are very similar to the base game. Mathematically, this system is the same as taking the mean of the rolls and comparing it to the DC recommended by the DMG. The biggest difference is that a player with a low modifier feels like they are contributing, since adding any amount to the total makes the total larger. This isn't true, but to a person there in the moment it feels that way.

This system could be interpreted as the party holding a conversation and everyone contributing what they know. If what they each individually know is enough, then the group pieces together the information to gain a larger picture.

  • \$\begingroup\$ I think this could work if you as the DM ask the players to suggest a skill; under the 5e rules the players have no right to declare they want to roll a skill (DMG, p. 237): the players declare what they want to do, and the DM decides if a roll is approrpiate, and if so what skill they roll. They cannot just say "I roll xyz". In my situation, they would say: "what do I know about this?", and I would tell them, "Roll an intelligence check. You can add your relgion bonus if you have one", so I'd have to look at their sheet and suggest a skill, they have no tactical choice, other than recommend. \$\endgroup\$ Commented Dec 17, 2023 at 7:13
  • \$\begingroup\$ Also, what stops a player to say; "OK, what do I know about this? Does my training in history tell me anything that this deity was involved in? I also happen to have studied Arcana - can I remember anything being mentioned about some effect or artifact related to the deity? Oh, and even though I never attended temple school, do I recall something from the general visits to the temple sermons with my parents about relgion?". Then you have to let them roll checks for all these things, and if all players do that, that just makes the problem worse. \$\endgroup\$ Commented Dec 17, 2023 at 7:19
  • \$\begingroup\$ Your homebrew mechanism is an interesting approach, but is very tough. This makes the chance for success worse than a simple helped action or a group check, or a single player check. See this question and answer for a detailed explanation why that is so. For it to work well, you have to calibrate how much you take off the multiple. \$\endgroup\$ Commented Dec 17, 2023 at 7:25
  • \$\begingroup\$ In my group at least, it is very normal for players to ask to roll a different skill. Player: 'I want to look for clues.' DM: 'Roll Perception or Investigation.' Player: 'Can I roll Survival to find tracks left behind?' The DM then decides if the player's justification is appropriate. If a player tried to roll multiple times for one goal, I would say each roll requires time to pass. If they are in a dungeon, spending time on this gives the enemy time to prepare. If they are in a safe location, I would treat their request as a downtime activity rather than a simple skill check. \$\endgroup\$
    – E Tam
    Commented Dec 20, 2023 at 5:17
  • \$\begingroup\$ The math of my homebrew is not as simple as that post suggests. For DC < ~10.6, it actually gets easier with more players. For DC above that, it get harder. (Source: anydice.com/program/339ab, click 'At Least' setting). This means that DC5 checks get easier, and they should be easy so the heuristic of 5*N works. DC10 just barely changes, so 10*N also works. A new rule is need for DC15. 30% of the time 1d20 > 15, so that is our goal. Fortunately, Nd20 > 12*N about 30% of the time for 3 < N < 10. TLDR: Group DC is number of players * individual DC, except for DC15 is replaced with DC12. \$\endgroup\$
    – E Tam
    Commented Dec 20, 2023 at 5:39

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .