How D&D Next almost made knowledge count (and then backtracked)

Have you ever seen the Antiques Roadshow on television? Folks bring in curios from grandma’s attic, and then an expert explains the history of each piece and assesses the item’s value. If the real world worked by the rules of Dungeons & Dragons, the show could dispense with the experts. The producers could simply round up a panel of yahoos from the Jerry Springer audience in the next studio, show them an 19th century jardinière (a flower pot holder, for those of you who just whiffed your knowledge check), and let some knucklehead roll a 19 or 20. “The distinctive crosshatching shows a genuine example of 1890s, New Orleans Art Pottery by George Ohr,” he would say before asking another panelist to flash her boobs.

In every D&D game, this pattern repeats with each check that allows the whole party to participate.

Edwaert Collier - Vanitas - Still Life with Books and Manuscripts and a Skull

I did not write this post to gripe about lack of realism; I’ve praised unrealistic game mechanics before.

I want the game to reward players who invest in knowledge skills. Instead, the moment the dungeon master asks for something like a history check, everyone at the table jumps in to roll. More often that not, the player who invested in history rolls too low to determine the nature of the ancient battle standard, while some bozo with an intelligence-8 dump stat rolls a 19 and starts reciting the history of the old empire’s vanished legion. Once again, the party’s scholar feels like a chump for staying in school.

If everyone in the party can attempt a knowledge check, the five or so rolls ensure that someone in the party will luck into all but the most difficult checks. Why bother investing in knowledge skills? Someone will hit anyway.

You could bar party members without training in an area of knowledge from making checks. Third edition imposed such rules, but I favor the fourth edition approach of allowing everyone to participate, even if they stand little chance of success.

The August 2, 2013 playtest packet included a Lore rule that offered a solution: Characters who knew some field of lore gained a +10 to intelligence checks rather than the +5 advantage typical for similar checks in 4E. The +10 bonus reaches high enough to grant the scholar a significant boost over the rest of the party. The rule yields two advantages:

  • The party’s expert stands a better chance of making a knowledge check than the rest of the party.

  • Knowledge checks can be hard enough to reward knowledgeable players with information that would otherwise be out of the party’s reach.

Despite these advantages, the published version of D&D Next will probably omit Lore for two reasons:

  • The designers favor a simple scale of difficulty classes that applies at every level, throughout the system. When characters gain a +10 for lore rather than the small bonuses for skill proficiency, the difficulty of knowledge checks must be set higher than suggested by the universal DC scale. If I wrote the Dungeon Masters Guide, I would simply coach DMs to favor harder DCs for knowledge checks.

  • The designers seem enchanted with the notion of using as few types of bonuses as possible. I suspect they would see a different lore bonus as clutter, not worth its benefits. The final playtest packet aggressively pushed all skill bonuses into a single proficiency bonus, while eliminating lore from the game. See “Proficiency and bounded accuracy” for more.

Assuming the design goes as I expect, and knowledge skills deliver the same, small bonuses as other skills, I plan to run knowledge checks using the following procedure:

  1. Allow everyone in the party to make the knowledge check.

  2. Give the players some minimal amount on information based on the (probable) success of someone’s roll.

  3. Ask for the check results from anyone with the applicable knowledge skill.

  4. If any experts succeeded on their rolls, give deeper information.

This method rewards players who invest in knowledge skills with an advantage, even though the rules as written rarely offer a benefit.

20 thoughts on “How D&D Next almost made knowledge count (and then backtracked)

  1. Chris Shaeffer

    Hmm, I like your solution. It’s simple and works.

    Another solution (albeit more complex) might be giving lore-invested players:
    1) The proficiency bonus
    2) Advantage on the roll

    You could make the difference even more profound by having uninvested players roll at disadvantage.

    1. Chris Shaeffer

      Some quick number crunching: This means an invested player would roll an average of 15 to 20 (depending on their proficiency), and that a non-invested player would roll an average of 7.

        1. DM David Post author


          Thanks for your insights. I’m always happy when someone else shares enough of my fascination with the game’s nuts-and-bolts to comment (and to perform math too)!

          Advantage and disadvantage differ from mere bonuses and penalties in some interesting ways. Unlike a bonus, advantage can never guarantee success; unlike a penalty, disadvantage can never guarantee failure. I like the way the mechanic can inject a bit of uncertainty into the game.

          As for the knowledge checks, another aspect of the mechanic applies: Advantage grants you a higher chance of success, but advantage doesn’t allow you to reach a higher difficulty than otherwise possible.

          The Lore rule allowed a DM to set the knowledge check difficulty to something like DC 25. Then, a smart, trained first-level character with a +13 on a lore check enjoyed a fair chance of success, while an untrained no-nothing cannot luck into success. Without lore, even with advantage, that expert character cannot hit a DC 25 and anyone of average intelligence can luck into a DC 20 success.

          Advantage doesn’t solve the core problem: that a trained character cannot reach significantly higher DCs than a lucky untrained character. And if five character’s roll along with you, even with advantage, you have little chance of rolling better than the luckiest of the other four.

          High DCs put expert knowledge out of the reach of non-experts, and allow only experts—and parties that include experts—to benefit from obscure bits of knowledge.

    2. Chris

      This was going to be my suggestion as well. I like that it utilizes an existing mechanic that all players will be familiar with.

  2. Don Holt

    Speaking of mechanics, one of the things that I dislike about D&D is flat distribution of the D20 roll. Every result has a 5% chance of happening.

    What if you roll 5D4 instead of 1D20. The range of this distribution is roughly the same, but the variance is very different. A 20 only occurs once in a thousand times, rather than 1 in 20. This would make slight differences in bonuses more important, as more results would fail in the 11-13 range, and few in the 16-20.

    And on a five, you give the player certainty of something which was dead wrong.

  3. Chamel77

    In my home brew campaign the stats are 4 physical: strength, endurance, agility and dexterity and 4 mental: knowledge, fellowship, willpower and perception.

    Wisdom and intelligence are too similar, “wisdom” is just a good fantasy term. Intelligence rolls are usually made for knowledge, not figuring out complex problems (that’s what role playing is for).

  4. Groumy

    I just read your post and the one on the bounded acuracy, and I think that quote from the DM’s Guidelines can be the answer :

    “Take into account the ability score associated with the intended action. It’s easy for someone with a Strength score of 18 to flip over a table, though not easy for someone with a Strength score of 9.”

    What I mean, if is flipping a table is easy for a character with 18 in Strength and hard for a character with 9, they clearly suggest to use different Difficulty Class.

    I would go as saying that flipping a table is Trivial (DC 5) for an 18 STR guy, and Moderate (DC 15) for a guy with a 9 STR . That’s an automatic success for the 18 STR guy, but an interesting roll for the 9 STR guy.

    The same could be applied for Knowledges Profiencies, to know that a troll regen stop to works with fire or acid damage could be :

    – An Easy (DC 10) task for someone proficient in Intelligence (Nature)
    – A Moderate (DC 15) task for someone proficient in Intelligence (Arcana)
    – A Hard (DC 20) task for someone with no profiencies.

    Basically, I’m suggesting to use Profiencies as Skills in Numenera. If you have a Profienciy that would help you in knowlegde field, it drops the difficulty for you by one (or grant a +5, as you prefer).

    As Extraordinary Abilities too, bellow 10 step up the difficulty, 16 and greater step down the difficulty.

    That’s my two cents.

    1. DM David Post author

      Hi Groumy,
      You suggest a clever work around. I may use this method when I want an element of uncertainty, but less than a strict interpretation of the rules provides.

  5. Jason

    Ive been thinking about this a lot, and realize that since the early days of D&D this has been sort of a broken mechanic, where the 18 strength fighter fails to break down a door by rolling bad, but then then 8 strength wizard rolls a 20. Same thing with Knowledge Lore checks, where the expert rolls bad but the brain-damaged halfling recalls the name of a ancient king while he’s drunk. How to approach this issue within the confines of D&D Next? I would suggest clear cut auto success, auto failure and then a middle ground “maybe.” That 18 strength fighter doesn’t need to roll to break down the door. Its a given. The 8 strength mage doesn’t need to roll, it’s impossible. The 12-13 strength thief…well, maybe he has a 50/50 chance. Sort of makes me miss the Open doors/Bend Bars percentages of 1e.

    1. Groumy

      You know, in D&D Next/5th there is 7 degree of difficulty, ranging from Trivial (DC 5) to Nearly Impossible (DC 35) and the ability bonuses ranges from -1 to +5, giving 7 basic degree of success.

      With that in mind, there’s probably some relation between those two that could serve as both a required level (auto failure) and automatic success.

      So we could match a required ability mod to a degree of difficulty like that :

      Trivial (DC 5) : -1
      Easy (DC 10) : 0
      Moderate (DC 15) : +1
      Hard (DC 20) : +2
      Very Hard (DC 25) : +3
      Formidable (DC 30) : +4
      Nearly Impossible (DC 35) : +5

      And grand auto success if you ability beats the requirement by 4.

      But that don’t include proficiency bonus and the like, it is just food for thoughts 😉

      1. DM David Post author

        Hi Groumy,
        Thanks for the food for thought. I know the 5E designers sought to free the game from the need to reference numbers like these, favoring DM judgement instead. However, many players favor a higher degree of resolution transparency, where they have a clearer idea of how an action’s outcome will be resolved. A system such as this could be instituted at the table without any changes to the rules as written.


    2. DM David Post author

      One of the early playtest packets included a rule for automatic success similar to the one you propose, but the designers dropped it. They left the matter to the DM’s judgement. As the game stands, a character’s abilities and proficiency have so little influence on the outcome of a check that I think the DM should award success or failure for high and low scores. A DM should reserve rolled checks for cases when the odds amount to something like a coin flip.

      This goes back to the original game without checks, where a DM might improvise a die roll to impartially resolve cases where judgement fails.


  6. newb

    Here are three things I use to make skills/non-weapon profs. more meaningful:

    – Tell the players to go wild with skill checks. The main reason why people complain that skills are more useless than combat abilities is because they don’t get enough chances to use them. Make skill checks as often as possible, and don’t be afraid to hand out bonuses for clever uses of skill checks. Throwing around a few extra +2 bonuses isn’t a game changer, but it does make players feel like there is a value to training characters in skills, it adds more depth to encounters/scenarios, and it gives them the warm flush of success. And it gets player’s creative juices flowing which is always good.

    It also has the added benefit of making you a better DM: When a player makes a good roll on a knowledge check, the player expects you to provide some rare and insightful information, which means that you cannot just present bland, cookie cutter foes. Now your opponents need histories, culture, strategy, back up plans, subterfuge, special abilities and unconventional technologies. And more importantly, ones that matter. In short, it forces you to add more depth to your games.

    Part of the reason why skills often feel useless is because in most conventional scenarios they are. Most prepackaged scenarios only provide a handful of opportunities to use skills per session. You’re lucky if you might get one chance to showcase that one skill you spent so many points boosting (and then roll a 1 or a 2). If on the other hand players use skills/non-weapon profs. as many times per session as they do their combat abilities, even small differences in training bonuses make an impact in the long-run.

    – Present intelligent encounters. A reason why skills/non-weapon profs. usually don’t matter, is because often encounters are so dumbed down that any degree of intelligence is redundant. However, if you present intelligent encounters instead, those skills suddenly become essential to success. Don’t misunderstand me though – you needn’t push the bar so high that a few failed skill check will necessarily doom the party to failure. Instead, players should be made to know that simply plunging headfirst into the fray will cost them significant advantages, while approaching a battle intelligently will give them a major edge. As I often say, players don’t need any statistical advantage in the game – the fact that they’re several minds opposed to just one (the DM) already gives them a major advantage. You just have to make them use that advantage.

    There is often a stupid preconception that combat abilities are for battles, and skills/non-weaps are for everything else. Except there is no reason for that to be the case. If you present intelligent combat encounters, you may just find players relying on their skills more often than their combat abilities.

    For example:
    Take a typical precon scenario: Party finds a group of enemies, kills them, takes the loot. You can run it that way, or you can make it more interesting: local history check tells you that the enemies are in fact a notorious city gang and the fact that they’re out in force in the middle of nowhere means they’re likely waiting for to do a trade with gang 2, who are their regular business partners. Military history check tells you they’re using the standard equipage and deployment of such and such military organization, which also customarily uses hidden snipers to cover retreats and usually has a second band of reinforcements hidden somewhere in the wings. Arcane history check tells you…You get the picture.

    The standard combat encounter, presented intelligently, typically offers dozens of opportunities for awesome skill checks. Present a fully three dimensional combat terrain (something which I find tiles often inhibits) and suddenly physical skills like climbing, jumping, sneaking, toppling, squeezing, pushing dominate combat. Present intelligent opponents with inobvious abilities, back up plans, tactics and deception and misdirection, and suddenly making regular mental skill checks become crucial.

    You may argue that making more intelligent combat encounters demands a lot more preparation from the DM. I think that it demands a little more preparation, but not much. For instance, a simple trick I use is to make a list of skills/non-weapon profs. and then to make a list of five or so general purpose ‘useful discoveries’ that a PC might get from a successful check, such as:

    Military history: snipers; hidden reinforcements; suicide attackers; a rare and powerful item; flee and then ambush at a later stage

    and if a PC makes a good check (or if the don’t – as a penalty), I pick one that works best for the situation at hand and work it in. And as I come across new and interesting ideas, I simply update the list.

    – Finally, no simple success/failure. A key change I’ve made in my homebrew rules is the introduction of degree of success in rolls. This simple mechanic works by altering the consequences of successes or failures by how far above or below the DC they are:

    DC +0 – +3 Partial success (success with a disadvantage)
    DC +4 – +7 Standard success
    DC +8 – +11 Exceptional success (success with an extra advantage)
    DC +12 or more Critical success (success with a major extra advantage)

    And the same with failures:

    DC -1 – -4 Partial failure (failure with an advantage (usually bonus on retry))

    …you get the idea.

    It might be argued that this mechanic can slow down gameplay considerably, but the point is not to use this with every skill check, just those you really want to showcase or the cliffhanger situation type.

    The benefit is that this mechanic really allows those small training bonuses to shine. And what seems to work even better is only to add the training bonus afterwards. (You rolled a 15 on a DC 10 – that’s a standard success. But wait, add your +5 training bonus and now you’ve got a critical success. Wowza!) It might be argued that it still doesn’t stop you rolling a 1 or a 2 in spite of your training bonus, but it does allow training to mitigate the damage of failures, and it does reward players for those extra high rolls which training bonuses make possible.

    1. DM David Post author

      Hi newb,

      You present a lot of good ideas here. Thanks for taking the time to post this! I really like your tactic of using the characters’ knowledge skills, representing of the game world, to bring useful background and information into play, even in a combat. This approach helps bring both the characters and the game world to life. I sometimes grumble that skills force players out of the game world by making the details of the imaginary world unimportant, but your strategy has the opposite effect.


      1. newb

        You say you like these ideas, but are you going to use them?

        I probably sound like an a-hole, but I’ve gotten all to accustomed to people either saying ‘yes, that’s nice dearie’ and then forgetting everything I said, or just telling me straight that I’m an idiot, whenever I pitch any new or vaguely innovative idea. Roleplayers in my experience are such a bunch of old maids whenever it comes to anything new or different from the same old stale fare they’re used to.

        I can remember enthusiastically pitching the idea of bounded accuracy back in the 90s (pretty much after my first time DMing high-level PCs and finding it to be completely impossible). The only responses I ever got were either non-committed or openly hostile. Now, 15 years on down the line, wizards pitches exactly the same idea and everyone thinks its the greatest thing to come out since velcro.

        You certainly don’t deserve my spleen, but you’ve got to agree that the typical conservatism of most roleplayers is simply ridiculous.

        1. DM David Post author

          Hi newb,
          I absolutely intend to add your method for using knowledge checks in combat to my bag of tricks.

          I suspect that most game masters weigh the amount a check exceeds a DC as a measure of success. As for using a scale of specific numbers, I don’t think they add enough to make me want to adopt the scale. The degree-of-success numbers substitute the DM’s judgement of how much better, say, a 2-above check is versus a 6-above check, with the DM’s judgement of what a partial success means versus a standard success.

          I like bounded accuracy too, but lots of folks will tell you that it will doom 5E. Check out the comments at the end of my post on proficiency and bounded accuracy in D&D Next.

          Again, thanks for your thoughts!


  7. Pingback: Smarter Intelligence Checks | Ludus Ludorum

  8. miklos

    Another solution is proficent players reroll 2,3,4,5 and keep the 2nd rolll….so nonproficent players reroll 16,17,18,19 and keep the 2nd roll

  9. Clay

    This brings to mind the fact that Intelligence is the least important skill in 5e (unless it is your spell casting skill). Here’s a thought: give the PCs Expertise (double proficiency) in one domain of knowledge, plus one domain for each point of Intelligence bonus (i.e. Intelligence 18 would get 5 knowledge domains). This would suggest having areas of knowledge as in 3e (Geography, Nobility, Local, etc.) not just Arcana, History, Nature, and Religion as in 5e.

    Too fiddly?


Leave a Reply