Let’s have a history lesson.

A long, long time ago, I was axed from a podcast I’d helped create by my a&$hole of a partner. On Thanksgiving weekend. Right before we interviewed one of my game design heroes. Bitter and spiteful, I decided to start my own podcast. But that was hard. So I started a blog instead. And after posting a series of articles about how to make solo monsters in D&D 4E more like the bosses in God of War III, which I’d just finished on my brand new Playstation 3—I said this was a long time ago—I had no idea what else to put on my blog. But then, one of my friends in the gaming community on Twitter—this was back when Twitter was still fun and the gaming community wasn’t spending all of its time screaming about imaginary racism and hating everyone in the name of tolerance—one of my friends claimed that D&D’s core d20 mechanic with its binary outcomes and limited skills made it impossible to run a good mystery game. I decided my next article would prove him wrong.

But I couldn’t prove him wrong. I mean, I could. But only on a technicality. And being right on a technicality is only a win if you’re an a$&hole. And this was so long ago, I wasn’t even sure if I was an a$&hole yet. See, I proved that it wasn’t the binary nature of skill checks or the d20 mechanic f$&%ing up good mysteries and investigations. It was actually the way D&D presented itself. And it was the way GMs had been taught to think. I realized that if I could just teach GMs how to think properly, they’d be able to unlock the full potential of the system. The end result was an article called 5 Simple Rules for Dating My Teenaged Skill System. And man did people eat that s$&% up.

And that’s how TheAngryGM.com was born. That’s why I started the site and that’s why almost all of my articles are either about fixing RPG systems or about fixing GM’s brains. And because I’m really good at breaking things down and understanding how things work and seeing things with a beginner’s mind and then explaining all of that s$&%, well, I’m still here. Meanwhile, no one even remembers that dumb podcast that did well for ten episodes and then faded into obscurity when one of the hosts mysteriously vanished and the remaining host ran it into the ground.

Why do I bring up all this ancient history bulls$&%? Well, it’s funny, but I never did go back and write the article I originally intended to write. The one about how you can run good mysteries and investigations in D&D. Now, I wouldn’t give a s$&%, but D&D really hasn’t improved on that front at all. In fact, it’s gotten worse. And meanwhile, I had to write some investigation encounters for this module I’m working on. And that means I’ve been dealing with the same bulls$&% that started this all.

Playing at Investigation

This article’s about action resolution. Again. I have to make that clear because once you start reading, you’re probably going to get confused by all my talk about running mystery adventures and investigation encounters. See, I don’t want you to think that all the s$&% I’m about to say only pertains to running mysteries. Nope. It actually pertains to how D&D uses its proficiency rules wrong and how you can use them right.

Now, I’ve written about some of this s$&% before. That can’t be helped. You write about the same role-playing game system for a dozen f$%&ing years and see if you don’t start repeating yourself. But I’m pretty sure I’ve never looked at this particular problem from this particular angle and made this particular case this particularly strongly. If I’m wrong about that and this really is just another f$&%ing rerun of s$&% that I’ve already written, then just add it to the pile of evidence I’ve amassed that the WotC designers know how to write really good mechanics and have no idea how the hell to explain them or use them.

Beyond that, maybe it’ll just help you run better mystery adventures and investigation encounters.

The story starts a few weeks ago when I was writing a couple of investigation encounters for my upcoming module The Fall of Silverpine Watch. And you’ve probably noticed I’m speaking very categorically here. I’ve become really f$&%ing careful about my language. I no longer use the words ‘scene’ and ‘encounter’ as synonyms, for example. And I have a growing list of descriptors for both scenes and encounters. I’m introducing a new one to the lexicon today. The investigation encounter.

An investigation encounter is one where the players get the chance to learn useful information by interacting with different encounter elements. Unlike, say, a combat encounter or a social encounter, an investigation encounter usually lacks a central conflict. It has a dramatic question though. Usually, “what information do the players learn from these clues?” This s$&%—the open-ended dramatic question and the lack of conflict—makes investigation encounters hard to build and run. And they’re paced by the players. They last until the players decide there’s nothing else to learn or give up trying to find anything.

Now, there may come a day when I write an article about investigation encounters. But today is not that day. Today, I’m writing about what problems investigation encounters reveal in D&D’s action resolution system and proficiency rules. And how to solve those problems and run better overall games.

Inviting the Players to Ask Questions

GMs suck at what I call ‘inviting the players to act.’ In my book—and all over this website too— I’ve explained that running encounters—and running entire f$%&ing games—runs thusly:

1. The GM sets the scene
2. The GM invites the players to act
3. The players describe their characters’ actions
4. The GM resolves the characters’ actions and describes the results
5. Repeat the steps above as necessary until the game is over

That second step’s easy enough when there’s a central conflict in the encounter. Something that’s directly opposing the characters. Nothing invites action like an ogre trying to kill you or a ravine between you and a pile of treasure. But when all you’ve got is a scene to poke around in trying to answer questions you don’t even know you’re asking, you don’t exactly feel pulled into the action.

And that’s why investigation encounters usually take the form of ‘the mysterious thing.’ The players are just walking along, minding their own business, and suddenly there’s a thing in their path. A thing that implies a mystery. The classic example is, of course, the dead body in the road. Even if the players’ current goal doesn’t involve figuring out what’s turning so many people into corpses around here, the mysterious corpse in the road inspires the players to ask some questions. Even they’re just variations on the theme of, “what did this, and are we next?”

Thing is, an apparently empty room can contain just as much information as a mysterious corpse. A careful examination of the floor, for example, might reveal the telltale scratches made by a large, clawed creature’s recent passage. Or it might turn up a few barbed, iron-tipped darts that suggest a nearby trap. But apparently empty rooms don’t draw the party in the way corpses and runed pillars and strange monuments do. Those things are invitations. “Hey,” shouts the corpse, “check me out! I probably have treasure in my pockets and useful secrets to share!” Not literally, of course. If the corpse literally says anything, then the party’s probably about to have a combat encounter, not an investigation encounter. Or at least, a very unpleasant social encounter.

Investigation encounters are passive. They imply questions and then give the players the chance to discover the answers. And, to make them work, most involve an interactable element that, when interacted at, yields useful information.

Now, that’s not the only way to structure an investigation encounter—or a bigger, broader investigation scene—but it is the easiest and most obvious. It’s the one that works best. But I’m not writing about that s$&% today. I’m writing about something else. So, let’s consider…

The Corpse with Two Secrets

Pop quiz, hotshot: there’s a thing in an encounter that can provide different answers to different people. What do you do?

What do you do?

Let me give you a ‘for instance.’ Say the heroes are traveling a road through a dangerous forest and they find a standard ‘body in the road.’ They decide to examine the body. Everyone can see how the corpse is dressed and equipped. And that its clothing is torn. And anyone examining the body closely can see that the corpse is unnaturally pale. And that the body’s uninjured. A character who’s exceptionally deductive, though—not just perceptive—might guess from the man’s clothing that he’s a merchant and note that, given how far he is from town, he doesn’t seem to be carrying enough supplies to make the trip. The medically trained character might note that the body’s not just pale, it’s bloodless. There’s no bruising or pooling. And might also find, with a careful examination, the nearly invisible puncture wounds on his legs and back. The survivalist might note that the body is being shunned by scavengers; that there’s signs that natural beasts have been by recently, but that they passed the body up.

See the problem? No? Let me lay it out for you. The problem lies in some of the assumptions D&D is designed around. Or at least presented around. You won’t see this s$&% explicitly spelled out. Not clearly. You’ll find bits and pieces mentioned in passing here and there. Mostly, though, this s$&%’s just sort of implied.

First, when it comes to the PCs investigating things, examining things, or asking questions about things, D&D wants you—the GM—to cast the widest f$&%ing net you can. While the rules in the book suggest that you dole out bonuses when the players are specific in their actions and questions, they don’t want you withholding ancillary information a character might stumble upon while looking for something specific. I’ll call this the ‘interaction button’ problem.

See, a long time ago, when you played a video game RPG on a console—like Dragon Warrior I, the first real attempt to fit a computer RPG experience in a console cartridge—when you played console-based RPGs, you had to be specific in your interactions. When you wanted to open a door, you had to select that specific option from a menu. And when you wanted to walk down a flight of stairs, you had to pick the option for that. Talk to an NPC? Select ‘talk’. Open a chest or search a pot? Choose ‘search’. But eventually, designers realized that you’d never talk to a flight of stairs or open a person. So they could just pick one interaction for each object and then have one button to trigger whatever that object’s interaction was. Thus was born the context-sensitive action button.

The D&D designers eventually realized similar tedium existed in D&D. When the players came to a door, for example, you ended up having to make a whole bunch of die rolls before the paranoid players who’d taken one too many traps to the face would actually walk through it. Roll to listen at the door. Roll to search the door for traps. And so on. Same with chests and 10-foot-sections-of-apparently-empty-rooms. So, by the time 4E rolled around, they codified a context-sensitive action button of sorts by establishing that all you had to do was specify you were searching and roll a Perception check, and then, if you were successful, you’d turn up everything there was to find. Just say you’re searching the room to spot all the traps, find all the secret doors, find the key hidden under the carpet, everything. D&D 5E attempted to roll this back a smidge with a sidebar in the PHB about making the players specify where they’re searching and rewarding them for that specificity, but the damage was done.

Second, D&D 5E has mostly done away with requiring that characters have specific proficiencies before they can take certain actions or gain certain bits of information. One major exception explicitly noted in the rules is that a character must have proficiency with thieves’ tools to pick locks and sabotage traps. But, by the rules, it’s up to the GM’s discretion whether you need to know how to actually use blacksmith’s tools to forge a greatsword. Generally speaking in D&D, nothing’s supposed to be locked away behind the need for specific training.

Now do you see the problem I’m talking about? Mechanically speaking, how do you deal with the body I described above. The one that has different kinds of information for different experts. How do you handle that in a game that assumes there’s a single interaction button and no actions can be restricted based solely on someone’s expertise in certain areas? How can you handle anything the party can interact with that can provide more than one clue depending on who’s examining it?

And before you all lose your s$%&, I’m not saying D&D’s assumptions are bad assumptions. I know why things changed. I’m all for removing the bulls$&% with the players making seventeen different Search checks just to walk through a door. And no, the reason was not that the majority of GMs were just waiting for the players to forget one of the seventeen checks so they could say, “you didn’t specifically say you were examining the hinge pins and the tiny piece of flint affixed to the lower pin strikes a spark as you open the door and ignites the odorless, colorless, tasteless, invisible explosive gas filling the hallway. Roll 1d100 to see how many fragments you’re blasted into.” That s$&% seldom happened. At the very least, it was far less common than it is for whiny, bratty children to complain about how much it happened in the old days of gaming that they hadn’t even been alive for.

But those assumptions have consequences. And the consequence here is that any examinable, investigable thing in the game—be it a door or a chest or a corpse or a 10-foot-by-10-foot section of wall or an entire room—must require one—und precisely vun—skill check and must reveal one—und precisely vun—infodump on a success. Without a big ole wall of text, you can’t have a corpse that tells medical experts one thing and nature experts another. And you sure as hell can’t require them to make different ability checks with different DCs.

“So f$&%ing what,” you might say. This is a really small, really esoteric problem. Maybe. Maybe it’s just the kind of thing that makes it hard to build interesting mysteries. Or interesting investigations. Maybe it’s just the kind of thing that you have to work around every time you want to reveal backstory information through environmental details or historical artifacts.

But it’s bigger than that. Because it downplays both what the characters know and what questions the players ask. It minimizes the players’ roles in solving puzzles. And it marginalizes characters in the party. How? Well, say the wizard investigates the body. He’s got the best Intelligence score in the party. The wizard’s player rolls a solid 18 and he gets the infodump. There’s no reason for him to turn to the healer and say, “what do you make of these wounds?” There’s no reason for the ranger to say, “you know what’s odd though…?” No one has to get involved unless the character with the best Intelligence score rolls a crappy check. And when the wizard does blow his roll, the best strategy is to invite the next best Intelligence score to take a crack at it. Every character just boils down to an ability check. Their skills don’t matter. In precisely the sort of situation when individual skills should be the most important thing.

What I’m saying is that the scene you see in every procedural crime drama where different experts all examine the same thing and each provides different information based on their skills, that scene is a scene that should happen in an investigation encounter or mystery adventure. It showcases how each character is different and drags multiple different characters into the action.

So how can you have that scene? How can you bring the players’ brains back into the game and spotlight the individual characters’ skills? Well, you can’t. Not unless you’re ready to reject those assumptions D&D is built around. But is that a surprising answer? It shouldn’t be. Everyone knows D&D is a really crappy system for running mysteries and investigations.

Fix #1: GMs Pick Abilities; Players Propose Proficiencies

I’ve said before that GMs need to recognize that skill checks aren’t a thing in D&D. There’s only ability checks to which the players get to add proficiency bonuses in the event they have relevant training. And I stand by that. When a player proposes an action, the GM should only concern himself with figuring out which ability score governs the described action and then ask for a check. So, when the hulking fighter throws the merchant to the ground, then picks him up and shoves him against the wall, the GM should ask for a Strength check to determine whether the merchant confesses his crimes.

It should then be on the player to say, “hey, I have proficiency with Intimidation. Can I add my proficiency bonus?” And, of course, the GM would say “yes, absolutely” because this act of physical intimidation definitely falls under the purview of the Intimidation proficiency by virtue of that being the f$&%ing name of the skill. Of course, it’s a lot easier to use this approach if you use character sheets that don’t tie every f$&%ing skill to one—und precisely vun—ability score. Like these gorgeous sheets my friend and supporter Alyssa made.

This approach removes the ‘interaction button’ and restores the menu-driven interaction system of old. The one that required you to think about what you were interacting with and how you were interacting with it. And decide which interaction was likely to yield the result you wanted. It also provides a language for talking about those interactions even if players and GMs lack the expertise necessary to understand how a medic might examine a corpse and how that might differ from a survivalist’s approach. Because guess what, different professionals look for different things in different ways.

In other words, when a player examines a body, they might not know how to say that their character is trying to obtain medical information from the body, but they sure as hell know their character is trained in medicine and they suspect the corpse has some medical information in it. Because, you know, it’s a dead body and it used to be a living body, and medicine deals with how one turns into the other. Or how to keep one from turning into the other. And this can be really useful with the weird-a$& Investigation proficiency. I mean, consider the difference between a burglar ransacking a room looking for loot and Sherlock Holmes examining the room to figure out who burgled it and what they burgled from it.

This actually adds a puzzle-solving element to investigation encounters. Just as the character has to search for the right answers the right way, the player has to recognize which of their skills is likely to provide them with the information they’re looking for. They have to think about the questions they have and how they’re likely to be answered. See, by default, investigating locations and objects boils down to saying, “I investigate the location or object” and then rolling a die roll to see if you investigated it good enough. There’s nothing interesting there. The GM just pointed to a thing and said, “there’s an interesting thing that’s probably full of clues” and the player said, “well, I check the thing for clues,” and then a plastic polyhedron said, “okay, he checked it good enough, give him the clues.” That’s not f$&%ing gameplay.

Of course, it might not occur to a player that one of their skills might be relevant. The ranger’s player might never consider checking into the scavenger activity around a body. And you probably see that as a flaw. But it ain’t. It’s actually a strength. Because it means that what the players think and perceive, what they guess, and what they figure out, that s$&% changes the outcome. More than the dice. And it gives the player something to learn through gameplay. The players can actually get better at using their skills as they keep playing. As they experiment with different skills in different situations.

Moreover, when a player suggests a particular skill might be relevant, it pushes the GM to improvise. To say, “you know what, I didn’t think that Nature would be useful here, but it makes sense, so what information can I offer?” And that also helps the GM get better at building interaction encounters and mystery adventures. The next time there’s a corpse to be found, that GM’s going to say to himself, “last time the party found a body in the woods, the ranger examined the area around the body with his Nature skill. I’d better make sure I have some answers for him ready this time.”

Beyond that, if you’re a savvy GM and you’re concerned that your players might not recognize a particular skill might be relevant, you can plant clues in your narration. You can describe the body as laying, “exposed to the natural elements, but so far untouched.” That weird turn of phrase might start the slow-a$& hamster wheel in the ranger player’s brain a-spinnin’. Subtle clues in the narrative are basically the TTRPG equivalent of using graphical and environmental details to guide a player in a video game. It’s the equivalent of Dark Souls using the environmental lighting to help you find the ladders in Blighttown or Horizon: Zero Dawn clumsily smearing all of the ledges with yellow dust so you can see where you can climb. Yeah, it’s a pretty advanced GMing trick, but running good mysteries is pretty advanced GMing. Not something every GM is up to.

Does this mean you—the GM—should never ask for a roll with a specific proficiency? Does it mean you should never suggest a proficiency might be useful? No. While you should default to letting the players decide what skills they want to apply to what situations, there’s always going to be some situations when you should prompt the players. If the players don’t know what they’re actually rolling for or if you’re rolling secret rolls behind your screen, for example. Though if you’re properly using passive scores, that shouldn’t be an issue. And when there’s no ambiguity or no element of mystery or puzzle-solving and when there’s a huge gulf between what the players know of a situation and what the characters in that situation would perceive, then you should prompt the players. Like, say you’ve got some mysterious runes carved on a pillar. Without the knowledge that their characters possess, it’d be impossible for the players to discern whether the runes are part of an arcane formula or whether they’re religious ideographs or whether they’re just the letters of an ancient alphabet. Any player who asked to use Religion or Arcana would just be guessing. Guessing isn’t interesting. So that’s when you say, “make an Intelligence check and, if you’re proficient with Religion, you can add your proficiency bonus.”

Except you wouldn’t necessarily say precisely that. Because…

Fix 2: Make Proficiency Matter Again

The above fix gets rid of the ‘interaction button’ and gives the players a way to define what questions they’re looking for answers to and how they’re looking. If they don’t recognize a skill might be appropriate in an investigative challenge—meaning the character just failed to see the pattern or put the pieces together—that’s akin to the player failing to solve any puzzle in the game. And since puzzle-solving is pretty much the central mechanic in a mystery game, that’s what you want. You don’t want to boil everything down to die rolling.

But mysteries are also skill-based challenges. It’s important that the players put the pieces together and that they ask the right questions. But what the characters know—the characters’ skills, experiences, and backgrounds—are also important. Investigation encounters really should shine a bright f$&%ing spotlight on each character’s particular skillset. You know, “thank God I dropped out of seminary to be a detective or we’d never have caught this Seven Deadly Sins killer.” S$&% like that.

Honestly, given how important roles, skills, and archetypes are in D&D’s character generation, it’s actually pretty f$&%ed up how D&D marginalizes all those choices by assuming everyone can do everything and experts just get a slightly better bonus. Fortunately, it’s easy to fix. Just reverse the assumption. Unless there’s a really good f$&%ing reason to think otherwise, assume only someone with training can do a thing. Everyone can try to sneak around, sure, but only someone with actual Stealth training can effectively instruct an untrained party on how to lay a good ambush. Anyone can climb a tree or a rope or a ladder, but only someone with Athletics training can free-climb a bare rockface or a castle wall. Anyone can spot the cause of death if the head’s no longer attached to the body, but it takes someone who knows their medicine to recognize any more mysterious cause of death than that.

Just gate s$&% behind proficiencies. Don’t be afraid. It’s okay. Let me tell you a secret. There’s not that many parties that are actually missing any vital skills. D&D’s skill list just ain’t that long. And if you let players talk things out before they start playing, you’ll discover they naturally cover all their skill bases anyway. Partly because, well, players like to cover their bases and partly because players like to feel special. They do their damndest not to duplicate skills. Which means you’ll always one dude who’s the Persuasion dude and one dudette who’s the Survival dudette and you’ll have a Religion chump and so on. Trust me. There’s eighteen skills in D&D 5E. Nineteen if you count thieves’ tools as a vital skill. Four players can cover almost all of them.

And while a player might get annoyed because you won’t let their meathead climb a 30-foot wall without tools, fixtures, or training, the guy who spent a precious proficiency on Athletics is actually going to feel pretty damned special. He can shoulder his way to the front of the group and say, “here, let me climb that and I’ll drop a rope down for the rest of you.” As it should be.

And now, when you design a mysterious thing, you can give it some info that any schmuck with two working eyeballs and a decent die roll can discover. And under that, you can assign specific clues for specific experts. Separate the s$&% that’s common knowledge from the specialized information. Anyone can see that the body’s strangely pale and that there’s no obvious sign of injury. But only the mediciner can categorically say the body’s been drained of blood and recognize the punctures through which the strange exsanguination happened. Only the survivalist can recognize that the body’s not just undisturbed, it’s being shunned by scavengers. And only the investigator can deduce, based on the man’s clothes and appearance, that he was a traveling peddler, and given the distance from town, he doesn’t have enough supplies on him to make the trip. Which means the party should be looking for his camp. Or his lost mule. Or his stolen backpack.

And when someone examines the strange runes, you don’t say, “make an Intelligence check and add your proficiency bonus if you have Religion.” You say, “are you proficient in Religion? No? Then piss off.” And that enables the cleric to step forward and say, “hey, Alice, what did you find; let me have a look.” Is it a little metagamey that the cleric only stepped forward to see what the wizard was looking at because the GM asked about the Religion proficiency? Who gives a single, solitary f$&%! A little metagamey is allowed. An RPG is supposed to be a fun game set in a consistent but ultimately fictional world; not an excuse to yell at your friends about simulation theory.

My Final Word

In the end, my point is that investigations and mysteries and exploration and discovery and every other f$&%ing mental challenge should be first-and-foremost about what questions the players ask, what conclusions the players draw, and what skills the players chose to put on their character sheets. D&D’s assumptions have taken all of those interesting things and jettisoned them into the Abyss so that all that’s left is die rolling. Die rolling isn’t interesting. It’s a completely random event that no one controls. When you plant a thing in the game world and you point at it and a player examines it and you ask them to roll a specific number on a die before you give them any information, the player’s choices and actions just don’t f$&%ing matter. You could send the player home. Even the choice to examine the thing wasn’t really a choice because you pointed out the thing. The act of narrating the existence of the thing is just telling the players to interact with the thing. They’re just moving from breadcrumb to breadcrumb, rolling dice at your command. And all they can do is try to be the best damned die rollers they can be. There is literally no way anyone can build a compelling mystery with that setup.

More important though is that the game’s designers—and the idiots who are going to argue with me about this s$&%—have forgotten something very important. An RPG is at its most interesting and most engaging and most challenging when the players can’t just roll dice to overcome a challenge. When everyone in the party just takes turns rolling a check to climb a wall and then everyone climbs the wall, that’s boring as f$&%. That’s watching people play Yahtzee. But when the players need to be on the other side of the wall and no one can just roll a die to be on the other side of the wall, that’s when the game gets good. That’s when the players start improvising ladders or grappling hooks. That’s when the barbarian asks how high he can throw a halfling.

And what kind of GM would rather say, “okay, good Strength checks everyone, you won the wall” when they can say, “okay, the halfling sails right over the wall and takes… let’s see… 5d6 falling damage. And if you survive that, roll initiative, because the guard dogs are really hungry.”