Balancing Player Skill and Character Skill

I would like to apologize for my recent lack of activity.  I started a new job recently, and it has cut significantly into my free time.  I think I will soon have my life back in order, and resume my previous posting schedule.  With that in mind, I feel this topic is below my usual standard.  It’s better than nothing.

Two weeks ago I described how character skills and player skills differ and contribute during different kinds play.

The biggest problem for incorporating the player skills is that the GM must take on a much bigger role in arbitrating what is “reasonable.”  In social scenes, the GM gets to decide what is a persuasive argument.  In puzzles, a GM gets to decide what is a straightforward solution, what is a creative solution, and what is just too strange to work.  In combat, the GM gets to decide what situational modifiers apply, and in many cases needs to make silly judgement calls about whether (for example) a “typical” lantern is an equivalent amount of light to a candle or a 40 watt lightbulb.

Do not immediately become too concerned about balance.  For me, the goal of balance is to ensure that all players can participate in every scene.  I believe it is just plain poor game design to make a character’s abilities just as useful as a player’s abilities, or to aim to make each character as effective at the same thing as each other.

Let us look at a simple way to encourage players to use both their own skills and their character”s skills.  This will be with a differential reward system (the name is harder to understand than the system).

Differential Reward System:

The differential reward system begins with the assumption that the players are rewarded for every scene, regardless of the type of challenge they face.  With that in mind, begin tracking the types of challenges each player overcome, and whether they were overcome with player skill or character skill.  Every time a player overcomes a challenge with the related player skill, the reward for overcoming that challenge with a character skill becomes larger.  Every time a player overcomes a challenge with the related character skill, the reward for overcoming that challenge with a player skill becomes larger.  Each player will need to be tracked separately, or this won’t work.

Since the way player skill is integrated into combat scenes is strongly integrated into most RPGs combat systems, this won’t work for combat.  The same is also true of integrating executive thinking.  I’ll let these topics percolate a bit, and see if anything comes of it.

Sorcerers vs. Fighters in 3.x DnD and Pathfinder

I have had enough of people telling me that fighters in 3.x are useless, especially with the many changes in pathfinder that further refine balance from 3.5. I’m going to compare a very simplistic fighter build and a very simplistic sorcerer build. Both of them are built for damage dealing, and I’ll compare how much damage they do per turn. I already know that from levels 1-5 the fighter does more damage, hands down. 6-10 it’s about the same, unless the party funnels a little bit of extra money into the fighters weapon. Meaningful tactical differences are that the mage does more damage whenever they can hit at least 3 targets, about the same if the mage can hit 2 targets, and less if they can hit only one.

What I’ll prove is that the fighter does more damage when taking a full attack action, and less damage when only using a standard action to attack. As such, at high levels the fighter remains highly immobile but highly damaging, and the mage is powerful at range and still has the option of a move action. This tends to fit with a common perception of the fighter as a very defensive character, as their roll will be to rapidly kill anything that gets close enough to the party to hurt the spellcasters. However, the way they fulfill this defensive roll is by being really good at doing a lot of damage, while also being too slow to effectively attack enemies that are not yet within 10 feet.

To be clear, there are many other ways to play spellcasters and fighters than these. I consider many of the other ways far more fun.

Sorcerer Build

As everyone who likes to play spellcasters knows, the spell school to use for dealing damage is evocation. Here are the spells, organized by sorcerer level (not spell level):

Sorcerer Level Spell
12 Chain Lightning: 12d6, ref half.
14 Delayed Blast Fireball. 14D6, reflex half. HOWEVER, Empowered Cone of Cold with Heighten Spell feat is better: 21d6, ref half.
16 Empowered, Heightened Chain Lightning: 25d6, reflex half.
18 Mazimized and Heightened Chain Lightning: 108 damage, reflex half. It’s better than meteor swarm for damage, but not area of effect.

The Sorcerer will require particular feats to benefit this spell list: Spell Focus (Evocation), Greater Spell Focus (Evocation). empowered spell, heightened spell, and (before level 18) maximized spell.

High intelligence helps increase save DC, which means enemies are less likely to reduce the damage I will thus assume the character started the game with 20 intelligence, and put every stat increase into intelligence.

Finally, the spellcaster will need an intelligence boosting hat. Headband of vast intelligence +6 is easily affordable by level 12, and the other mental stat boosting hats won’t increase intelligence more than that.

Sorcerer Damage Per Turn

All of these sells are reflex saves for half, and use six sided dice for damage. I’ll treat each dice as 3.5 points of damage, and then use half of that value for a failed save. Then I’ll multiply the value of a failed save by the percentage chance of succeeding at the saving throw, add it to the value of half the damage multiplied by the percentage chance of failing the saving throw. That will give us the most likely amount of damage per turn.

To determine the enemies save bonuses, I just used the average of all monsters in the Bestiary (the first one) of that CR. Then I used an average. I think this is the part where a small sample size will introduce the greatest possibility of error. Especially given the results of level 18, which make no sense. Of course, all the level 18 monsters are absolutely gigantic, and large monsters have low dex and more HP to compensate.


Reflex Saving Throws


Rounded Average


10, 9, 7, 8, 5, 13




13, 15, 9, 10, 12, 9




13, 13, 17, 11




13, 13, 12



Sorcerer Level

Save DC




(42*0.85)+ (21*0.15)= 36 damage



(49*0.8)+(24.5*0.2) =44 damage



(87.5*0.75)+(43.75*0.25) =76.56 damage



(108*0.85)+(54*0.15) =99.9 damage


A fighter build intended to maximize damage will require feats to maximize their attack rolls, and maximized strength (ie. 20). At high level, equipment will matter a lot. Of course, normally a fighter with maximized strength would want to use a 2-handed weapon. I’m not going to do this, in part just to drive my point home, and in part because the barbarian’s non-damage related powers are (in my opinion) even better with 2 handed weapons. This is about fighters, so I’m not worried about that.

The required feats are:

Weapon Focus (lvl 1)

Greater Weapon Focus (lvl 8)

Weapon Specialization (lvl 4)

Greater Weapon Specialization (lvl 12)

Critical Focus (lvl 9)

It is very important to remember that the fighter gets Weapon Training. This will provide +2 to attack and damage during the fighters periods at levels 12, +3 to attack and damage during levels 14 and 16, and +4 during level 18.

I’m going to place about 25% of the fighters wealth or less into their weapon. Many people would want to place more. I’ll also buy the belt of Gaint Strength +6 with all of these builds, which costs some 36 000. It is a substantial investment at level 12, but still only amounts to about a third of the character’s wealth at that point.

PC Level Wealth Weapon
12 108 000 Longsword +4
14 185 000 Keen Longsword +4
16 315 000 Keen Longsword +5
18 530 000 Keen Longsword +5 of Flaming Burst

Fighter Damage per Turn

Calculating the fighter’s damage per turn is a lot more work than the sorcerers. First, I’ll calculate the damage per hit, discounting critical hits, and multiply by the liklyhood of it happening for each hit. Then I’ll calculate how much more damage occurs on a confirmed critical hit, and multiply that by the likelyhood of it happening on each of the individual attacks. Then add it all together.

I will use the average AC across all monsters in the bestiary of the same CR, just as I did for determining reflex save values. I’ll treat each d6 as 3.5, each d8 as 4.5, and each d10 as 5.5.




Rounded Average


29, 27, 23, 26, 27, 25




29, 29, 29, 29, 29 27




38, 38, 35, 32




37, 37, 32



Likelihood to hit

If you are critiquing my numbers, don’t forget to count the +4 to confirm critical hits from critical focus.

PC Level Attack Bonus Base + str + feat + enhancement + weapon training) Chance to Hit With Each Attack Chance for Additional Critical Damage with Each Attack (chance of falling in threat range * chance of hitting again)
12 12+9+2+4+2= +29 100% / 95% / 70% 10% / 10% / 9%
14 14+9+2+4+3= +32 100% / 95% / 70% 20% / 20% / 18%
16 16+10+2+5+3= +36 100% / 75%/ 50% / 25% 20% / 19% /14% / 9%
18 18+10+2+5+4= +39 100% / 100% / 75% / 50% 20% / 20% / 19% / 14%


Now for damage:

PC Level Damage per Hit (base + str + feat + enhancement + weapon training + elemental damage) Number of Hits

(sum of chances to hit)

Damage, not including critical hits
12 4.5+9+4+4+2= 23.5 damage 2.65 62.28
14 4.5+9+4+4+3= 24.5 damage 2.65 64.93
16 4.5+9+4+5+3= 25.5 damage 2.5 63.75
18 4.5+9+4+5+4+3.5=30 damage 3.25 97.5


Please note that, before, even counting critical hits, the fighter does more damage until level 16, and at level 18 does only 2.4 points of damage less.


Now, we do damage for critical hits, using the same technique. Because I’m only doing the additional damage from critical hits, it’s okay that the threat rolls overlap with the rolls that hit for normal damage. Since the d10 damage from the elemental burst weapon replaces the d6 damage from the normal damage roll, I included it as the difference between the average d10 result and d6 result: 2.


PC Level Damage per Crit (base + str + feat + enhancement + weapon training + elemental damage) Sum of Critical Hits Additional Crit Damage
12 4.5+9+4+4+2= 23.5 damage 0.29 6.82
14 4.5+9+4+4+3= 24.5 damage 0.58 14.21
16 4.5+9+4+5+3= 25.5 damage 0.62 15.81
18 4.5+9+4+5+4+2=28.5 damage 0.73 20.81


Now we add all the damage up:

PC Level Normal Damage / full attack Crit Damage / full Attack Total Damage / full attack
12 62.28 6.82 69.1
14 64.93 14.21 79.14
16 63.75 15.81 79.31
18 97.5 20.81 118.56


At most levels, the fighter can do much more damage to a single target than a sorcerer can. The notable exception is at level 16, where the difference is quite small. For wizards, this will begin at level 15. I do not know if it lasts through level 17, but it clearly ends by level 18. I think it would mostly depend on how long it takes the fighter to get a better weapon.

At most levels, sorcerers and wizards need to hit two targets to equal or surpass the fighter in damage. Fighters, for their part, only do this much damage if they don’t need to move more than a five foot step. When a fighter must move and attack they do as much damage as a single hit is worth, which is never close to the amount that a sorcerer could do in that turn. I feel that this is a very balanced combination, although it does make mobility a key element of game balance. This means that it is only balanced if you play with a battle grid.

Speaking strategically, it makes the fighter the biggest threat to the most dangerous monster, and an even bigger threat to any adjacent monster. As such, it makes very little strategic sense for those monsters to ignore the fighter in favour of the squishy sorcerer. As such, they’ll either be constantly running away from the fighter, or attempting to gang up on the fighter. It’s a plausible form of “aggro control” that doesn’t use any threat control powers. When the monsters are far away and using ranged powers, they can safely target the mage and other squishy members of the party. In so doing, they can safely kill the Mage without being killed by the fighter.

From the point of view of concept execution, this makes the fighter the best to stand toe-to-toe with a giant monster. The fighter does the most damage and can take the most damage. In fights like these wizards are better off buffing the fighter than damaging the dragon. Once again, this fits heroic fantasy fiction quite nicely. The heroic knight fights the dragon with the help of their supernatural allies; the heroic knight does not cower while the supernatural allies attack the dragon!







69.1 33.1



79.14 35.14



79.31 2.75



118.56 18.66

So the next time someone tells you “fighters are useless,” just refer them here. Unless you find an error in my math. That would be emberassing. If anyone good at statistics wants to check this for errors, please go ahead. Something feels a little off to me about how I handled critical hit damage, although crits are actually irrelevant to my argument at all levels except 16. That makes me wonder how the math would look without the Critical Focus feat, and just taking bonus elemental damage earlier instead. It would probably be more effective. But I’m done with math for tonight, if anyone else wants to figure that out, go ahead.

Player Skill vs. Character Skill

As a GM, I run into some rather unfortunate combinations of player skills and character types. Here’s a short list:

  1. A player with terrible people skills, playing a high charisma character.

  2. A player who is bad at making tactical decisions, playing a commando.

  3. A player who is poor at teamwork, playing a healer.

  4. A player who is poor at “outside-the-box” thinking making a rogue or other skill based character.

In order for a player to be able to make use of a character’s strengths, the player needs to be skilled in a certain way. This can leads to situations where the GM might fudge how an NPC reacts to being treated disrespectfully, because the character being disrespectful has high charisma and the player has no idea they are behaving in a manner that most people would consider a bit insulting. Or the bad guys need to make stupid tactical decisions because the commando is supposed to be a tactical genius. Or everyone dies because the healer just didn’t notice people need healing, and wants to spend their time hitting bad guys with sticks instead of healing. Or the rogue seems really underpowered because they never can think of a way to earn a surprise round. If a player’s skill set doesn’t line up with how a character’s abilities imply that character should be played, then that character will either be ineffective or force the GM to dumb down the game.

I find myself questioning lately what happens in the opposite situation: if player’s skills are excellent for a type of encounter, but the character’s abilities for the situation are quite poor, will the players’ skills be ineffective? I’m inclined to think this mostly has to do with the GM’s style. Regardlesss, I’m going to argue that it is better if the GM designs their game with the intent of appealing to player skills more than character skills. This has strange repercussion for the role of skills/abilities in a game.

Role-playing Scenes and Player Skill

Lets begin with role-playing scenes. These scenes often are about the players either deciding on a course of action, or convincing an NPC to do something. A player with strong interpersonal skills will generally be persuasive. They will be able to read into the motivations of others accurately, and make use of this information. This will give them a great deal of control over the decisions the party makes, regardless of their charisma scores. Does it give them a similar amount of control over NPCs?

That depends on the GM. If after providing a highly persuasive argument the GM says “that’s very persuasive. Roll diplomacy.” then it doesn’t help. If the GM says “that’s very persuasive, so they agree with you.” then it is incredibly helpful. If a player lacks interpersonal skills, or is highly disinterested in these kind of role=playing scenes, then a quick roll could be used to resolve the discussion instead. Note that if the player has strong interpersonal skills and the GM doesn’t require rolls for persuasive arguments, the player may note that there is no reason for the character to also have high interpersonal skills.

Combat Scenes and Player Skill

Combat tends to have a few different tactical things to consider in an RPG: “action economy,” countering, and situational modifiers. Action economy is just about using one’s turn efficiently. There’s not too much to say about it. Countering is about knowing how to identify enemy weaknesses during a game based on the enemies’ description, and then being able to use that weakness. Obvious examples include attacking ice monsters with fire, or knowing to “kite” enemies in heavy armour because they move slowly. Accruing situational bonuses is acquired by a combination of rules knowledge, positioning, and in-universe reasoning. For example, if a player knows that opponents get a -2 on their attack rolls if the target is obscured by darkness (rules knowledge), can move their character into a dark hallway (positioning), and is able to argue that the candle at the end of the hall doesn’t provide sufficient illumination to eliminate the penalty (in-universe reasoning), then they have just effectively increased their defenses by 2.

Some players really dislike situational modifiers because they can bog the game down with checking tables. I find this is only a problem if the players are bad at using situational modifiers. The players who are good at it already know all the content of the tables, or at least are looking for something very specific when they need to refer to them.

If a character has strong enough combat abilities, then they don’t need to worry about situational modifiers, enemy weaknesses, or their action economy. They do a ton of damage and can take a ton of damage, so they can get away with being inefficient.

Very few game systems have rules that allow the players and the GM to take advantage of situational modifiers or counter enemies to enough of a degree that player skill can be a replacement for character builds. In GURPS situational bonuses can easily be as high as +10 or more if the game is set in the future, which can certainly compensate for lack of combat skills. Mage: the Ascension has a very open ended approach to stacking modifers, but tends to cap the total bonus to -5 (minuses are good). Both have plenty of options for countering.

Creative Thinking and “Skill Puzzles”

When it comes to designing open-ended puzzles, the GM is already in a mood to agree to unorthodox and creative solutions. A GM who uses narrow, closed-ended puzzles is probably just a bad GM, and it makes no sense to plan for GMs who don’t want to give players agency and problem solving ability during a puzzle. This means that, in practice, the players don’t even need appropriate skills to solve an open ended puzzle. They just need to be creative and capable of in-universe reasoning. However, if a player is bad at that kind of thinking, it is really easy to be successful if the character has a high score in the related skills.

Imagine this very straightforward, puzzle: the GM asks the players “how will you break into the office building?”

  1. The character who has high charisma skills wants to disguise themselves as a janitor. That’s a bit creative.

  2. The character who is good with guns wants to shoot a grappling hook to the roof, and then cut through the door on the roof slowly and methodically with hand tools. That’s pretty creative too.

  3. The character who can pick locks decides to pick the lock to the back door. That’s not very creative, but it will obviously work.

Executive Thinking and Support Roles

Executive thinking is an important ability for a lot of character types, because there are whole classes in games that are support classes. Executive thinking is chiefly about prioritizing between multiple priorities. Choosing to buff defenses or attack abilities is an easy example. Bards and Clerics in 3.x, and any leader class in 4.0, all come to mind. Making use of this as a skill tends to require that the support player knows what each other players main strengths and weaknesses are, and what is needed at the moment.

Some games are well designed to make use of this kind of thinking, others are not. In 3.x games, players are better off using aid another to assist a character do what they specialize in than to attempt it themselves, so long as the party level is below level 5. For example, fighters specialize in doing damage to single targets, and wizards specialize in doing damage to multiple targets. If there is only one target, the party is more likely to do more damage by the mage using aid another to give the fighter +2 to hit than trying to hit themselves. Above level 5, though, aid another becomes completely useless. At this point, only spellcasters who choose to use buff spells can make use of this kind of thinking in a useful way.

This kind of thinking is, in most RPGs, the kind that relies the most on player skill. If used effectively, it is often downright overpowered. Do the math on the Haste spell. If used in the first round of combat, if the combat is against one or two monsters, Haste will be responsible for more damage over the course of a combat than cone of cold, and haste is two spell levels lower. Against 3 or more monsters, cone of cold will be more effective. The big problem with these support roles is that they rely on player skill and character skills lining up. It is not possible for a player to use their skill alone to be good at this role, nor is it possible for a character to be built in away that makes them good at this role in the absence of player skill. I suppose leader classes in 4e can be good at this with build alone, but that game is so dumbed down that anyone can be good at anything by build alone.

Methodical Research and Character Creation

I would like to clarify that a player who is skilled at methodical research can rapidly and efficiently use the rule books to create powerful character builds. As such, it might be more accurate to think of mechanically powerful characters as extensions of a player’s particular skill set. This introduces a substantial choice for a GM: how much to favour one type of player skill over another. The powergamer may usually be the only player to have their skills represented in mechanics, and they are also often said to miss out on the subtler and more fulfilling parts of role-playing games. However, if one considers character optimization to be an expression of a player’s abilities, and (more importantly) an expression of the things about themselves they like to bring to a gaming table, then it becomes incredibly bizarre that most game systems overly appeal to this skill, make an effort to reduce the impact of other skill sets, and make “take the high road” style arguments in favour of other kinds of play.

Counter-intuitive Character Design

If the game is working this way, players should design their characters to be good at the kind of scenes they are bad at or don’t enjoy as players. That way they can participate in those scenes, but do so with minimal effort. Alternately, players could build their characters to get abilities that are otherwise imposssible for a person. For example, a character can’t be an engineer if they don’t know engineering. It doesn’t matter if the player is an engineer; the character does not get to use engineering to solve puzzles unless they know it. This latter alternative would mean that characters just need the minimum degree of competency to make something possible for that character.

So, when a game is designed so that player skills will usually be sufficient to be effective in a campaign, there are two reasons for a character to develop skills and abilities as normal:

  1. Achieve fuller participation in a type of scene than the player’s skill allows.

  2. Obtain novel abilities that create new options for the character.

Why Games Should Appeal to Player Abiliy

Designing scenes with player ability (not character ability) in mind has several benefits. It lets players do what they’re good at, to the extent that they want to do it. It also lets players get around their personal limitations to participate in scenes they find interesting, but lack the skill to participate in to the same extent as other players. It lets the social players be as social as they want, and quieter players can always roll a diplomacy check instead. It lets the creative players come up with ingenious solutions as much as they want, and the player who thinks break-ins are cool but can’t think outside the box can always build their character to be cat burglar. It lets the tactical genius make all the brilliant decisions in the world during a battle, and the characters who can’t memorize the illumination modifiers can just stand in a corner and make attack rolls.

Kinds of Player Skill

Here is a non-exhaustive list of player skills, the kind of scenes they usually apply in, and what character ability can be used to compensate for the player lacking that particular strength. The character skill section is left blank, because it varies from game system to game system. Some additional rows are there, also. I encourage you all to think of what you would add to the table..

Player Skill

Game Scene

Character Skill

Interpersonal Skills

Role-playing scenes


Executive Prioritizing

Any scene type, but only by helping someone else achieve their objective


Creative Thinking



Memory and Rules Knowledge

Combat (situational modifiers and countering)


Spatial Thinking

Some Puzzles, Combat with Grids





Methodical Research

Character Creation











Computer RPGs and Player Skill

The ideas I had here started with a criticism I have of most computer RPGs: if the player raises the difficulty rating of the game, the enemies usually just do more damage and have more hit points. This forces the player to grind more, but doesn’t actually increase the skill required to play the game. As such, it doesn’t actually increase the difficulty of the game. It just increases the length of the game.

This got me thinking about how player skill, as opposed to character skill, effects the game. In the end I came to the conclusion that player ability should matter far more than character ability, and very few RPGs are designed with that in mind. I’d say GURPS and Mage: the Ascension are the only ones I know of.

Closing Thoughts

I actually feel like there are substantial benefits to the players, as people, if all the players are constantly working on increasing the player skills in a game. I’ll be writing a post about house rules for balancing player and character skills for my next article. I feel these rules will be most useful in a “sandbox” game. I’ll get into why later.

Dragons and Detectives: Thrillers

Psychological thrillers are a lot like mysteries. They both engage the reasoning ability of their audience, and have plots that are principally about uncovering “the truth.” The key difference, however, is that thrillers aim to create a mood of doubt and paranoia. This is accomplished by making the characters and audience question the integrity of the investigation. Questioning the integrity of the investigation changes the mystery of the story substantially: instead of discovering “who committed this crime?” the question becomes “what is wrong with this investigation?”

This post assumes familiarity with the other “Dragons and Detectives” posts.

Integrity of the Investigation

By the “integrity of the investigation” I mean the phenomena where all people and institutions supporting the investigation are acting in good faith and aim to help uncover the truth. Thrillers challenge this notion, and in so doing create a sense of paranoia. Here is a (non-exhaustive) list of ways to challenge the integrity of the investigation:

  1. A detective is either highly immoral or has a tenuous grasp on reality.

  2. The organization that supports the detective (the police department, a P.I.’s network, etc.) contains some traitors, saboteurs, or otherwise corrupt individuals.

  3. The moral or legal authority of the organization that supports the detectives is drawn into question, or they withdraw their support.

The first and second one are the most commonly used in psychological thriller movies. Think of any thriller you’ve seen where a friendly secondary character becomes progressively more sinister as the movie goes on. Meanwhile, the main character’s behaviour becomes increasingly more erratic and bizarre.

Events for Psychological Thrillers: Sabotage and Lies

The events in a psychological thriller involve clues changing or going missing, initially for no readily apparent reason. The evidence is misplaced, or an interviewee denies that they ever spoke to the players. This sabotage of the investigation rapidly escalates to threats and “indirect” attempts on a person’s life (like lighting their house on fire). Only near the end will there be a “direct” violent engagement.

This changes the way the players will re-examine their theories, because they need to start thinking about who has the opportunity to sabotage their investigation. This will usually seem beyond any of their suspects, and will result in all kinds of paranoid speculation. It is possible, of course, that the person sabotaging the crime has a motive that is completely removed from the original investigation. In psychological thrillers, however, this does not happen because there are no coincidences. In a psychological thriller everything is connected.

PvP and Sabotage

What will make the players feel a lot more secure in their theories is that they principally interact with each other. When an interviewee says “I never spoke to those guys” all the PCs say “yes you did; we we’re all there!” The way around this is to make the PCs distrust each other.

Here is what I suggest: at the end of the first socratic session in a mystery, have each player write the GM a secret note. In it, the player writes down how their character could manipulate the investigation in a manner that benefits them. Examples could include framing a professional rival, rushing through the investigation irresponsibly to “juke the stats,” or turning the investigation into a propaganda campaign for the local police department. The GM puts a check mark on each proposal that seems plausible, and an x on each proposal that is not. If the player gets a check mark, they may choose to pursue the investigation honestly or they may choose to sabotage it. A player who got an x must investigate honestly.

I would suggest letting each player take one action secretly between each socratic seminar and event. Honest investigators can use this action to try and track down the saboteur. Manipulators can use this to fabricate evidence.

Now the game could go in a few possible directions. It could be an honest investigation if everyone chooses to solve the mystery. It could be an intrigue game if everyone is sabotaging the investigation. It’s a thriller if some are manipulating the investigation, and some are pursuing it honestly.

The manipulators would need to get the biggest reward if they can manipulate the investigation just enough to get their objectives met, the honest investigators still catch the right criminal. Of course, if the manipulator gets caught and must be punished, odds are good the character will wind up in jail or dead.

In order for the honest investigators to be able to secretly investigate their colleagues, I would want to have each player identify at least one skill or ability that they keep secret, and one part of their background that they keep secret. Presumably, this part of their background is related to how they plan to benefit from corruption. By clearly identifying these to the GM at the beginning of a campaign, the GM knows what to reveal to PCs when they investigate each other. The background provides psycho-social clues, and the skills provide opportune clues.

Doing this does break the genre convention of making everything connected. However, it allows a group to run multiple mysteries that are very close to psychological thrillers without everything becoming overly predictable.

Surprise Endings

For movies, the benefit of a constant mood of doubt and paranoia is that it keeps the “feeling” of mystery alive in the absence of reasoning. It can stimulate emotional responses from the audience in place of rational responses. Thriller movies tend to aim for a surprise-ending, so they rely on the audience not having any time to reason about the films except at the very beginning and the very end. The very beginning sets up the detectives initial, non-paranoid expectations. Then everyone gets paranoid for 1.5 hours. Then there’s a dialogue wherein the detective and the audience are walked through all the reasoning needed to discover “The Truth.” It’s a big surprise ending and everyone is awed.

To pull this off in an RPG, the GM will need to keep interrupting the players socratic seminars with threatening action scenes, distracting rp scenes, and dangerous threats best avoided with neat puzzles. They get one socratic seminar at the beginning, then the GM interrupts them until they have all the clues they need, and then the GM lets them have a socratic seminar again. They put it all together, and it’s a big surprise ending.

The group would probably need to finish this whole plot in one session, because if the players have free time in the middle of the plot they might just put it all together too soon.

The Inquistion: a Campaign I Would Want to Run

If I was to run a thriller campaign, I’d want to set it in a low-fantasy version of europe and make the players work for that world’s version of the inquisition. This makes the players into investigators, of a sort, but instead of looking for breaches of the law they are looking for breaches of morality. They aren’t even looking for breaches of public morality; they are ferreting out “corruption” in people’s private lives. They are certainly allowed to enforce public morality if they want, and it is through enforcing public morality that the players end up dealing with high-profile murder mysteries.

This means, right from the beginning, the integrity of the organization that supports the PCs is cast in doubt. The inquisition will obviously abuse its power. Some of it’s more ruthless agents will punish people for “moral crimes” that shouldn’t be crimes in the first place. Others (like the PCs, I hope) use the scope of their authority to try and benefit people by removing corrupting influences, and supporting nurturing ones.

In the course of removing “corrupting influences,” the players (and other inquisitors), would have the authority to be as fascistic, xenophobic, arbitrary, and generally evil as they want to be. This enables, but does not require, the players to threaten the integrity of the detectives (themselves). When some of the clues take the form of forced confessions, and some of the reasoning involves obvious bigotry (“if he’s a heretic, he must be a murderer too!”), the players may begin to doubt their own ability to find the real criminal.

Finally, if bishops and nobles are frequently prideful, conceited, and deceitful, the PCs will almost certainly not trust the integrity of the people “above” them. Other inquisitors can be prone to obvious bigotry and corruption, and political clout can be used to shield the powerful from investigation. Paranoia will natural ensue.

Depending on players experience with bigotry in real life, this game could be very upsetting for them. I’ve had some unpleasant encounters with neo-nazis in my lifetime, but they were few and far between. Many people have had it much, much worse. Also, if people have been abused by police in any way this game might be really unpleasant. This could also be a really heavy game for people because of how concepts of public and private morality can be severely impacted by gender or sexual orientation. If you want to use this idea, be careful; it could make a lot of people really uncomfortable.

Dragons and Detectives: Clues and Events

In a straightforward detective story, the clues necessary to reach the correct explanation of the crime are presented at the beginning of the story. The problem is that these clues can be used to create multiple theories. Then the story progresses, and the detectives find more information. This information is communicated through events, and it allows some interpretations to be disproved. Sometimes, an event occurs that changes a key assumption about the case, forcing all theories to be re-evaluated.

So, for creating a mystery game, we’ll look at clues and events. I’m going to assume that the players are using some kind of socratic seminar-style discussion to come to conclusions about the mystery.

The Clues

For the sake of games, I think it’s useful to divide clues into three categories: physical, psychosocial, and opportune. Physical clues establish what actually happened at a scene of a crime. A bottle of poison at the scene of a murder is a good example. Psychosocial clues are used to determine who would have a motive to commit the crime. They are largely gathered by interviews and observation of people, and reveal properties of a person’s personality and relationships. Examples might include that the victim was having an affair, and their significant other had just learned of the affair. Opportune clues are clues about the opportunity to commit a crime. They include where a person was at the time the crime was being permitted, or skills or ability required to commit the crime.

It is important to always include at least one clue in each category that has nothing to do with the crime. This is important for increasing the number of possible theories the players will come up with, and it makes using events later easier. It also rewards players who are particularly good critical thinkers, who might figure out what is a coincidence and what is not with their own wits. It is also common to have a forensic clue of some kind that needs to be analyzed. If the players have the appropriate skills themselves, they can get this information effectively immediately, but otherwise they need to wait. This waiting allows the GM to introduce information later in an Event that will result in the players making new theories.

Physical clues tend to be the first one’s analyzed. This is important for when the GM is planning events, but at this precise moment it doesn’t matter. Analyzing physical clues will mostly be used for determining supporting psychosocial and opportune clues.

Imagine the site of a murder has furniture knocked over, an empty bottle of poison, a broken window, and a bunch of blood splattered on the table. There are also no cuts or visible wounds on the victim. A player may deduce from this that the person was poisoned, thrashed while dying (knocking over the furniture), and then someone broke the window and took something from the table. A different player might imagine a scenario where the victim was poisoned with the murderer in the room, who then gloated, and the victim attacked the murderer in a rage (knocking over furniture and spilling blood on the table in the process).

Acquiring psycho-social clues is done through interviews, so the players need a way to learn about important people from the victim’s life. Once that is done, they can interview them. The goal is to be able to understand the victim’s place in other people’s lives, so that the players might figure out who would murder them. Since the murderer presumably had a motive, the victim is presumably not quite a saint. This is often represented by at least one person coming right out and saying “I didn’t kill him, but I wish I did!” One of the interviewees will withhold information that they find embarrassing. This is important for later in the investigation, as it will make a person suddenly seem much more suspicious regardless of whether they fit a player’s theory or not. A third type of interviewee will be missing, refuse to be interviewed, or postpone the interview until later. This will allow the GM to feed the PCs information later on in the investigation if the players are completely on the wrong track, and if they are on the right track the interview in the future would constitute an Event. The list of people who are important in the victim’s life can be narrowed down substantially with a bit of common sense thinking about opportunity. This means the players shouldn’t be provided with or make a list of every single person the victim knew, but should instead make a list of people with whom the victim had contact recently.

Opportune clues are the last kind, and rely on the players already having a theory about what happened. It is by finding (or not finding) opportune clues that they can support or weaken the various theories that they came up with. It will often be up to the players to ask questions like “where were you on the night of january 7th between seven and eight in the morning?” to get this evidence. It is largely up to the players to ask the right questions and look for the right kind of evidence. To make this easier, if a player asks a broad question like “I look around the room for clues” the GM can ask them to narrow it down. For example, if the character wants to know if the person is fit enough to overpower the victim, they might clarify that they want to determine how physically powerful the character is. The GM can then respond easily with “the suspect has a cane in the corner of the room” or “the suspect has anabolic steroids in their medicine cabinet, and many boxing trophies.” However, the GM will most likely want to plan future events so that they release more evidence about the opportunity to commit the crime.

Opportune clues are therefore used mostly to disprove theories.

To help with this, here is a table you may use.







(Hates the victim)

(Withholds information)


(In event 1)

(In event 2)

(In event 3)


There are two very different kinds of events: events that introduce new clues and events that change existing clues. An example of the former is when some hired goons show up to threaten the detective, and (after beating the tar out of the goons) the detective learns they hang out at Miss Kitty’s Gentleman’s Club. An example of the latter is when a forensic scientist calls the detective and says “You won’t believe this! The bullet casings weren’t from the same pistol found at the scene of the crime.”

There’s another purpose to events, though. It breaks up the scenes of role-playing (interviews) and reasoning (out-of-character or in-character Socratic sessions) with other types of play. The sudden arrival of hired goons tends to lead to combat. A new but uncooperative interviewee might need to be caught in a chase scene. Perhaps the victim’s loved ones are being uncooperative, so the detectives decide to sneak into their home and look for more clues.

Since a mystery game will already have frequent opportunity for acting and role-playing, there’s no need to try and introduce more of this with events. Feel free to make new NPCs uncooperative to the point of violence, to include more combat. It also will include a large amount of in-universe reasoning, so there’s not really a need to include more really open-ended puzzles. If any puzzles come up, they should be more about spatial reasoning than logical reasoning. They would thus apply more to sneaking around.

Often, the players will be choosing what events are going to happen based on their attempts to prove/disprove their theories. The GM has a responsibility to make sure the events will provide the information needed to disprove wrong theories. If the players have largely failed to create a theory that correctly explains the crime, the GM should make sure to introduce clues that will help steer them in that direction. Often, this will occur because the players are over-focused on a single clue, so use the event to change that clue a little bit.

The GMs goal with events is twofold: give the players enough to think about that more socratic seminars are meaningful, and slowly steer the players towards the correct solution. This isn’t to say that the PCs are guaranteed to reach the correct solution eventually; it’s simply a reminder that the GM needs to make sure that the PCs have a good enough chance for the game to be “fair.”

The Red Herring

The red herring is the theory that is wrong, yet still as plausible as the correct theory for most the story. The players will almost certainly create their own red herring during the socratic seminars. However, it is important to not immediately disprove the Red Herring with events. Instead, make sure that new clues support the correct theory, without disproving the red herring. As such, only players who are particularly apt critical thinkers will notice that they have better reason to believe the correct theory than the red herring. If the GM is particularly good, there will be a period near the end of the mystery where the players divided between just the red herring and the correct theory, and both theories will seem very reasonable. One last piece of evidence will end up proving which theory is correct,

The Correct Theory

The funny thing about a mystery is that the GM doesn’t actually need to decide what actually happened. I suggest the GM decides about half the time what actually happened. The rest of the time, the GM just creates the clues and improvises some events that will support or disprove some of the PCs theories. Then the GM picks one of the PCs theories and decides to make that theory the correct one. If the GM is feeling a little bit perverse, the GM can wait until the PCs end up deciding between two plausible theories, and the GM makes whichever one the party chooses the Red Herring. This is contrive a particular type of climatic encounter: the PCs accuse a person, that person is innocent, and then the real criminal reveals themselves and tries to kill everybody. In theory, the real criminal does this because the detectives “are too close” and “it’s only a matter of time.”

Making it so that half the time the correct theory has been decided in advance makes for fair play. Choosing a theory from among the players’ theories will make tend to make the players feel more smart, because the investigation will go unusually smoothly. This gives the GM and players more time to focus on role-playing and the action scenes. The perverse way is good if the players start getting a little bit full of themselves; just don’t tell them about it. If the GM is good at making play fair when they make the theory in advance, the players should be successful at least 50% of the time. If the players are successful more than 75% of the time, consider using the perverse method to make it so they must fail some of the time. They’ll probably be happy and more engaged in later sessions, so long as they don’t know that the GM decided to make it impossible for them that one time. It’s a dirty trick, but it works.