Blades in the Dark Is Innovative, but Difficult to Run

A cloaked man, lit by fire light, with a dagger in one hand.

Blades in the Dark is a new roleplaying game, designed by John Harper and published by Evil Hat Productions. It is a game of desperate criminals in an industrial dystopia, taking heavy inspiration from TV shows like Peaky Blinders, video games like Dishonored, and novels like The Gentlemen Bastards. Players take on the roles of a criminal crew trying to survive on the streets of Duskwall as they are squeezed by the law on one side and rival criminals on the other.

The game’s design history is rooted firmly in Apocalypse World, but it is a completely original system. It puts a heavy emphasis on fiction-first gameplay but also has a high level of crunch. How well does that combination work? Let’s find out!

Character Generation Is Fast and Easy

Blades in the Dark hits the ground running with character creation. Once players pick their class,* nearly everything they need to know is right there on the sheet. Every skill* in the game is on the sheet, as are all the special abilities. This speeds up the process considerably and makes everyone’s life easier.

The classes provide a good balance between structure and free-form creation. Each class dictates what special abilities a character has access to, and players are then given skill points to customize their characters. One player’s Cutter might be good at smooth talking in addition to the default fighting while another player’s Cutter has a strong mechanics hobby. The game gives players just enough points to make competent characters, and rules on how those points can be spent prevent the risk of super-specialized PCs who can only do one thing.

The only real complaint I have about character creation is that some items on the character sheet are confusingly worded. For example, the “Playbook” track looks identical to the attribute tracks, making players think that Playbook is something they’re supposed to roll. But really it’s for tracking experience points. It’s also not immediately clear what the difference is between “Pleasure” and “Luxury” in the vice section.

That small gripe aside, Blades gets an A+ for its character creation. Players can make what they want and it doesn’t take a long time. Excellent.

The Faction Rules Are Excellent

Speaking of excellent, the shining gem of this system is its faction rules. In addition to their character sheets, players also have their crew sheet to look after. The crew represents the PCs’ communal holdings and the NPCs who work for them. It’s small at first, little more than a lair and a goon or two, but through hard work and skulduggery, it can grow to rival the great powers of Duskwall.

Building up the crew is the default motivation for everything the PCs do. The GM is free to craft other storylines of course, but at the end of the day there will still be cutthroats to pay and street-wars to wage. The narrative has a three-step process. First, the PCs plan scores so they can afford to increase their crews’ assets. Then the PCs have to deal with whoever their score angered. Finally, the Bluecoats come sniffing around to see who pulled that score.

And the cycle continues. The crew upgrades are both flavorful and mechanically relevant. Getting a canal boat lets the PCs move more cargo, and it adds to the feeling of a labyrinthian city where criminals can literally sail below the law’s notice. Training up the crew’s goons means they have a better fighting roll and that other crews start to take notice.

The mechanics for all this are complex, but not overly so, and most groups should be able to master them without trouble. This in itself is a major accomplishment. The only other roleplaying game I’ve seen with rules similar to this is A Song of Ice and Fire, and the faction rules in that game are so dense they bring the game to a screeching halt.

Beyond the fun of managing a criminal crew, these rules have narrative value as well. I’m normally extremely skeptical of the idea that mechanics can generate a story with little input from the GM, but Blades in the Dark manages it. Every score has a complication that the PCs must deal with, and if they fail, their crew takes a hit. This creates instant investment, because the crew is something the group can actually lose. There’s a lot of fun to be had just pulling scores and climbing the criminal ladder, without ever needing a main plot.   

The Stress and Harm Rules Are Elegant

Stress is both the game’s meta currency and a major consequence of failed rolls. PCs give themselves stress to get extra dice or to help out their friends, and they gain stress when resisting harm. They must play a dangerous balancing act, deciding if they should just accept the broken arm or risk the stress needed to avoid it. The broken arm comes with serious penalties, but getting too much stress can take the character out of the action completely.

This resource-management mini-game adds a tactical element to play, making sure the more mechanically oriented players always have something to do. It also shows that a character is deeply affected by their experiences. Even if a PC never takes a direct hit, the buildup of stress that comes from being a successful criminal can be their downfall.

Between scores, PCs have a chance to reduce their stress by indulging in vices. These vices run the gamut from traditional favorites like wine and drugs to truly bizarre options like locking oneself in a room and staring at a bronze idol for two days. Whatever the vice, indulging in it carries dangers. A PC might overindulge or get picked up by the Bluecoats. There’s even a possibility that the character could go on a multi-week-long bender, and their player will need a backup PC for the next session.

But PCs who ignore their vices accumulate more and more stress until they suffer truly traumatic consequences. It’s both fun to roleplay and offers mechanical depth.

The Flashback Rules Are Fun, but Limiting

Blades in the Dark highly discourages players from planning their score ahead of time. Instead, the GM is supposed to start things off when the PCs run into their first obstacle: a locked door, a sentry bot, an angry ghost, etc. The PCs deal with each obstacle as it arises, and they can add planning via flashback.

For example: If the PCs encounter a locked vault-door, one of them can say “I totally got the key for this door before we went on the mission.” Then they’d narrate how they stole the key, get the GM’s approval, and roll to see if they were successful.

The flashbacks are fun to use, no question about it. They make the score feel like an Ocean’s Eleven film. And since each flashback costs a point of stress, they add to the stress management system as well. Players must decide if an obstacle is difficult enough to justify a flashback or if they’ll deal with it in the present.

Unfortunately, flashbacks do present some difficulties. For one thing, it’s hard to completely avoid paradoxes within the fiction. The rules say that flashbacks can’t override anything that’s already been established, but that’s easier said than done, especially if a PC fails their roll in a flashback. If they take an injury as a result, does that mean they retroactively had the injury the whole time? Would they really have gone on the score with a sprained ankle?

Flashbacks can also be frustrating to players who enjoy planning out every detail of a score. These players can end up feeling like they’re being charged stress for something they would have taken care of ahead of time if the rules had let them.

Even with these problems, flashbacks are fun enough to be a net positive. They’re worth using as long as GM and players both are on the lookout for potential pitfalls.   

The Time Rules Are Useful, but Awkward

A major component of Blade’s rules is the progress clock. If the PCs want to accomplish something big, the GM creates a clock with between 4-10 sections. Each time the a PC succeeds on a roll, they fill in some of the sections. Once every section is filled in, the PCs accomplish their objective.

For example, let’s say the players want to get a Bluecoat’s badge from inside a locked safe. The GM rules this is too big to do at all once and creates a six-section clock. First a PC rolls to sneak in the front door, which fills in two sections. Then they roll to knock out the guard, etc.

Progress clocks allow the GM to break up difficult or complex goals into manageable sections. They can also be used to add a time limit, with the GM creating a clock to represent when enemy reinforcements arrive or the McGuffin is moved to a different safe house. Each time the PCs make a roll, the GM fills in a box on the clock. This is great for increasing tension, and the idea is worth house ruling into other systems.

The problem with progress clocks is the game expects them to be used for nearly everything. A number of important mechanics only work if there is a progress clock to affect, to the point that rolling only once for something is extremely awkward. This can slow the game down at the worst time, because suddenly even simple objectives must be rolled for multiple times.  

The Core Die Mechanic Is Cumbersome

Okay, now for the parts of the system that are just bad. Blades’ core dice mechanic is technically a dice pool system, but it manages to be the worst dice pool I have ever seen. It has all the problems of a system like Dungeon World, and it creates some new ones too.

PCs roll a number of d6 equal to the skill they’re using, with a few opportunities for extra dice. The only die that matters is the highest one. If it’s a six, that’s a complete success. If it’s a 4-5, that’s a partial success. A 1-3 is a failure.

First, this a static difficulty system. That means the likelihood of success is dictated only by the PC’s skill, no matter what they’re trying to do. Picking a rusty padlock has the same difficulty as opening a bank safe. Once players get 4-5 dice, something they can do in character creation if they set their minds to it, their chances of total success exceed 50%, no matter what they’re trying to do. This creates a nonsensical atmosphere at the table, with successes having little to do with what’s happening in the fiction.

In what I believe is an attempt to alleviate the problems of a static difficulty system, Blades adds two sliders to every roll. The first is position, and the second is effect. The worse a character’s position is, the bigger their penalties for failure, whereas effect is a general measure of how much impact their roll has.

These two sliders mean every roll has a bunch of possible outcomes, and there are no good guidelines for deciding which outcome to choose. A partial success can mean the PC suffers a wound, or it can mean they suffer a wound, a complication occurs, their effect is reduced, and their position is worsened. It’s entirely at the GM’s discretion.

With so many options, it’s impossible to set the stakes of a roll ahead of time. Instead, PCs and GM argue about the results after the roll is made. This does technically allow more difficult tasks to carry greater risk, but only in what happens if the character fails. A six is still always a success. As far as letting the GM adjust difficulty to reflect what’s happening in the fiction, it feels like someone has taken away my hammer and replaced it with a rock, a stick, and some duct tape.

Effect is also difficult. It’s not always clear what extra or reduced effect even does, and then there are multiple factors to consider when deciding if a character’s abilities will grant them extra effect. The whole system is overly complex and provides little utility.

The Examples Are Contradictory

Blades’ dice mechanic is bad, but the examples of how to play are worse. They seem to have been written with little thought to how they would actually work. For example, on page 30 the rules say not to negate a PCs goal with the consequences of a partial success. That’s good advice, but then it goes on to say that if the PCs goal was to corner their enemy, the consequence could be that the enemy is in the corner, but now they have the PC’s gun.

What? That is a level of semantic trickery I would expect from the mythical Sphinx. If a GM did that to me, I would certainly feel  like my goal had been negated or that the GM was messing with me. Sure, I technically put my enemy in the corner, but they have my gun now so that doesn’t even matter!

This is hardly the only bizarre example in the book. Page six says that the players have final say over which skill can reasonably be used to solve a problem. That sounds ripe for abuse, but okay, that’s how the game wants to play it. But then page 25 contains an unintuitive workaround through which a GM can declare that an action simply won’t work. So players don’t actually have the final say; they’re just told they do. That’s a recipe for conflict at the table if ever I heard one.

On the bright side, none of these examples dictate how the game is played. They’re examples, and an experienced group can just ignore them. But less experienced GMs won’t know it’s a bad idea to manipulate their players’ intent or tell the players they have final say over something when they do not.

The Setting Is Predisposed to Evil

This last issue isn’t about the system at all, but rather the setting and the premise. By default, the PCs in this game are bad people. They push drugs, murder for money, intimidate civilians, or all three! In some way or another, they make money by promoting human misery.

This can be a problem, because PCs already trend toward the dark side. There’s a reason the term “murderhobo” is so often used to describe them. It can be a struggle to keep the party from dipping into heinous behavior, and that’s assuming the basic premise of the game is to accomplish a noble goal.

Blades in the Dark is the exact opposite. It encourages the PCs to be bad people. This has a good chance to end up with a group that’s less like the characters of Peaky Blinders or Gentlemen Bastards and more like a group of serial killers.

Not all groups will succumb to this, and not all GMs will be bothered by it, but if you’re considering this game and don’t want to run the Evil League of Evil, spend some time thinking of a premise that will give your group more noble ambitions. The PCs could be part of a revolutionary movement or a group of Robin Hoods, so long as it provides some pushback against evil.

Blades in the Dark has a lot to recommend it. The faction rules in particular are amazing and worth reading even if you never plan to run the game. GMs who can tap dance around the dice mechanic can have a lot of fun with this game. But those issues are not small ones. At first, players will enjoy their over-abundance of success, but eventually it will get boring. The game makes it difficult to provide effective opposition, and a story without effective opposition isn’t interesting. At the very least, this is not a system for inexperienced GMs; it demands too much.  

Treat your friends to an evening of dark ritual murder. In a fictional game scenario, of course. Uncover your lost memories and save the day in our stand-alone game, The Voyage.

Read more about



  1. Neil Smith

    A decent review.

    One technical clarification: flashbacks don’t _require_ stress. If you had an easy opportunity to do something, there’s no stress cost. Stress is only invoked if the flashback represents an unlikely opportunity or a complex task. The alternative is that the player spend a downtime action (and perhaps Coin) to set something up before the job starts. When I ran a playtest campaign, something like two thirds of flashbacks didn’t cost Stress to invoke.

    The warning about games becoming evil is appropriate: groups should have a discussion about tone before starting the game.

    As for the other concerns you raise (and they’re all valid questions to ask), all I can say is that they didn’t come up in play for my group. Part of that, I think, is that the game both requires and fosters a higher level of trust between GM and players than many others. The assumption that the PCs are already highly competent scoundrels means the GM should be saying “Yes” a lot to their suggestions. The flashback and equipment systems only work if the GM agrees to virtually every request, allowing the dice system to create the costs and complications that make play interesting. Once you have that atmosphere, the game tends towards the GM and players co-operating to make interesting characters and situations that play out in interesting ways.

    I strongly suggest you give the game a few sessions of play. You may revise your opinions of some of the mechanics once you’ve seen them in use.

    • umbralAeronaut

      Gonna agree with Neil here. I have GM’ed a campaign of Blades in the Dark for approximately 15 sessions (through Roll20 in fact) and have to say I think it’s among the finest RPG systems in existence at what it does. I have found the core die mechanics to be quite good and scale shockingly well with player experience (I too was concerned about how the system would handle dice pools of 4 dice or higher, until I ran it). Another technical clarification, contrary to the review it’s mechanically impossible for a player to have more then 3 dice in an Action after character creation. You should double-check that.

    • Oren Ashkenazi

      You’re correct that a starting character can’t have more than a 3 rating, but unless I’m reading something wrong, five dice on a roll isn’t too difficult right from the start.

      2 Dice from the character’s rating, raised to 3 with the correct Crew benefit (like Deadly for the Assassins), +1d from Pushing Yourself or a Devil’s Bargain, and +1d from a teammate’s help.

      • umbralAeronaut

        Ah, you have it right then. Fair enough. It should be noted that to achieve that dice pool the PCs are spending 1-3 Stress (1 from Assistance, 2 from Pushing), or accruing a lot of Devil’s Bargains which are basically ‘free’ Moves for the GM that a player agrees to as price for a bonus action die. And that is for the single skill that the PC has decided his character is the best at at start. Spending the stress isn’t sustainable forever, and moreover a massive die pool is still no guarantee of a 6 (the Succeed with no Consequences result).

        The game system breathes best in the 4-5 success zone for the Players, where outcomes are happening that simultaneously advance the player’s goals and bloody the PC’s noses. It’s a thing you have to see in action to believe, perhaps.

  2. dizzy

    i’ve run three games of blades so far, and while largely i agree with the assessment, i’m comfortable with the drawbacks you’ve identified. the resolution is static and a little loosey-goosey, but my players and i have been doing this a while, so we can play with the particulars until they feel right, and the whole game just drips with the kind of flavor we adore. also, i find riding the heat for kills clause real, real hard keeps things from getting *too* dark, altho having the players explicitly and necessarily start as criminals is risky, in terms of nasty behaviour.

    from the rules: “add +2heat if killing was involved (whether the crew did the killing or not–bodies draw attention)”. heat matters, and two heat is a fair amount. after the first heist, my crew realized that they needed to be real judicious about drawing blood, and that alone reigned in the sort of darker thematic elements; once killing is off the table a sort of baseline morality is established.

    • Jim Montgomery

      I just started watching the RollPlay Blades series on Youtube and the crew is a group of Assassins, and the first advance they took was Crow’s Veil, which eliminates the 2 heat for killing people. Those PCs seem pretty monstrous so far. The GM/Players -definitely- need to be on the same page tone-wise.

  3. Patroclus

    I’m a bit confused what you mean by the nature of a rusty Padlock and a Bank safe, as the rules have very clear ideas what to do with both and they are very different from each other.

    For the rusty Padlock, It is a lower tier, so you have greater effect, in this case, a single success might get you in, and you might get a bonus die to deal with it depending on the tools you bring.

    For a bank, you put up a wheel of 8 ticks. If they didn’t bring proper equipment, that means the GM can declare No Effect and the player must push themselves to achieve even 1 tick on a success. Even with the right equipment, the higher tier of the bank versus their tools might mean limited effect, which takes much longer to open the bank safe. Which means more chance to get caught. Even a 4-5 could be deadly, as you make progress but can also alert any guards or other security measures.

    And once someone is alerted? Make a clock with 4 ticks, that’s how long you have to get out with the loot before you are surrounded with little chance to escape.

    The padlock is the small obstacle to the real heist, the bank’s safe would be the heist and should not be attempted unless the players are confident in their plan and abilities.

  4. Bobross

    I’ve GM’d 6 or so game sessions now and am finally feeling solid with it. The difficulty sliders are largely in the GM’s head and it takes some familiarity to use them correctly. Things like scale and tier took me a while to be able to implement well, but once I did, I found the players taking wounds and struggling through missions, seeing real stakes for themselves.

    Buying dice is nice, except when the crew starts at tier 0, they go against a tier 2 opponent and then suddenly they have to use stress to boost their effect size, since they’ll do nothing with a single 6. Basically, you don’t treat it like DnD – where it’s very important not to let the players get in over their heads. You can throw them in the deep end and watch them slowly drown with tiny victories keeping them going. They can succeed in every score and still go slowly insane from all the stress, or find the characters completely incapacitated from wounds, or in jail, or hopelessly entangled in a mess of inter-faction struggles.

    What I love most about the game is that when I can’t think of stuff, I crowd-source it and have the players help with world building. I had one session end in an art-show, where each player described a piece of “morally correct” (according to the Church of the Ecstasy of the Flesh) piece of art. The players become invested in the story, and so they willingly take setbacks, because the setbacks themselves are interesting. I’ve had no arguments with players, rather moments of mutual glee as they saw how truly screwed their characters were becoming. I leave the players in charge of their own exp and they handle it fairly too.

    Currently one character is recovering from a gut wound and another has spiritually fused with an eel and is becoming feral. Meanwhile a demon has taken up residence in their hideout and is cultivating an evil rose garden – with all the murders attracting law enforcement attention. Also they just started a war with a fencing school. The story grows more and more intricate with each session and it’s this – not the leveling progression that keeps the players coming back, although the leveling is fun as well. I keep getting better at gauging difficulty, and so the game stays challenging. If it loses its challenge (which it might after 15 sessions or so), I can imagine the players willingly surrendering their characters and crew and starting over as some minor faction in the same world. To me that would be an RPG success.

  5. Greg

    In TV, movies, and books, evildoers can be protagonists as long as they maintain audience sympathy. I’ve often thought there should be some kind of Sympathy Mechanic for roleplaying games that feature evil protagonists – something that doesn’t stop players from being evil, but maybe stops them from being the kind of evil jerks I hate when I find myself GMing them.

    I guess that wouldn’t be a bad mechanic even for regular good-guy RPGs, for that matter.

  6. Justin Alexander

    The resolution mechanics are actually trickier for experienced GMs than new GMs, because they’re asking the experienced GM to make a paradigm shift. In most RPGs, you imagine what success looks like and you set a difficulty number based on how hard you think it is to achieve that success.

    In Blades in the Dark, what you do is look at the situation and imagine what a typical successful outcome looks like (taking into account quality, tier, potency, and scale). That’s your EFFECT.

    It’s almost an inversion of what you do in other RPGs. But the advantage is that it allows the PCs to do a number of things to manipulate that effect (pushing themselves, getting better tools, creating an advantage, seeking assistance, choosing a different approach) before committing to the action.

    And then there’s the other aspect of the GM’s ruling: POSITION. This describes how much you’re putting at risk by attempting this action (or already at risk in some cases). You say it’s impossible to set stakes before the roll, but that’s not true: Setting POSITION is where you set the stakes.

    This breaks down pretty simply: If your position is Controlled, then you’ll either succeed, succeed with a minor consequence, or be able to withdraw. If your position is Risky, then you’ll succeed and/or suffer a standard complication. If your position is Desperate, you’ll either succeed and/or suffer a serious complication.

    Note that each position has a specific level of consequence which is on the line. So you can absolutely define what that consequence will be before you roll the dice.

Leave a Comment

Please see our comments policy (updated 03/28/20) and our privacy policy for details on how we moderate comments and who receives your information.