Every base roll is Good Luck minus Bad Luck
Each side throws two six-sided dice, one marked Good Luck and one marked Bad Luck, and subtracts the second from the first. A Good Luck 4 against a Bad Luck 1 is +3; the same faces the other way round are −3; any double is zero.
The subtraction matters because many combinations cluster near zero. Extreme base results are possible but less common. A plan that needs the maximum face difference is fragile, while a plan that benefits from several neighboring results is resilient.
An opposed contest compares final totals, not one player's dice
Both sides roll Good Luck minus Bad Luck. Applicable situational, role, stance, equipment, protection, target-state, and contract modifiers are applied according to the selected mechanic. The actor's final total is then compared with the opponent's final total.
A positive roll can still lose
The actor throws 5 and 3 for +2. The defender throws 4 and 4 for zero, then adds +3 from a defense that applies, finishing on +3. The actor's roll looked good on its own, but the defender takes the contest by 1.
Always read the displayed breakdown. A stat or item is not a universal modifier merely because it sounds relevant; it joins only when the action or defense rules say it applies.
The final margin describes the strength of the contest result
Neutral
The final totals tie. The selected action defines what a neutral margin means.
Standard
The winning margin is 1–2. The winner has a clear but limited edge.
Strong
The winning margin is 3–4. The action may produce a more decisive effect.
Critical
The winning margin is 5 or more. The action's rules decide the critical outcome.
Below zero the bands run the other way: a negative final total is a Setback, and −5 or lower is a Botch. Each band names a strength, not an effect, so an investigation, a strike, a guard, and a persuasion each decide for themselves what their own Strong or Setback means.
The base dice favor modest results over extremes
There are 36 equally likely Good Luck and Bad Luck face pairs. Six pairs tie at zero. Five pairs produce +1 and five produce −1; the number of combinations decreases as the absolute difference grows. Only one pair produces +5, and only one produces −5.
| Base result | Face pairs | Chance of exactly this result |
|---|---|---|
| 0 | 6 | 16.7% |
| +1 or −1 | 5 each | 13.9% each |
| +2 or −2 | 4 each | 11.1% each |
| +3 or −3 | 3 each | 8.3% each |
| +4 or −4 | 2 each | 5.6% each |
| +5 or −5 | 1 each | 2.8% each |
Read as thresholds, a base result of +1 or better comes up 15 times in 36 (41.7%), +3 or better 6 times (16.7%), and +5 exactly once (2.8%). A base result of −5 is a Botch if modifiers do not lift the final total above it.
Modifiers improve a contest without promising it
Anything that can tilt a contest, such as where you stand, your stance, your gear, the target's protection or condition, your role's rules, or a contract, is applied to either side by the mechanic you chose. The more situational bonuses pile up, the less each extra one adds, though a few signature bonuses skip that curve. When the breakdown and your own arithmetic disagree, trust the breakdown.
Which stat you roll depends on what the ability does, not on your role. Strikes, arrests, and crime-economy jobs use Offense; reads use Perception; guarding and rescuing use Defense; covert jobs use Stealth; persuading uses your Influence standing; and sabotage draws on Offense and Stealth together. An ability's Known Factors line shows which one it will use before you commit.
Hunt
An offensive, watchful posture sharpens attacks and investigations while increasing exposure and the story the player must later explain.
Lay Low
A defensive posture steadies a guard or a rescue and protects a vulnerable plan, but surrenders attack pressure.
Blend In
A stealth posture strengthens covert work and can reduce visibility without making the player undetectable.
The best modifier may be a different legal target or a later timing window. One additional point against strong protection can have less strategic value than collecting an independent clue first.
What one effective point is worth
The table assumes the other side adds nothing, which you will rarely know in a real match, so treat it as the price of a point rather than a forecast. Your effective modifier is what is left after diminishing returns. Going from even to +1 lifts your chance from 44% to 56%, and each further point adds a little less: the fourth buys about eight points of odds where the first bought twelve. Ties are not wins here; a tied contest is Neutral, and the action decides what that means.
| Your effective modifier | Chance to win |
|---|---|
| −2 | 24% |
| −1 | 34% |
| ±0 | 44% |
| +1 | 56% |
| +2 | 66% |
| +3 | 76% |
| +4 | 84% |
Bonuses a seat earns during a match
Three bonuses come from how your match has gone rather than from your role or kit. Each one is small and bounded, so the useful question is when to spend it, which the point-value table above answers.
Learned the hard way
Two failures in a row with one kind of ability (reading, guarding, covert work, or persuading) earn a single +1 on your next try at that kind. It comes once a match and never from attacks. If your reads keep missing, the next one is the best-odds read you will get, so aim it at the seat you most need to clear.
Push
From the second cycle you may push one declared night ability a cycle for +1 at a cost of 1 Heat, and pushing an attack marks you Spotted. Pushing a contest you expect to be even is worth about twelve points of odds; pushing one you already lead is worth less, while the Heat costs the same.
Scars
Surviving a wound that leaves you on your last life gives you a permanent trade for the match, one stat up and another down, chosen by where the wound landed. Plan your remaining cycles around the stat that went up.
A push or a lesson adds nothing, and a push costs nothing, when a stance passive or a banked follow-up already gives that roll +1.
The Learn guide lists all four scars. Read the contest breakdown after resolution to see which bonus, if any, actually joined the roll.
Judge actions by expected information and objective value
Before submitting, list what a win, tie, and loss each accomplish. A protective action may preserve a key voice on success and reveal hostile pressure on failure. An investigation may generate a strong clue or merely narrow a category. An attack may create exposure even when it misses.
Avoid all-or-nothing plans when another action creates value across more bands. The goal is not to eliminate randomness; it is to choose contests whose uncertainty still advances the seat's objective.
A failed roll is not proof that the declaration was false
Players often remember the actor's visible die and ignore defense, modifiers, or margin. That creates bad courtroom evidence. A declared investigation can legally fail; a positive attack roll can lose to a stronger defense; a successful action can still have a limited effect.
Use the result to update probability, not rewrite history. Compare the declaration made before resolution, the legal target, the final breakdown, and the later public story. Contradiction is evidence; ordinary variance is not a confession.
Worked contest examples
Strong actor win
The actor throws 6 and 2 for +4 and has no modifier. The opponent throws 3 and 2 for +1. The actor wins by 3, a Strong margin, and the chosen action decides what a Strong result does.
Neutral contest
The actor finishes +2 and the opponent finishes +2. The margin is neutral even though neither base roll was zero. Read the selected action's neutral outcome.
Modifier changes the winner
The actor's base result is +1 and the defender's is 0. An applicable +2 defense moves the defender to +2, reversing the apparent one-point actor lead.
Common dice mistakes
- Reading only Good Luck. The base result subtracts Bad Luck.
- Reading only the actor. Opposed contests compare both final totals.
- Assuming every stat always applies. The selected mechanic controls modifiers.
- Promising a Critical result. Better odds never become certainty.
- Turning failure into a role verdict. Evaluate the declaration and breakdown before alleging deception.
Dice FAQ
What is the most likely base result?
Zero has six of the 36 face pairs. Results near zero have more combinations than extreme positive or negative differences.
Does a Critical margin always mean the same thing?
No. The selected action defines what Standard, Strong, Critical, Setback, Botch, or Neutral actually does.
Where can I plan around rolls?
Use the strategy guide, then connect results to public proof with the evidence guide.
Revision history
Dated changes to this guide, newest first.
- — Full editorial review. Added the base-roll table, what one effective point is worth, and the bonuses a seat earns during a match, and rewrote the worked examples so they no longer repeat the Learn guide.
- — Each ability now rolls the stat its lane names.
- — First published.