Tutorial

Rewritten lessons and recipes are labelled pilot until human and browser review. Old URLs have explicit migration pages; source pending is not rules certification.

By part

A — From a die to an answer

• One die: a map of possible damage — Read one probability row and check the range (pilot)
• Two dice: totals are not equally likely — Count equally likely pairs behind a total (pilot)
• Modifiers: move the results, not their chances — Use names to parameterise a modifier (pilot)
• Success: include the target itself — Build and check an inclusive target query (pilot)

B — Comparing choices and building reports

• Roll-under checks: low can be good — Ask an inclusive roll-under question · beginner (pilot)
• An average is not a success chance — Distinguish mean total from success probability · beginner (pilot)
• One table, several questions — Build one table of labelled probability questions · beginner (pilot)
• A loop builds a comparison, not a simulation — Use a loop to construct a useful comparison · beginner (pilot)

C — What information does the rule need?

• Name the outcome, not just the total — Name the outcome, not just the total · intermediate (pilot)
• Exactly this outcome, or this good or better? — Exactly this outcome, or this good or better? · intermediate (pilot)
• Keep the faces until the rule is finished — Keep the faces until the rule is finished · intermediate (pilot)
• Advantage changes the shape, not a fixed bonus — Advantage changes the shape, not a fixed bonus · intermediate (pilot)

D — Reading a handful of dice

• Keeping three equals dropping one — Keeping three equals dropping one · intermediate (pilot)
• Any and none are complementary questions — Any and none are complementary questions · intermediate (source pending)
• A count is itself a distribution — A count is itself a distribution · intermediate (pilot)
• Join unlike dice without adding their damage — Join unlike dice without adding their damage · intermediate (pilot)

E — Custom dice and transformations

• A custom face list describes a die — A custom face list describes a die · intermediate (pilot)
• Reusing a distribution is not reusing a face — Reusing a distribution is not reusing a face · intermediate (pilot)
• Armour cannot create negative damage — Armour cannot create negative damage · intermediate (pilot)
• Round each result, not the average — Round each result, not the average · intermediate (pilot)

F — Writing rules of your own

• A function returns a reusable calculation — A function returns a reusable calculation · advanced (pilot)
• Test natural-face exceptions before totals — Test natural-face exceptions before totals · advanced (pilot)
• Inspect the same pool faces in one callback — Inspect the same pool faces in one callback · advanced (pilot)
• Define what a tie means before comparing rolls — Define what a tie means before comparing rolls · advanced (pilot)

G — Decisions, rerolls, and open-ended rules

• A later save can depend on earlier damage — A later save can depend on earlier damage · advanced (pilot)
• Rerolling once is a policy, not filtering — Rerolling once is a policy, not filtering · advanced (pilot)
• Conditioning changes the question — Conditioning changes the question · advanced (pilot)
• An unbounded rule needs an explicit cap — An unbounded rule needs an explicit cap · advanced (pilot)

H — Build and defend a report

• Continuation rules matter as much as the die — Continuation rules matter as much as the die · advanced (source pending)
• Choose a policy before seeing its future roll — Choose a policy before seeing its future roll · advanced (pilot)
• Brindlewood Bay: a report must distinguish eligibility from odds — Distinguish Theorize eligibility, accounted Clues, and result queries · advanced (pilot)
• Build and defend a complete scoped model — Build and defend a complete scoped model · advanced (pilot)