The Beat Board

A beat board is a story laid out on index cards, one event per card, arranged so that the order can be changed. A beat is one event-sized unit of the story: something happens that changes a situation. That unit goes on a single physical card, written by hand and squeezed to a logline, which is the few words that still say what occurs.

The point of cards is that they can be moved, and every rule about the format exists to protect that. If an event will not compress to a line, the card is holding two events, or a decision the writer has been avoiding. Once the beats are on cards, the order becomes something to try rather than something to defend.

A board describes the story only while it matches what has actually been written. A board that has drifted from the draft will point at the wrong problem.

Core takeaways

• One event per card, compressed to about seven words. A beat that will not compress is two events, or a decision that has not been made.
• Compression is the card’s first job and storage is the second. If a squeeze pass changes none of the beats, the board is only storing them.
• The board checks events. It does not invent them. Blank cards and a marker do not produce a story.
• The joint between two neighbouring cards has to read but or therefore. If the honest connective is “and then”, the sequence has failed.
• A scene flips a value from one polarity to the other at a nameable turning point, or it is not a scene.
• The board stops being accurate unless it is brought back in line with the draft after every session. That sync is the recurring cost of keeping a board.

The card

Compression is the first job and storage is the second. A pass that forces already-written beats down to card length should change some of them. If every beat survives the squeeze at seven words with nothing changed, the board is filing beats rather than testing them.

The rules exist so the cards can still be rearranged. A prose outline cannot be rearranged that way. The working ceiling is about seven words. It is a ceiling rather than a fixed rule, and it comes from one feature-card practice as that practice circulates, not from a published statement.

Labels stay generic. “Battle in swamp” can move to another slot. “Final showdown” has already fixed itself to one slot.

A changed beat gets a new card. The old card is not patched. Software makes patching cheap, and that is what the rule is there to stop. On a laptop the move is to write a fresh card and throw the previous one away.

Under that same feature practice the whole story fits in fewer than about fifty cards. A second published sheet uses forty. A novel port sits between forty and sixty. Those numbers belong to the practices that published them.

One hour-drama room used a three-by-five-foot corkboard, three-by-five index cards, and Sharpies, and would not switch to a whiteboard. The person running that room said the squeeze-to-card step was what produced the breakthroughs. That is attributed theory, not a trial.

One circulating practice spreads the cards flat on a table instead of pinning them. The stated reason is that taking a pin out and putting it back makes people leave the order alone. That is plausible, and nobody has timed it. It is a choice about furniture, not a law.

How the board is laid out

Cards that small put the whole story in one look, which is why they are that small. A board can also run two tiers at once: a season row along the top, with the beats of individual episodes hanging under it. One season board ran thirteen cards across the top because that season had thirteen episodes. The slot for the last of those episodes held a single card that said “BOOM.”

A placeholder card is a named but unbuilt beat holding a slot. Holding a slot before the beat is built is allowed on a board. Seeding and Payoff is the same discipline run the other way: a named but unbuilt beat reserves a place the story has not yet earned.

One feature sheet is forty cards in four rows of ten, for Act 1, 2A, 2B, 3, and those forty cards contain that sheet’s fifteen named beats. The fifteen are not fifteen cards. The card anatomy on that sheet is a location header, a one-line description, a plus-or-minus pair, and angle-brackets for conflict. A novel port of the same idea runs forty to sixty scene cards in four columns, with the surface split into eight sections. An older practice puts one scene on each three-by-five card and publishes no total. That older count arrives here secondhand; the flag is in Sources.

Initials in a corner let a glance show who has vanished from a stretch of episodes. Spotlight Rotation is that scan made physical: a rotation failure is an initial missing from a run of cards. A second card color, reserved for one category, makes that category’s distribution visible without reading. One practice reserves the second color for action.

What the board does not do

Every method named here assumes the cards already have events on them. Blank cards and a marker do not invent a story. Invention happens somewhere else: in a sentence, a walk, or a page of draft. The board is the check. Compression tests events. It does not supply them.

Testing each joint

One practice maps three acts across a wall, then fills in the beats. The joint between neighbours has to read but or therefore. If “and then” is the honest connective, the sequence has failed. The seminar states it this way: “But, because, therefore: that gives you the causation between each beat, and that’s a story.” Premise and Controlling Idea is that claim run at seam scale.

Writing the connective onto the board turns the joint into an object that can pass or fail. The seminar that gave the test does not call it a board object; making it one is this page’s addition. A beat that will only take “and then” is cut, or rebuilt until it is a consequence of the one before it. Doing this at every joint should expose the places where prose had smoothed a sequence into plain succession. That is the checkable yield. When neighbouring cards belong to different plotlines, the legal joint may sit two cards down. Causation is tested along each line, not only along the wall.

Value turns

A scene flips a value from one polarity to the other at a nameable turning point, or it is not a scene. The value turn is that flip: good-for-this becoming bad-for-this, or the reverse, at one discrete moment. One spreadsheet writes the rule as columns: value in play, polarity at open, polarity at close, and the exact moment of the flip. No turning point means no value shift. No value shift means it is not a scene. The plus-or-minus pair on a feature card is the same test at card scale. One one-page variant, the Foolscap, plots each scene’s one-line event above or below an axis according to polarity, so the board reads as a zigzag.

The cull

Take a card off and read the sequence without it. If the story still stands, the card goes. After that, the survivors are rearranged until each one reads as the consequence of the one before it. Necessity is settled first and order second. Cards the story did not need stay unnecessary however well they are arranged.

Card sorting from user research

User-research card sorting uses thirty to fifty cards, one item each, and never the same wording on two cards. Matching labels make people sort by the matching words instead of by the relationship underneath. An open sort lets the sorter invent and name the groups, and that is how a mental model is found. A closed sort files cards into categories already chosen, and that is how a structure is checked. Loose scene cards sorted open may surface the story’s own shape. That is a plausible transfer, not a finding from the UX literature, because UX sorting is about membership rather than sequence. Dealing the same cards into a pre-printed four-by-ten grid is a closed sort in the same analogy. It is useful. It is not a finding.

Scene, Sequel, and the handoff

Scenes run Goal, then Conflict, then Disaster. Sequels run Reaction, then Dilemma, then Decision. A Motivation-Reaction Unit is the still-smaller pair inside both: something happens, and the character reacts. The Decision of one Sequel is the Goal of the next Scene. That handoff is the joint that holds the sequence together. Each unit forwards a named object that the next unit has to receive. A card that writes its Decision has already said what the following card must take. Two popularizers teach the pair. Their names are in Sources.

Growing a story one level at a time

Snowflake is an optional ladder: each unit at one size becomes one unit of the next size down. A logline card is the squeezed form of the paragraph it will grow into. The ladder itself does not begin as cards. Joining it to the board is this page’s move.

The times below are that author’s own clock, not a field average: an hour for a one-sentence summary; another hour to grow it into a five-sentence paragraph; about an hour per character for a one-page summary; several hours to explode each of those sentences into its own paragraph, which is the one-page synopsis; a day or two for character synopses; a week to grow each of those paragraphs into a full page, which is the four-page synopsis; a week for character bibles; up to a week for a scene-list spreadsheet; optional one- or two-page scene notes; then the draft. The middle rungs each take a week, so the ladder spends weeks of structure work before a draft line exists. Those are one author’s numbers, not a norm. What the time buys is that no level is invented from nothing.

What is unknown about a season row

A two-tier season board has no published card cap. The fifty-card feature ceiling does not travel onto a season row. The seven-word rule comes from feature practice. Whether it holds on a season row has not been checked.

Keeping the board matched to the draft

Each drafting session moves the manuscript a little farther from the board. That drift is why the second discipline exists. Sync is one of two mechanics: bind each card to the document it stands for, or rebuild a list of what was actually written and compare it to the plan. The class is any card bound to its document, or a reverse outline on paper. One software corkboard binds each card to the real document, so dragging a card reorders the binder. Order then syncs by the drag. Content still does not. The synopsis field is the mini-outline that has to be rewritten as the scene changes.

A reverse outline is written after a session. It states each new paragraph or scene’s actual function in about twelve to fifteen words, lists those notes in the order they now stand, and compares the list to the plan. Twelve to fifteen is this page’s working count for “a sentence or so”. It is not a published figure from the writing-center handout. If the comparison shows a split and the cards are not brought across, the board no longer describes the manuscript. Skip the sync and the board keeps describing the pre-draft intention, and every diagnostic run on it is then reading a story that is gone. One spreadsheet practice updates after every session so that the sheet describes the pages rather than the intention.

What the board costs

Up front: forty-odd cards written by hand to logline standard, a table or corkboard, and a rewrite instead of a patch every time a beat moves. Recurring: a twelve-to-fifteen-word line for each scene the session produced, every session, for the life of the draft. The recurring charge starts tomorrow if the board is kept.

The board is the wrong instrument for some writers. It locks structure before drafting and locks it again after every session. A writer who finds the story by drafting gets the relationship inverted, because the plan freezes before the discovery happens. For that writer the reverse outline is the piece worth keeping. The full board is a commitment made too early.

Why the board has to stay true

Seven movable words are cheap enough to rewrite, and a card cheap enough to rewrite is cheap enough to keep true. The tests on this page only pay back on a board that matches the draft. If the recurring sync stops, the cards and the corkboard are still there, and the diagnostics no longer report anything about the manuscript.

How to practice this

1. Write each beat of your story on one card, by hand, at about seven words. Notice which beats will not compress. Each of those is holding two events or an unmade decision.
2. Squeeze beats you have already written down to card length. Notice whether any of them change. If none change, the board is filing beats rather than testing them.
3. Lay the cards in order and write but or therefore on every joint. Notice any joint that will only take “and then”. Cut that beat, or rebuild it until it follows from the one before it.
4. For each card, name the value in play and its polarity at open and close. Notice any card with no moment where the polarity flips. That card is not a scene.
5. Take one card off and read the sequence without it. If the story still stands, the card goes. Rearrange the survivors only after the unneeded cards are gone.
6. After your next drafting session, state each new scene’s actual function in twelve to fifteen words. List those notes in the order they now stand and compare them to the board. Move the cards across wherever the two have split.

Related pages

• Spotlight Rotation: the character-initials overlay is that page’s mechanism made physical. A rotation failure shows as an initial missing from a stretch of board.
• Seeding and Payoff: the placeholder card is that page’s discipline in reverse. A named but unbuilt beat reserves a slot the story has not yet earned.
• Premise and Controlling Idea: causation between each beat is that page’s claim tested at seam level.

Open questions

• Whether a membership sort from user-research transfers to sequence. This page already doubts it.
• Whether the ten-step ladder’s time budgets hold for anyone but their author.
• Whether a season row has a published card cap, and whether the seven-word rule holds on it.
• Whether testing but/therefore only along the wall fails an interleaved board.

Sources

• Brett Martin, Difficult Men (2013): 3×5 / Sharpies / 13-episode row / “BOOM.” Martin in the room, season 4 of Breaking Bad.
• John August’s card practices as circulated (Final Draft blog recap of August on index cards / Scriptnotes): about seven words, generic labels, under fifty cards, table layout, rewrite, initials, second color. Craft testimony, not a trial.
• Matt Stone and Trey Parker, NYU seminar, Speakola transcript and the YouTube cut: but/therefore; “and then” fails.
• Blake Snyder, Save the Cat!: forty cards, four rows of ten, card anatomy.
• Alexandra Sokoloff, Screenwriting Tricks for Authors: forty to sixty, four columns, eight sections.
• Syd Field, Screenplay: one scene per 3×5; cull then reorder. Secondhand on this page.
• Robert McKee, Story: value-turn. Story Grid, Value Shift 101 / The Spreadsheet / Foolscap: columns and zigzag. The cadence is used; the paid method is not required.
• Dwight V. Swain, Techniques of the Selling Writer; Jack Bickham, Scene & Structure; Randy Ingermanson, Writing the Perfect Scene; September C. Fawkes: Scene/Sequel, Decision→Goal.
• Randy Ingermanson, Snowflake Method: ten steps and his time budgets.
• Nielsen Norman Group, “Card Sorting: Uncover Users’ Mental Models”: thirty to fifty, one item, no duplicate wording, open/closed. Transfer to sequence is this page’s, and doubtful.
• Literature & Latte, Corkboard documentation: card-to-document binding. The class is any bound card, or a reverse outline on paper.
• UNC Writing Center, “Reverse Outlining”: the method. The twelve-to-fifteen-word count is this page’s.

The two-discipline pairing, compress and keep true, is this vault’s (2026-07-27).