The ten tests are for revision, not drafting: applied while writing they produce the beat-shaped lifelessness the taxonomy warns against. Run them on a pair rather than a scene, and in order, since tests 1 to 5 (Goal, Descent, Disaster, Cost and Chain) are structural and their failures make the later ones meaningless. Record a sentence rather than a verdict: “the goal is find out why the locks were changed, and it descends from the drive across town” is useful, and “passes test 2” is not. Before tests 5 (Chain) and 6 (Three connectives), strip the connectives (and so, because of this, which meant that), since a chain asserted in connecting sentences is not yet dramatised in events. A failed test is a question about a choice, and each test here is failed on purpose by a book this chapter admires.
Test 1 — Goal
What does the viewpoint character want at the start of this unit, in one sentence with a verb of pursuit and an object?
Failure. No engine, or only a book-sized goal.
Exception. Pure reception, where bad news reaches someone who wanted nothing.
Test 2 — Descent
Which decision produced that goal, and can I point to the sentence?
Failure. A free goal, supplied by the plot’s needs; the fault is usually in the previous sequel.
Exception. A goal descending from another character’s decision, or from a parallel chain.
Test 3 — Disaster
Can I state how the unit ends in a sentence containing but or worse?
Failure. A clean win, and a next goal that will have to be invented.
Exception. The run of clean wins handed to the losing side (The Count of Monte Cristo).
Test 4 — Cost
What did the character pay, and in what currency (time, money, trust, an ally, a self-image)?
Failure. Drama without expenditure: both parties end where they began.
Exception. A deliberately weightless incident in a comic register.
Test 5 — Chain
Written as bare events, could each of three links have happened without the one before it?
Failure. The chain has become a list.
Exception. The picaresque and the episodic satire, where succession is the point.
Test 6 — Three connectives
Outlined with only and then, but and therefore, how many and thens are there?
Failure. Three or more in a row: the chain has failed there. All therefore and no but: no resistance.
Exception. A novel built on refusal of causal logic (Catch-22).
Test 7 — Dilemma
Can I state each option’s cost in one sentence, and would a sympathetic reader disagree with the choice?
Failure. A false dilemma; the reader finishes the weighing first.
Exception. The dilemma false from outside and real from inside (The Remains of the Day).
Test 8 — Criterion
At which decision would the protagonist’s earlier self have chosen differently, and what changed?
Failure. None: a working plot and no arc. Several close together: the arc spent too fast.
Exception. A book whose subject is a criterion that cannot move (Manchester by the Sea).
Test 9 — Reversibility
Could the character drop this decision between chapters at no cost?
Failure. A preference, not a decision; the sequel has not ended.
Exception. Early decisions carried by momentum, where foreclosure comes later.
Test 10 — Beat weight
Does each beat’s page-space follow its weight in the story?
Failure. The register is on autopilot.
Exception. A deliberate flatness of affect that the book makes its subject (Blood Meridian).
The and then, but and therefore test is popular craft lore rather than theory: it circulates as a formulation of the South Park writers Trey Parker and Matt Stone, and descends from Forster’s king and queen. It works because it forces the writer to state the connective, which is the one thing prose allows a writer to avoid. The exceptions are discussed among the failure patterns.
Reading the results together
Failures cluster, and the cluster points to the first move. Tests 1 (Goal) and 2 (Descent) failing together mean scenes assembled by plot requirement: rebuild forwards from the last decision that was made on the page. Test 2 (Descent) alone means a free goal, whose cause is a missing or decisionless sequel before it. Tests 3 (Disaster) and 4 (Cost) together mean incident without cost: ask what each victory should have spent. Tests 5 (Chain) and 6 (Three connectives) mean the chain has become a list: rewrite the disaster before the first and then so that it forces the next goal. Test 8 (Criterion) alone means a working engine and a static protagonist, and the remedy is to find the one decision that should have been made on new grounds and make it. A book that passes 3 and 8 and fails 1 is probably sequel-dominant by design; confirm the register rather than repair it.
The beat-weight audit
Test 10 (Beat weight) is the slowest and finds what the writer cannot see, because a default register is invisible from inside. Take ten consecutive pairs from the middle of the draft. For each, count the words given to the scene (goal, conflict, disaster) and to the sequel (reaction, dilemma, decision), and rate the disaster’s weight from 1 to 5 by how much it changes what the character can do. Then compare. Sequel length that stays roughly constant regardless of the weight means autopilot: cut the sequels after minor disasters to a clause before expanding anything. Length that falls as weight rises means the writer is more at ease with small feelings than large ones, which is common and correctable. The longest sequel following a minor disaster is usually a fossil of drafting, where the writer found the book’s voice; keep the prose and move its function. Several pairs with no sequel at all mean incident without consequence.
See also: the Reference protocol, with each test’s procedure and a table of combined failures; the audit’s interpretation table.