RingBRP/docs/scenarios/CLEAN_GROUND_PLAYTEST_4.md

CLEAN GROUND — desk playtest 4 (the clock)

Scenario version: v0.13, at commit 00e5a99 Date: 13 September 2026 Method: full session, GM chair, playing the proven quartet of player archetypes. Seeded roller, seed 6142, banding taken from rules.mjs rather than estimated, every roll logged below. Earlier passes were 4271, 9314, 5580/7731. This pass exists for one reason: v0.13 cut Act Two from 65 minutes to 45 and no pass had ever run against those timings. Desk pass 1's 4:25 describes a scenario that no longer exists.

What a desk pass can and cannot settle, stated first because it governs everything below. It cannot produce a wall clock. Six strangers talk at a speed no simulation predicts, and any minute figure here is an estimate with reasoning attached, not a measurement. What it can test is whether every scene has a mechanism that ends it — something in the text that stops the scene other than a GM deciding to move on. That is what v0.13 changed and that is what this pass is for.

Verdict: the clock holds, and it holds for the reason it was built to. Every scene in Act Two now ends on something written down. But the run landed at ~3:46 against a 3:40 budget, and it found three places where a roll the dice actually produced has no answer in the text — one of them in Act One, which has nothing to do with the recut.

---

The rolls, in order

seed 6142
  Bhattacharya  Research — the Sike Head file            94 vs  63   fail
  Braithwaite   Medicine — the body                      81 vs  53   fail
  Renshaw       Persuade — Swinburne                    100 vs  63   FUMBLE
  Pollard       Spot — the stump                         57 vs  58   success
  Okonkwo       Spot — the treeline, crossing            54 vs  52   fail
  Ashcroft      Insight — that memory is new             32 vs  63   success
  Renshaw       Persuade — the column will believe        7 vs  63   special
  Braithwaite   Spot — the same six at the back          32 vs  48   success
  Ashcroft      Insight — it is offered, not remembered  94 vs  63   fail
  Pollard       Listen — something on the flank          84 vs  55   fail
  Bhattacharya  Anomaly Lore — the depot entry           98 vs  58   FUMBLE
  Renshaw       Insight — Ivy on the almanac             35 vs  63   success
  Pollard       Spot — the lorry, twice a week           50 vs  58   success
  Braithwaite   Xenology (Baseline) — the six            47 vs  45   fail
  Okonkwo       Spot — the understudy is wrong           26 vs  52   success
  Ashcroft      Insight — a colleague who is not         75 vs  63   fail

  COUNTDOWN STEP 5 — opposedContestFor, the real rule
  understudy POWx5 75          rolled 100   fumble
  Ashcroft    85 -> 42 (Diff)  rolled  49   failure
  → winner none · settled FALSE · margin 0

This is a hostile seed and that is useful. Two fumbles, eight failures, and the party's one reliable Act Four test misses. A pass that only tests a clock under lucky rolls tests nothing: failure is what generates table time.

---

Act One — Registry and the Sike Head

Planned 45 · ran ~50 · over by 5

Briefing capped at ten and the cap held without effort — Registry's other appointment is a clean exit and the lore-digger stopped mid-question, which is exactly what it is for.

Then the seed took the act apart. Research fails, so the Sike Head file gives nothing. Medicine fails, so Braithwaite's own read of the body gives nothing. And Renshaw fumbles Persuade on Swinburne.

• ⚠ THE SWINBURNE FUMBLE HAS NO WRITTEN ANSWER, and H01 depends on her. The handout's failure case reads: "if nobody asks for the note, Swinburne has a photocopy of it in a drawer because she was not satisfied with what she was told, and she will produce it." That is written for a party who did not ask. It is not written for a party who asked and was actively offensive about it. A fumbled Persuade on a grieving farmer who "will not be fobbed off" should not be answered by her helpfully producing the document anyway — but the text gives the GM nothing else, so the improvised ruling was to have her produce it and say nothing. That works and it is flat.
  • This is not a timing defect and it predates v0.13. It is the third pass in a row to find that the failure cases are written for absence rather than for bad rolls.
• Timing cost: about five minutes of the table circling for another route into the body, which is honest play and not waste. It simply was not budgeted.

On the clock, Act One remains the act with no ending mechanism of its own. It ends when the GM says they are at the stump. That has never caused a problem in four passes and is noted only for completeness.

---

Act Two — Over · the act this pass exists for

Planned 45 · ran ~48 · over by 3

Beat	Budget	Ran	Did it end on something written?
The crossing	5	5	Yes — the minute-five Geiger cue
The column arrives	5	5	Yes — Ivy walks forward
The line rests	25	~28	Yes, and it was needed — see below
The line moves on	10	10	Yes — H03a handed over

The crossing cap works, and the seed proved it the hard way

Okonkwo's Spot on the treeline failed, so there was nothing to find and nothing to describe. Under v0.12's "let the silence sit until a player breaks it" that is a scene with no floor: nothing to see, nobody rolling, and a table that has not yet worked out it is allowed to speak. The minute-five Geiger cue fired and moved it. A cap is worth most precisely when the dice give the GM nothing, which is the case the old text handled worst.

THE OFFER inside the column scene costs nothing — confirmed

Run during the rest while the other three talked to Ivy. Ashcroft's Insight to notice it is being offered rather than remembered failed, so he took it for a real memory, which the text says is "worse, and is the point" — and it was. The aside cost zero wall clock because the table was occupied, which is the whole of the twenty-minute saving.

• One thing better than expected: because the rest of the table was mid-conversation, the aside had no audience, and Ashcroft's player came back into a scene that had moved on without him. That is thematically exact and it is an accident of the scheduling change. Worth keeping deliberately.

⚠ The rest clock held, but a SPECIAL fights it

Renshaw rolled a special on Persuade — 7 against 63 — and the column opened up generously. The table had more material than it could use at exactly the moment the clock wanted to end the scene, and the honest simulation is that they pushed: 28 minutes, not 25.

• The text handles the ordinary case well. "The old ones stand up, because they always do" ends a conversation that is winding down. It does not obviously cover a conversation that has just become more productive because somebody rolled well.
• ⚠ A GM who has just rewarded a special by opening the column up will feel that standing the line up is punishing the good roll. They will let it run, and the three minutes become ten.
• The fix is not to shorten the rest. It is to say what a special buys — and it should buy something portable rather than more conversation. Ivy giving up the almanac early, or a name, travels into Act Three; another ten minutes of talking does not.

The rest of the act

Braithwaite spotted the same six at the back. Pollard's Listen on the flank failed, so the thing keeping pace stayed unnoticed — no cost, the scenario is explicit that this is atmosphere and not a clue.

Act Two's recut is sound. Every beat ended on something written into the text rather than on GM judgement, which is what it was built to do and what the 65-minute version never had. Three minutes over on a hostile seed with a special in the worst possible place is a good result.

---

Act Three — There is no north

Planned 50 · ran ~45 · under by 5 (with the northern-seam cut taken, −15)

Renshaw's Insight on Ivy and Pollard's Spot on the lorry both landed, so the loop proved itself cleanly and the almanac pair did its work with no help from the GM. H03b beside H03a remains the best thing in the scenario and needed nothing said.

Then Bhattacharya fumbled Anomaly Lore on the depot entry, and the text has a clause for it — the agent concludes with complete confidence that the peg can simply be re-cut.

• ⚠ The fumble clause is excellent and it is a time bomb with no timing guidance. The document itself says it is "the most attractive wrong idea available, and a table will chase it". This table chased it for about ten minutes, which is the clause working as designed. But it sits in a 50-minute act that the Pacing Note has already cut by 15, and nothing tells the GM what to do when a fumble adds ten minutes back.
  • It came out even here only because the seam cut was taken. On a table that fumbles this roll and has not taken the cut, Act Three runs 60.
  • What it needs is one line saying how long to let the wrong idea live, and what ends it. The natural ender already exists in the text — the depot's paperwork is what proves cancellation is the only lever — so it is a matter of saying let them reach the depot with the wrong idea, and let the document be what corrects it.

---

Act Four — One of us

Planned 45 · ran ~45 · on budget

Braithwaite's Xenology (Baseline) failed — the party's one reliable test, missed. This is the single-point-of-failure the open questions have flagged twice and it finally came up in play.

• It played better than it reads. A party with a test that fails is not a party with no scene; they are a party that has to decide without one, which is the act's actual subject. Okonkwo's Spot landed and gave them a wrongness they could not justify, and Ashcroft's Insight failed, so the one person who might have named it did not.
• No timing cost. A failed test is faster than a successful one.

The unsettled countdown fired on its first ever test

Step 5, with Ashcroft on the Difficult branch because he accepted THE OFFER in Act Two: the understudy rolled 100 — a fumble — against Ashcroft's failure. Neither side succeeded. Unsettled.

• This is the 14.5% outcome, hit on the first roll it was ever asked for, and it is the case that did not exist before v0.12. The read-aloud is written and it landed: the wrong face for a moment, then a colleague again, apologising for the interruption.
• It is also precisely the aggressor-fumble case the rule's author raised and left to this document to price. The document prices it and the price is right. The reach fails, the row is spent, the clock still turns to step 6, and the party has seen the mechanism fail without being able to prove anything to anybody.
• ✔ No fix needed. This is the one thing in this pass that was written blind and worked first time.

---

Close — Cancellation

Planned 25 · ran ~28 · over by 3

Ivy at the stump, the question, the silence. The decision took longer than budgeted because Ashcroft had accepted THE OFFER and his player had to decide what to say about it, which is the private box doing its job at the exact moment it was built for.

No roll resolves this and none was made. The 2 Coherence for accepting landed at the Close as written.

---

Player-hat verdicts

Archetype	Verdict	Score
Methodical veteran note-taker	Act Two's rest clock read as fair because it was announced. Act One's collapse under two failures and a fumble was frustrating in the way the scenario intends	8/10
Roleplay-first talker	Had the special, had the column open up, and then the line stood up. "I rolled a 7 and got less time." The one sour note in the session	6/10
Quiet rules-literate tactician	The opposed roll coming back unsettled was the best mechanical moment of the night — a rule that admits nothing happened	9/10
Lore-digger	The re-cut-the-peg fumble gave them a theory to chase for ten minutes and the depot took it away with a document. Ideal	9/10

---

Timing summary

Act	Planned	Ran	Δ
One	45	~50	+5
Two	45	~48	+3
Three	50	~45	−5 (seam cut taken)
Four	45	~45	—
Close	25	~28	+3
Play	210	~216	+6
Break	10	10	—
Wall clock	3:40	~3:46	+6

Six minutes over, on a seed with two fumbles and a special in the worst place, with two of the three Pacing Note cuts taken. The third cut — the column's hospitality once rather than twice — was held in reserve and would have covered it.

That is the first time this scenario has finished a pass with a cut still in hand. The v0.13 recut did what it was for.

But the margin is six minutes, not twenty. Anyone reading the Pacing Note's "about ten minutes of genuine slack" should read this table first: on a hostile seed the slack is four minutes and one unused cut.

---

Fix list

Numbered, prioritised, nothing rewritten here.

1. ⚠ Say what a special on the column Persuade buys, and make it portable. (Act Two, the rest) A special currently buys more conversation, which fights the rest clock and makes a good roll feel like a penalty — the one sour note in the whole pass. Write it to buy something that travels: Ivy gives up the almanac early, or a name from the 1962 arrival, or the fact that six at the back do not eat. Then the line can stand up on schedule and the roll still paid.

2. ⚠ Give the re-cut-the-peg fumble a duration and an ender. (Act Three) The clause is one of the best things in the document and it adds ~10 minutes with no guidance. One line: let the wrong idea live until they reach the depot, and let the paperwork correct it. Without that, a table that fumbles here and has not taken the seam cut runs Act Three at 60.

3. ⚠ Write an answer for a fumbled Persuade on Swinburne. (Act One) H01's failure case is written for a party who did not ask, not for one that asked badly. Predates v0.13. Suggested shape: she does not produce the photocopy, she produces the dog. She takes them to where it died, which is better staging than the drawer and puts them at the stump without the note — and the note then reaches them through the coroner, late, when it confirms rather than reveals.

4. Correct the Pacing Note's slack claim. It says about ten minutes of genuine slack. This pass says four minutes and one unused cut on a hostile seed. State the hostile-seed figure, not the friendly one — the friendly one is the number a GM will plan against and then miss.

5. Keep, deliberately, the accident in THE OFFER's new placement. Running it inside the column scene means the aside has no audience and Ashcroft returns to a conversation that moved on without him. That was a side-effect of a timing change and it is thematically exact. Write it down before someone "fixes" it back.

Not a fix, recorded as a result: the unsettled countdown outcome, written blind in v0.12, fired on its first test via an aggressor fumble and needed nothing. Leave it alone.

---

POST-PASS — a correction to finding 2, found while applying it

Fix 2 above says the re-cut-the-peg fumble has "no timing guidance" and needs "a duration and an ender". That was wrong, and I only found out by going to the text to change it.

The clause already carries both. It reads: "a table will spend twenty minutes on it. Let them try it exactly once. Cutting the notch does nothing whatsoever." Twenty minutes is the duration and trying it once is the ender. Both were there before this pass ran, and my own simulation costed the chase at ten minutes — half what the document predicts.

So the finding was right about the symptom and wrong about the cause. The problem is not that the fumble is unbounded. It is that the document knows the fumble costs twenty minutes and the timing table does not: Act Three is budgeted at 50, the text predicts 50 + 20, and nothing connects the fumble to the northern-seam cut that would pay for it.

Applied as that instead: a timing warning at the head of Act Three saying the twenty minutes exist, are deliberate, and make the seam cut compulsory the moment the fumble lands. The clause itself is untouched — it was already right.

Kept as the pass's own lesson. Pass 4 read that clause, formed a view about what it lacked, and did not re-read it before writing the finding down. The finding then sat in a committed document until the fix went to apply it. A playtest is a reading of the text, and a reading made while looking for faults finds faults that are not there as readily as ones that are — which is an argument for applying fixes against the source rather than against the notes.

---

POST-PASS 2 — six of this pass's sixteen ratings were wrong

Found while setting up desk pass 5, by pulling the four cut sheets out of roster.mjs instead of typing the numbers from memory. This pass typed them from memory.

Roll	Used	Actual	Effect
Pollard · Spot, the stump	58	53	success → fail
Okonkwo · Spot, treeline	52	35	none, fails either way
Braithwaite · Spot, the six	48	35	none, succeeds either way
Pollard · Listen, the flank	55	53	none
Bhattacharya · Anomaly Lore, depot	58	63	FUMBLE → plain failure
Braithwaite · Xenology (Baseline)	45	53	fail → success

The d100s are unchanged — they come from the seed — so this is purely a matter of what they were compared against.

Three things this pass reported did not happen.

1. The re-cut-the-peg fumble never fired. 98 against Anomaly Lore 63 is a plain failure; 96–99 never succeed and only 00 fumbles. So the ten minutes this pass spent chasing the wrong idea, and the whole of finding 2, rest on a fumble the seed did not produce.
2. Braithwaite's Xenology succeeded. The party's one reliable Act Four test did not miss. This pass's claim that the single-point-of-failure "finally came up in play" is false, and it remains untested after five passes.
3. Pollard missed the stump, making Act One worse rather than better.

What that does to the timing verdict. Without the fumble chase, Act Three runs about 35 with the seam cut rather than 45, and a successful Xenology is marginally faster than a failed one. The same seed, read correctly, lands near 3:34 — six minutes under budget rather than six minutes over.

What survives. Findings 1, 3 and 5 are untouched: the special on the column Persuade was a real 7, Renshaw's fumble on Swinburne was a real 00, and THE OFFER's staging is not a dice question at all. Finding 2's conclusion also survives, because it never depended on the roll: Act Three is budgeted at 50, the text itself predicts the fumble costs 20, and nothing connected that to the cut that pays for it. The reasoning was right and the evidence was invented. The scenario has been corrected to say so rather than citing a playtest that did not happen.

The lesson, which is the same one twice in one document. Post-pass 1 recorded that this pass formed a view about a clause without re-reading it. This is that again at the level of data: sixteen ratings, six of them wrong, none of them checked against the sheet they came from, in a document whose entire authority is that the rolls were honest. A playtest that invents its own target numbers is not a playtest, it is a story about one — and it is worth noticing that the errors were not random. Every wrong value was close to the real one, which is exactly why nothing looked wrong on the page.

Desk pass 5 pulls every rating from roster.mjs before rolling.