CLEAN GROUND v0.11: Dodge 68, and Ashcroft is busy rather than sidelined

Two holes a peer session's per-agent measurements pointed at. Their scratch
figures are not used; both edits are sourced to published numbers or to the
guard that now holds them.

EXPOSURE gains a fifth point: the party's sheets read about twice as effective
as they are, and the number doing it is the column's Dodge 68. That is
arithmetic off two published values, not a measurement -- Dodge 68 on The
Quiet Neighbours (tools/content.mjs, printed by the bestiary) and one defence
spent per incoming blow, so a 40% attack against a fresh dodge connects about
one time in eight. Nothing in the document said so, and grep for "expect to
hit" / "hit rate" / "lands about" returned nothing at all.

The point carries its own way out: DEFENCE_STEP is -30, so a defender who has
dodged once meets the next blow at 38%, then 8%, then the floor. Spend their
dodges rather than out-shooting them. That is also the other half of why the
four-player cut is a different game -- three attackers cannot spend three
defenders' dodges and six can, so the missing attacker is the one who would
have made everybody else's shots land.

The recorded landing rates agree and are now citeable (16-18%, above the naive
one-in-eight because most blows are not the first of their round), so the
sub-bullet cites tools/attackers-baseline.json rather than asserting.

The scaling note's Ashcroft warning gains the sharper reading: he is not
sidelined, he is busy and ineffective -- 9.2 attacks a fight of which 1% land
and 0.1% disable anybody, both held by check-attackers. A player who cannot
act knows to do something else; a player rolling once a round for an hour
thinks he is fighting. So the note now says what to give him instead, using
only what is on his sheet: First Aid 40, Dodge 63, Insight 63, the anchor.

Verified rather than relayed: Dodge 68 at tools/content.mjs, DEFENCE_STEP -30
and the per-round reset in rules.mjs and simulate.mjs, the three fighters'
skills at 40/40/35, and the per-attack rates re-measured independently across
three seeds before the baseline existed (17.6-18.3 / 17.4-17.6 / 15.5-16.2 /
1.0%, against the recorded 17.8 / 17.3 / 16.2 / 1.0).

npm run check: 14 guards, exit 0.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>

docs/scenarios/CLEAN_GROUND.md

@@ -689,7 +689,7 @@ node tools/simulate.mjs --creature the_understudy \
 range bands or fire modes, and every one of those makes a fight worse for whoever is
 losing. Both of the first two are live in this scenario.
 
-#### The four things a GM needs from that table
+#### The five things a GM needs from that table
 
 1. **Do not fix the understudy.** It is a 1d3 punch with a 60% chance to land, and it goes
    down in two rounds. That is correct and it is the point: the horror is that it passed
@@ -749,6 +749,32 @@ losing. Both of the first two are live in this scenario.
    the shape aloud, not the average:** "nearly half of all parties die" is the honest
    warning, and "you will probably lose two people" is not.
 
+5. **Their sheets read about twice as effective as they are, and the number doing it is
+   Dodge 68.** This one is arithmetic off two published numbers, not a measurement. The
+   column dodges at **68%** (`The Quiet Neighbours`, `tools/content.mjs`; the bestiary
+   prints it), and a defence is spent **per incoming blow** — so a skill roll is only the
+   first half of the test. A 40% attack against a fresh dodge connects about **one time in
+   eight** (0.40 x 0.32), against a sheet that reads 40%. The three who can fight in the cut
+   carry Firearm and Melee between 35 and 40. **Nobody at the table will guess that from
+   their own sheet**, and the same gap is Ashcroft's defect one level less severe: the
+   number on the page is not the chance of doing anything.
+   - **The way out is to spend their dodges, not to out-shoot them.** Each defence after the
+     first in the same round is at **-30** (`DEFENCE_STEP`, `rules.mjs`), so a defender who
+     has already dodged once meets the next blow at 38%, then 8%, then the 1% floor. The
+     second agent to connect in a round is worth more than the first — that is what
+     `check-focus` measures across the bestiary, and it holds here.
+   - **Recorded, and it agrees:** the three land **16 to 18%** of their attacks —
+     Bhattacharya 17.8, Renshaw 17.3, Braithwaite 16.2 — which is *above* the one-in-eight
+     the arithmetic gives for a fresh dodge, precisely because most blows in a fight are not
+     the first of their round. `tools/attackers-baseline.json`, held by `check-attackers`.
+   - **This is the other half of why the cut is a different game.** Three attackers cannot
+     spend three defenders' dodges; six can. Point 4 says the cut is short an attacker.
+     Dodge 68 says what that costs — the missing attacker is the one who would have made
+     everybody else's shots land.
+   - **Do not fix it by lowering the dodge.** Their 68% is why Act Two's answer is medicine
+     passed hand to hand with both parties off the ground rather than marksmanship. The
+     column being uncatchable is the point; see their tactics line.
+
 #### What to do instead of a fight
 
 - **Act Four's tension is a test, not a threat.** The question is which of two people is
@@ -1032,6 +1058,15 @@ twice.*
     Braithwaite on a knife. **That is a quarter of the party out of any fight**, and it is a
     large part of why EXPOSURE says violence at four players is party-ending. Substituting
     Agyeman fixes the arithmetic as well as the lane.
+    - **He is not sidelined in the fight — he is busy and ineffective.** He takes his
+      actions — **9.2 attacks across a fight, of which 1% land** (`check-attackers`) — and
+      **one attack in a thousand disables anybody**. That plays worse at the table than
+      being obviously out of it: a player who cannot act knows to do something else, and a
+      player rolling once a round for an hour thinks he is fighting. **If a fight starts and Ashcroft is in it, give him something that is not
+      an attack roll** — First Aid 40 on whoever went down, Dodge 63 to get to them through
+      the line, Insight 63 on whether the column is still shooting to kill or only to be
+      left alone, and the anchor. His entry lists the same four; this is where they earn
+      their keep.
 - **Never drop Braithwaite.** Xenology (Baseline) is the only test that works. If a
   player insists on a sheet that hasn't got it, hand the lane to **Dr Chidi Nkemdirim**
   `pc_nkemdirim` and accept that Act Four gets harder, or give the understudy a visible