# PAGE 62 SCORING: how every kill is scored (coordinator, 27 Oct 2026)

The user (inbox #34): "Brutality and cinematicity, nice script is very low in so many of the weapons ... quality of brutality and
cinematic scripting is worse in the end. Especially mini gun animation and scripts are so bad ... low brutality instead of extreme.
There is a lot of weapons like this. I only seen the starting stage weapon got good animation, after that quality getting dropped.
Score brutality and cinematic scripting rating for all animation as of now and yet to build, create a page 62 with these scores so that
I can go and check the video and tell you are correct or not. Then let's think of rescripting."

The user WILL check our numbers against the videos. The point is to find the weak kills, not to defend them. No inflation.

## FOUR scores since 27 Oct 01:30 (the user: "i need two more metric of rating in the 62: Script, Animation")
Each kill now gets BRUTALITY, CINEMATIC, SCRIPT and ANIMATION, each 1-10 with its own one-line reason:
- SCRIPT: the idea and story as written: how original the concept is, how clever the setup and payoff are, how well it fits the
  weapon, how memorable it is. Judge the concept even if the execution is poor. Scripts that aren't built yet get a SCRIPT score too.
- ANIMATION = ANIMATION DETAILING (the user, 27 Oct 01:40: "by animation i mean animation detailing, i think we lack detailing"):
  how much detail is drawn and animated: wound and gore anatomy (fibres, bone, organs, wet highlights, torn edges), faces and
  expressions, bodies (clothes, hands, secondary motion), props, debris and fx detail, the set and lighting detail. Flat shapes and plain
  blobs score low. Built kills only (null for scripts). The minigun re-rated on detailing averages 4.0.
- CINEMATIC now means camera and staging only: shots, angles, cuts, slow motion, framing, pacing. The story belongs to SCRIPT.
- BRUTALITY is unchanged.
Schema additions: `script`, `script_why`, `anim` (null if not built), `anim_why`. Calibration examples: _tools/scores/minigun.json
(minigun averages: script 5.4, animation 5.0).

## The original two scores (integers 1-10)
- BRUTALITY: how extreme and horrifying the XT level is: how bad the wound is, its anatomy and detail, how the body and face react, the
  aftermath (blood, pieces, the dead face). A tidy hole and a pool is 3-4. Visceral, specific, memorable horror is 9-10.
- CINEMATIC SCRIPTING: story, setup → escalation → payoff, camera (purposeful angles, cuts, close-ups, slow motion at the right moment),
  timing and pacing, readable staging, use of the arena, how unique the idea is. One static side shot with no build is 2-4. A mini film
  with a hook, tension and a perfectly timed payoff told by the camera is 9-10.
- Same absolute scale for films, skins and mids. Hold each tier to its tier: a weak MYTHIC or LEGENDARY says so in its reason.

## The scale (calibration: look at these before you score anything)
- Anchors (early weapons the user liked): img/62/_calib/ar.last-stand.jpg ≈ 8 / 8, sawedoff.gibbet-road.jpg ≈ 8 / 8,
  ar.human-shield.jpg ≈ 7 / 7.
- The minigun, scored by the coordinator: _tools/scores/minigun.json and its sheets img/62/minigun/*.jpg. Average 5.0 / 5.0; RELOAD 6/7,
  HEADS UP 7/6; THE PRESS 3/4, SOLID 3/4, THE CIRCLE 4/3. Open 3-4 of those sheets next to their scores before you start.
- Bands: 1-4 weak (rework), 5-6 average, 7-8 good, 9-10 extreme (the bar).

## How to score one weapon
1. The kill list: every ledger line for the weapon (`grep -oE '^- <w>\.[a-z0-9-]+' _tools/58k_LEDGER.md`): films, skins, mids, built or not.
2. BUILT kills, a filmstrip each:
   `RENDER_SLOTS=4 bash _tools/render_slot.sh node _tools/ix_film.mjs 58k_<w>.html <id> /Volumes/Workspace/Scratch/scores/<w>/<id> 12 844 390`
   then the contact sheet:
   `python3 /Volumes/Workspace/Scratch/scores/tools/sheet.py /Volumes/Workspace/Scratch/scores/<w>/<id> img/62/<w>/<id>.jpg "<ledger title> · <len> s"`
   (`{"error":"no kill ..."}` in run.json = not built: score it from its script instead.)
   The strip runs from gameplay to the ROUND OVER card. The held final is around frames 9-11; the last frame is the recap card over it.
3. LOOK at every sheet yourself (Read the jpg) and score what is actually on screen, not what the script promised.
4. NOT BUILT kills: score the ledger script (CORE / MOTION / CAMERA / FINAL) with status "script", and say so in the reason.
5. Write `_tools/scores/<w>.json`, an array with one entry per kill:
   `{id, weapon, title, tier, kind: "film"|"skin"|"mid", status: "built"|"script", brut, cine, brut_why, cine_why, fix, scorer, sheet, len}`
   - brut_why / cine_why: one concrete line each about what is on screen (about 15-25 words).
   - fix: the single biggest change that would take it to 9+.
   - sheet: "img/62/<w>/<id>.jpg" (null for scripts); len: the film length from run.json (null for scripts).
6. `python3 _tools/make_62.py`, then a one-line report to the coordinator: weapon, kills, average BRUT / CINE, the worst 3.

## Rules
- Never score a weapon you built.
- Score the kill as it is now. Note rule breaks you spot (no groin wounds, no neck hanging, no puff VLs) in the fix line.
- Building is paused: don't edit any 58k page while scoring.
- Every Chrome run goes through render_slot.sh. Delete nothing; scratch goes under /Volumes/Workspace/Scratch/scores/<w>/.

## MID-ROUND KILLS: score brutality STRICTLY (the user, 27 Oct 01:35)
The user: "the mid round kills is having very low brutuality and gore, but u put good score for those". The coordinator's first minigun mid scores were too generous and are now lowered (2-4). A mid is short and small in frame: judge the gore the player actually sees in the zone. Most current mids are 2-4 on BRUTALITY unless the wound is big, bloody and held on screen. Same absolute scale as the films; never grade a mid on a curve.