SirStone/SirRoboGarage
1
0
Fork 0
Code Issues 28 Pull Requests Actions Projects Releases Wiki Activity

movement final: pre-register the ship criterion before the confirmation battles

Browse Source
This commit is contained in:
Davide Cappellini
2026-09-26 02:21:09 +02:00
parent 47ce244740
commit ff03e81591
2 changed files with 71 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/movement_campaign.md
+32
View File
@@ -1164,3 +1164,35 @@ the default flip is a separate, explicit decision and was NOT made here.
|---|---|---:|---|---|
| `/tmp/ab/j119_b3` | `1256357` | 270 (0 failed; 1 excluded: Ascendant/strafe r1) | strafe, tfil, fast_flip, slow_flip, wide_spread, narrow | nothing beats strafe; wide_spread +0.11 NS (p=0.55) |
| `/tmp/ab/j119_b4` | `1256357` | 270 (0 failed, 0 excluded) | strafe, tfil, bullet_strong, field_strong, field_off, wall_tight | nothing beats strafe; field_strong and field_off detectably WORSE |
---
## Final confirmation + SHIP — PRE-REGISTRATION
> **PRE-REGISTERED BEFORE THE BATTLES (this section is committed, and its commit
> is the frozen binary's source, before the first battle is launched).** Session
> `/tmp/ab/j120_final`. Arms file `tools/ab/arms_movement_final.txt`, panel
> `tools/ab/panel_movement.txt` (frozen), **3 arms × 15 opponents × 5 runs × 3
> rounds = 225 battles**, conc 6, `--reference tfil`.
**THE PRE-REGISTERED SHIP CRITERION (fixed before fighting).** Ship the flip of
`TR_MOVEMENT`'s default from `tfil` to `strafe` **only if**, on this frozen
panel and this one frozen binary, `strafe` beats `tfil` head-to-head on the
PRIMARY metric *round wins/run* with **both**:
1. the **95% CI on Δwins/run excluding 0**; **and**
2. the **cross-opponent sign test favouring `strafe`** (p < 0.05).
This is stricter than the campaign's §1 rule 2 (which would let a survival win
carry a damage-neutral arm): the ship decision requires the *win* metric itself
to clear both bars. If it does not pass, **do not ship** — say so and stop; that
is a fully successful outcome. This batch uses **5 runs/arm (vs 3 in every
previous batch)** for higher power; `wide_spread` (the only positive point
estimate from Batches 3–4, +0.11/run, NS) rides along as the single challenger.
**Pre-registered prediction (before the battles):** `strafe` passes both bars
and ships; `wide_spread` ties `strafe` (no separable win); `tfil` is not
preferred by any primary metric. A wrong prediction is recorded as wrong.
**Outcome:** pending — filled in below after the session.