Arena round 6: revert railgun_m damage, correct playtest record

Arena re-run after the playtest-1 adjustments:

- spam-cruiser arena landed at par (default +9%) - the range buff alone
  priced out the small-gun-spam meta, so revert railgun_m damage 16 -> 14
  (range stays 80). The buff had regressed drone-swarm-vs-cruisers to
  +47%, largely via a hit-count breakpoint (60 HP drone: 5 hits at
  14 dmg, 4 at 16).
- battleship +30% / dreadnought +29% vs pure railgun_s fleets accepted
  as reach-doctrine texture (the l gun's 130 m standoff working as
  intended); repair kept at 4 HP/s (below par in-fight is the correct
  price for free between-wave sustain).

Also corrects the playtest-1 record: it was two FULL playthroughs won in
~40 min each with cruisers only (not two pushes) - a 2.5-3x run-length
gap. Pacing knobs deliberately deferred to playtest 2, since those runs
predate the repair nerf and the l-gun siege range.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DyCu8vwChKMbLJQ3xosYEN

docs/balancing/README.md

@@ -21,11 +21,12 @@ REQ-* ids in [../requirements.md](../requirements.md).
 
 First full balancing round complete (2026-07-06): targets → tree →
 numbers → threat-calculator parity → combat stats (arena-converged) →
-pacing. Playtesting in progress: playtest 1 (first two pushes) found the
-railgun_s-spam-cruiser meta and confirmed repair as overpowered — combat
-stats adjusted, see `history.md`. Next: re-run the arena suite against
-the adjustments, then continue playtests against the run-shape targets
-in `targets.md`.
+pacing. Playtesting in progress: playtest 1 (two full ~40-min wins,
+cruisers only) found the railgun_s-spam-cruiser meta, confirmed repair
+as overpowered, and exposed a 2.5–3× run-length gap; combat stats
+adjusted and re-checked in arena round 6 (see `history.md`). Next:
+playtest 2 with the new stats — pacing knobs (station scaling, threat
+rate, win pacing) wait for its result.
 
 ## Open action items