59 lines
3.6 KiB
Markdown
59 lines
3.6 KiB
Markdown
# Faction allocation in the lobby
|
||
|
||
## Sources and scope — 28 October 2026
|
||
|
||
**Prototyped / Partial.** The user-authorized revised Dune rulebook, physical
|
||
page 4 (printed page 2), gives two initial methods: randomly deal faction
|
||
prediction cards, or draw distinct turn-prediction numbers and select factions
|
||
from highest to lowest. It also permits players to swap factions. The original
|
||
publisher rulebook describes the same selection methods; this prototype exposes
|
||
both in Basic and Advanced and retains the existing direct-choice option.
|
||
|
||
A host may deal once from all factions belonging to currently enabled families,
|
||
without replacement. Alternatively, the host deals distinct numbers 1–10 to
|
||
seated faction seats. The public draft follows descending numbers, rejects a
|
||
picked or disabled faction and applies all choices together after the last pick.
|
||
These cards represent the setup draw and are not removed from later Bene Gesserit
|
||
prediction. Existing temporary faction selections do not reserve faces during a
|
||
draft. The host may cancel the draft; the previous factions remain. The engine
|
||
uses the existing cryptographic shuffle, event-bound actions and room version
|
||
fencing, and clears human readiness when allocation begins and ends.
|
||
|
||
For the provisional two-person/four-faction setup, each faction seat receives
|
||
its own draw and pick. That is a software interpretation of the two-faction
|
||
case, not a separate publisher clarification. Seats, credentials, host and
|
||
player-circle positions remain attached to the same participants. AI seats pick
|
||
a legal available faction automatically, preferring their existing faction
|
||
when available. They retain their difficulty and a faction-appropriate name.
|
||
Selection does not bypass any existing expansion/profile start gate.
|
||
|
||
Nine related unit files pass 43 cases; four isolated HTTP cases include a
|
||
four-seat/two-controller draft, seat selection and game start. These do not
|
||
establish complete setup or deployed acceptance. An isolated browser room also
|
||
exercised a random deal and a human/AI draft to completion.
|
||
|
||
Human seats may offer an exchange of their currently assigned factions. The
|
||
recipient accepts or declines and the proposer may cancel; each action is bound
|
||
to the current offer. A pending offer freezes other lobby changes and new seats
|
||
until resolved, but players can switch between their own seats or exit. Only
|
||
acceptance changes the assignments and clears human readiness. Seat IDs,
|
||
credentials, circles and host remain attached to their participants; faction
|
||
leaders and Ixian elites follow the new faction. A human may offer to swap
|
||
with an AI seat; the AI immediately accepts as a provisional software policy.
|
||
The AI seat keeps its difficulty, identity and controller, while its name and
|
||
faction assets follow the new faction. AI-initiated offers remain unfinished.
|
||
Broader player-count interpretations, AI strategy,
|
||
complete setup combinations, saved-game assurance and deployed acceptance
|
||
remain open.
|
||
|
||
For the swap checkpoint, ten related unit files pass 48 cases and two isolated
|
||
HTTP cases exercise allocation and authenticated swaps across independently
|
||
controlled seats. This evidence does not certify every setup combination.
|
||
An isolated browser room also exercised an offer between two controlled seats,
|
||
including switching to accept, cleared readiness and an available exit control.
|
||
|
||
The AI-target follow-up passes 50 cases in ten related unit files and three
|
||
isolated HTTP cases covering allocation and authenticated human/AI swaps.
|
||
An isolated browser room confirmed an immediate human/Hard-AI swap, retained AI
|
||
difficulty and the human readiness reset.
|