Election Mechanics

Elections in A House Divided run perpetually. Every office has continuous cycles: a cycle resolves, the next one spawns immediately, seats never sit vacant. This page is the top-level reference for how elections work — phases, timing, scoring, vote math. Deep dives for primaries, generals, and country-specific rules get their own pages.

The Two Phases

Every election has two phases:

1. Primary phase — intra-party competition. Candidates from the same party compete for one nomination slot. Resolution uses a primary score (no votes accumulate). Highest score per party advances.
2. General phase — inter-party competition. Nominees from different parties compete. Votes accumulate turn by turn; the final 4 turns land 25% of the pool.

When the general resolves, winners take office and the next cycle spawns.

Duration by Race Type (US)

All durations are in real hours = game weeks.

|| Office | Total | Primary | General | | --- | --- | --- | --- | | House | 96h (2 game years) | 48h | 48h | | Senate | 96h (2 game years) | 48h | 48h | | Governor | 192h (4 game years) | 144h | 48h | | State Senate | 192h (4 game years) | 144h | 48h | | President | 192h (4 game years) | 144h | 48h |

Senate classes are staggered: Class 1 -> Class 2 (+32h) -> Class 3 (+64h), so one-third of the Senate is up every ~1.3 real days.

UK, DE, and JP use different cycle lengths — see each country hub for specifics.

The Primary Score (State Races)

Primary scores go up to 100 points across four components, then a final infamy penalty applies:

Final score = raw * (1 - 0.05 * infamy/100) — a 5% reduction at infamy=100.

Of the 40 alignment points, 25 measure how closely you match your state's politics and 15 measure how closely you match your party platform. State match is the bigger factor. If a state has no cached lean, alignment falls back to a single 40-point check against the party (preserving the pre-rework formula).

NPP candidates get a 50% primary-score multiplier when at least one player is in the same party's primary — a 50-point NPP lands at 25, which makes competent player candidates heavy favorites. NPPs aren't subject to the infamy penalty.

The Primary Score (Presidential)

Presidential primaries are national, so they weight party platform and national reach instead of state position:

Same infamy penalty applies on the final score. Presidential primaries reward NPI accumulation and home-state party-org investment, where state primaries reward favorability, in-state political influence, and state-level policy fit.

Presidential primary stagger waves (the final 6 turns of the primary) use a delegate accumulator. NPPs in stagger waves receive an extra 0.6 multiplier on top of the usual handicaps when a player is in the same party.

Vote Accumulation (General)

During the general phase, votes accumulate each turn via the Total Appeal Pipeline (also used by polling and NPP dropout calculations).

Per-turn weighting

• Early turns: 0.75 * totalPool / (totalTurns - 4) per turn.
• Final 4 turns: 0.25 * totalPool / 4 per turn.

In other words, the last 4 turns are worth ~25% of the entire election. If you're within 3% going into the last day, every campaign and ad action in those 4 turns compounds into the heaviest-weighted snapshots.

The Total Appeal Pipeline

For each demographic group, for each candidate, per turn:

1. Reach — normalizeNPI(influence). Sqrt curve, capped at 1.0 once influence reaches 100. State and presidential general elections both use this curve. (Presidential primaries use a diminishing-returns curve 1 − exp(−NPI/45) instead — leaves more room for Support, Favorability, and Org to compete.)
2. Appeal (per group) — (50 - |econDiff|*5 - |socialDiff|*5)^2 / 100 + APPEAL_POSITION_FLOOR(0.5) + DIRECTION_BONUS(tribal, up to +30) + normalizeNPI(influence) * 12.5. Quadratic position scoring with a floor, directional bonus, and influence multiplier. Max ~55 in both state and presidential general races.
3. Approval scalar — effectiveFavorability / 100, where effectiveFavorability = favorability + archetypeApproval * 0.5 (clamped 0–100). 0% effective favorability = 0 votes.
4. Party org scalar — orgVoteWeight() normalized share model: (partyOrg / totalStateOrg)^0.5. Ranges from near 0 (tiny org share) to 1.0 (monopoly). Independents use a flat 0.5. Presidential races use the same normalized share, damped when the field is crowded: max(0.2, 1 - (n-2) * 0.15).
5. Infamy scalar — 1 - 0.05 * (infamy/100). Player characters with high infamy lose up to 5% of their per-group weight. NPPs aren't affected.
6. Turn pool scaled by party strength modifier:
   • State races: (1 + (stateGovernmentApproval/100 - 0.5) * 0.2) * officeStrength. At 50% approval the modifier is 1.0x; at 100% it is 1.1x; at 0% it is 0.9x.
   • Presidential races: (1 + (stateGovernmentApproval/100 - 0.5) * 0.5) * officeStrength. At 50% approval the modifier is 1.0x; at 100% it is 1.25x; at 0% it is 0.75x.
7. Group-level allocation — Each group contributes to the turn pool proportional to its size. Within each group, candidates split that contribution by relative (appeal * reach * approval * partyOrg * nppPenalty * infamyMult).
8. Votes summed across all groups. That's your turn total.

Office strength multipliers:

• Governor: 1.0
• State Senate: 0.85
• House: 0.9
• Senate: 0.8
• President: 1.0

Because the approval modifier is a flat +/-10% around the base for state races and +/-25% for presidential races, presidential elections feel state approval much more strongly.

NPP General Weight Penalty

In general elections where at least one player is in the race, NPPs receive an 0.8x weight penalty on their per-group vote weight. This is separate from the 0.5x primary-score penalty and applies only during the general phase. It ensures NPPs remain competitive but lose their structural advantage over active players.

FPTP vs RCV — The Spoiler Effect

Most states use First Past the Post (FPTP). A few run Ranked Choice Voting (RCV) — this is a state-level legislative choice.

In FPTP states, after the per-turn group allocation:

• For each third-party candidate, 4% * (their group-level votes) is drawn from the ideologically nearest major-party candidate and transferred to the third party.
• In presidential races, the rate is halved to 2%.
• "Nearest" = Manhattan distance on the econ/social grid.
• Major parties are determined per region: Democrat/Republican in the US, Labour/Conservative in England, SNP/Labour in Scotland, etc.
• The transfer is organization-aware: a third party with stronger state org than the nearest major party amplifies the bleed (factor 0.25x–2x).

This models real-world spoiler dynamics: a Green candidate bleeds Democratic votes, potentially handing the seat to the Republican.

In RCV states, no spoiler adjustment applies. Third parties compete on equal footing. Switching a state from FPTP to RCV requires legislation.

Reg, Support, and Swing-Flow Mechanics

The general-election pipeline includes three swing-flow factors that modulate how votes move between parties:

1. Registration resistance multiplier — regResistanceMultiplier = 1 + 0.3 x (Reg / 100). Range 1.0x-1.3x. Higher own-registration makes the party harder to peel away through persuasion. Independents and parties without a registration entry get neutral 1.0x.
2. Support mood multiplier — supportMoodMultiplier = 0.6 + 0.8 x (support / 100). Range 0.6x-1.4x with neutral 1.0x at support=50. Captures short-term candidate mood / momentum (debate performance, scandals, endorsements). New candidates without a stored support value default to 1.0x.
3. Transferable share — The fraction of a party's vote pool that is open to persuasion in a given turn. Base transferableShare is scaled by regResistanceMultiplier (higher Reg = smaller transferable pool) and candidate persuasionResistance (some candidates have innate resistance to being peeled away from their base party).

These factors mean a candidate with high Support, low opponent Registration, and low persuasion resistance can swing larger vote shares per turn than one facing entrenched opposition in a high-Reg state.

Multi-Seat Races

House and State Senate (US), Bundestag constituency seats, UK Commons regions, and JP Shugiin use proportional allocation:

• Largest-remainder method distributes seats proportional to vote share.
• Minimum threshold: 20% of votes for House and Commons; 10% for State Senate and Regional Council (larger districts with more parties splitting the vote).
• 2-seat special case (House only): winner takes both unless the runner-up reaches the threshold.
• Seats estimate: updated each turn — voteFraction * totalSeats.

Candidacy Rules

• One race at a time. Active candidacies block new declarations (cancel first, then re-enter).
• Home state / region only. US home state, UK home region, DE home Land, JP home prefecture — except President, which is national.
• Party required for primaries. Independents run in the "independent" party primary (same rules as any other party). No general-phase bypass.
• Country restriction. Your countryId must match the election's countryId; cross-country entry returns 403.
• Primary deadline. Declarations close at primaryEndTime. Miss the window and you wait a full cycle.

Withdrawal

You can withdraw any time before the election resolves:

• Candidate status -> withdrawn, withdrawnAt timestamped.
• Your votes are permanently removed from the tally (not redistributed).
• Your campaign document is deleted.
• You cannot re-enter the same election.

Withdrawal during a general you're losing is sometimes strategic — it frees actions for a better race next cycle. During a primary it's almost always wrong unless you're pivoting to a different office.

Suspend & Endorse (Presidential General)

During the general election only, a presidential nominee may suspend their own campaign from the Campaign Manager page (/campaign/[id]) and endorse another active nominee:

• You stay on the ballot — this is not a withdrawal.
• Campaign operations stop: turn income, upgrades, rallies, travel, canvassing, contributions, and queued Support drips halt.
• One-time campaign-strength transfer at the moment of endorsement: 25% of your campaign strength is debited from you and credited to your endorsed nominee. This is a single transfer — no per-turn campaign-strength transfers occur after this.
• Org boost: 25% of your per-state character org investment (primary-cycle characterStateOrg levels — not the shared statePartyOrg.organization party-machine scalar) is debited from you and added to your endorsed nominee's effective org for vote distribution in each state, for as long as you stay suspended.
• Your existing vote tally is preserved — votes already earned across every state stay counted. You remain on the ballot but accumulate no further votes in any subsequent turn.
• Suspended nominees forfeit all passive bonuses (swing-state ground game, VP home-state effect, campaign strength multiplier).
• Irreversible for the rest of the general. Nominee-only (not campaign manager).
• If your endorsed candidate withdraws, transfers are already complete and you remain frozen on the ballot — no further action needed.

Perpetual Continuity

No seat sits vacant. Mechanics:

• House: Next cycle spawned immediately on resolution.
• Senate / Governor / State Senate: Perpetual check each turn; replacement spawned for any slot with no active/upcoming race.
• Governor bootstrap: If no state has ever completed a governor race, one is spawned per state so cycle 1 exists.
• President: Spawned via the canonical LARP schedule when no active/upcoming president race exists; the cycle window plus a 24h primary / 24h general floor keep spawns aligned to real-world presidential election years.
• New elections inherit duration from the most recently completed equivalent race.

Presidential Election Specifics

• National race. Any home state can run.
• Electoral College (538 votes, 270 to win). ME and NE split by congressional district (UNIT_LEAN modifiers).
• Independent penalty: 0.3x vote share on the general path (70% reduction).
• Running mate (VP): After the primary, each nominee picks a VP. Cannot be the current President; cannot be the same person.
• Presidential Travel: Active candidates can travel to a US state during the general. Cost is 3–10 actions based on the state's electoral-vote count (<=5 EV = 3 actions; <=10 EV = 5; <=20 EV = 7; >20 EV = 10). +1% Favorability per turn while in-state, one state at a time.
• Governor Endorsement: A sitting Governor who endorses a presidential candidate in their home state applies a 1.5% bonus (GOVERNOR_ENDORSEMENT_STATE_BONUS = 1.015) to that candidate's vote total in the state.
• No EV majority: If no candidate reaches 270 (including a 269–269 tie or a third party blocking a majority), a contingent election runs. The House elects the President from the top three EV finishers (one vote per state delegation; DC has EVs but no House vote; 26 states needed). The Senate elects the Vice President from the top two running mates (51 votes needed). Representatives and senators vote primarily by party, then ideology proximity; tied state delegations abstain. Chamber composition is snapshotted when the contingent ballot first runs so same-turn House/Senate flips cannot change the ballot mid-resolution.
• MVP ballot: Contingent races resolve on a single simulated House/Senate ballot per cycle. If neither chamber reaches its threshold, a deterministic deadlock breaker seats a winner (multi-turn contingent ballots with an acting VP are planned for a later version).
• Party-weighted positions: In presidential general elections, candidate positions are blended toward their party platform with weight 1/3 ((partyPos + 3 * charPos) / 4). This keeps nominees aligned with their coalition while preserving individual identity — 75% candidate, 25% party.
• Swing-state ground game: In swing states (|lean| < 0.5 for modern elections, |lean| < 1.0 for legacy cycle-1 elections), candidates with invested ground game receive +3% votes per ground-game level.
• VP home-state effect: A nominee whose running mate hails from the current state receives +3% in that state. Stacks with ground game.
• Campaign strength multiplier: Player campaign contributions build campaign strength, which applies a multiplicative vote boost. Each contribution adds 0.75 * NPI strength. The curve is 1 + 1 * (1 - exp(-strength/50,000)), soft-capping at +100% (2x total votes) at very high strength. Each campaign turn also pulls the strict campaign-strength leader in an active race back toward that race's average by up to 175 strength.
• State lean multiplier: Each electoral unit applies a lean multiplier: 1 + lean * epSign * leanStrength (leanStrength = 0.3 for ME/NE districts, 0.25 for full states). The multiplier is floored at 0.5x so unfavorable lean never zeroes out a candidate entirely.

Polls vs Simulated Votes

• State races (House, Senate, Gov, State Senate, Commons, Bundestag, Sangiin, Shugiin): player polls use the same group-level competitive allocation and FPTP spoiler rules as real vote accumulation. Poll numbers reflect the actual math.
• Presidential: simulated votes accumulate per electoral unit using a different model (party-weighted positions, steep party-org curve, state-lean, swing-state ground game, FPTP spoiler not applied). Player polls remain home-state projections only and should not be treated as Electoral College forecasts.

Related

• Primaries — Declaration rules, NPP dynamics, primary-phase tactics.
• General Elections — Vote accumulation, final-turn weighting, tie-breakers.
• Campaign Strategy — Tactical decision-making.
• Primary vs General Tactics — How to pivot between phases.
• Demographics & Targeting — Appeal math and group composition.
• Snap Elections — Parliamentary dissolution in UK, JP, and DE.
• Formula Deep-Dive — Every formula with derivations.