@mcptoolshop/sovereign 1.1.0 → 1.1.2

package/README.ja.md

@@ -19,58 +19,17 @@
 
 ---
 
-## 概要
-
-Sovereignは、**アメリカの公的信用に関するハミルトン・システムを基にしたボードゲーム**であり、さらに、**完全なソロ/デジタル版**も付属しています。このデジタル版は、同じルールに基づいて、ブラウザ上で2つの決定的なAI対戦相手とローカルで動作します。
-
-- **ボードゲーム** — 印刷可能な34ページ版。40マス、22の物件 + 4つのルート + 2つの機関、8つのカラーシステム、7つの議会法（固定された歴史的順序）、4つのプレイヤー役割、3つの共通トラック（公共信用、公共抵抗、工業生産力）、12+12枚のイベントカード。財務省以外にも、商人道と製造業者道という2つの経済的戦略が存在します。
-- **デジタルモード** — 単一の自己完結型のHTMLファイル。完全な状態遷移システム、決定論的な mulberry32 乱数生成器、スクリプトによるAI対戦相手（財務省/金融、商人/インフラ、製造業者/産業）、ハッシュ整合性によるセーブ/ロード機能、リプレイ機能、バッチシミュレーションツール、ローカルバランステレメトリー。
-- **バランス基準** — v0.10のコアバランスがv1.1.0まで維持されています。財務省：60.0%、商人：23.5%、製造業者：16.5%（v0.18の改良版におけるCANONICAL × 400。すべてのプロファイルで目標範囲を達成）。
-- **システム障害** — 3つの段階：**信用危機**（公共信用 ≤ 4、警告）、**反乱**（公共抵抗 12、大惨事）、**デフォルト**（公共信用 0、大惨事）。v1.1.0で新しく追加されました。大惨事の状態はv0.10から変更されていません。
+**ステータス — v1.1.2 (ベータ版)。** v1.1.0は、リリースされた当日（2026年5月20日）に配布が中止されました。v1.1.1では、ゲームプレイの改善、12ラウンドのゲーム進行、勝利条件の変更が行われましたが、それでも短すぎると感じられました。v1.1.2では、ラウンド数の上限を設ける代わりに、**回路ベース**の終了条件に変更されました。ゲームは、あるプレイヤーが自勢力を共和国内を4回周回したときに終了します。平均ゲームラウンド数は約23ラウンド/67ターン（v1.1.1の約1.9倍）。ゲーム終了を引き起こしたプレイヤーが自動的に勝利するわけではなく、最終的な集計で最も高い影響力を持つプレイヤーが勝利します。印刷版のボードゲームは、v0.2で安定しています。詳細な変更点とベータ版に関する注意点は、`CHANGELOG.md`ファイルを参照してください。
 
 ---
 
-## v1.1.0の新機能
-
-### システム障害の基礎
-
-v1.1.0では、3段階のシステム障害の階層が実装されました。公共信用が0になった場合のデフォルト状態は、依然として壊滅的な金融崩壊の状態（各プレイヤーにつき50%の現金と1つのアップグレードを失う）です。公共抵抗が12になった場合の反乱状態は、依然として壊滅的な政治崩壊の状態（収入のアップグレードが破壊される）です。その間、新しい段階的なイベントである**信用危機**が発生します。公共信用が4以下になったときに初めて発生し、抵抗力を+1増加させ、システムログに「システム」の行を記録します。信用をリセットしたり、資産を破壊したり、ゲームを終了したりすることはありません。
-
-システム障害を実際にゲームプレイで可視化するために、現在、信用を低下させる4つのプレッシャーカードが追加されています。
-
-| カード | 効果 |
-|---|---|
-| 銀行の取り付け騒ぎ | 公共信用 -1、工業生産力 -1 |
-| 投機熱（信用 ≥ 7） | 公共信用 -1、抵抗力 +1、未所有の収入/国債のオークション |
-| 投機熱（信用 ≤ 6） | 公共信用 -2、抵抗力 +1、未所有の収入/国債のオークション |
-| 反連邦主義のパンフレット | 公共信用 -1、抵抗力 +1、収入システムを持つ物件ごとに30のTN（税金） |
-
-1周目の資金調達法は、依然として公共信用を+2増加させます。大惨事であるデフォルトの状態は、バランス目標ではなく、劇的な制限として維持されています。信用危機が、実際の状態を示す指標となります。
-
-CANONICAL-400の証拠（シード値：2026 – 2425）：財務省：60.0%、商人：23.5%、製造業者：16.5%。信用危機が発生する回数：2 / 400。デフォルトが発生する回数：0 / 400。反乱が発生する回数：0 / 400。抵抗力 ≥ 8 の状態が維持される回数：0 / 400。決定論：合格。詳細な証拠は、`experiments/v0.18-failure-pressure-candidate/sovereign-v0.18-evidence-sweep.html`で確認できます。
-
-### ゲーム全体の視覚的な改善
-
-プレイヤーが目にするすべての要素は、統一された連邦主義の財務省製品としてデザインされています。
-
-- 上部のバーに表示されるロゴ + モード表示 + 控えめなバージョン表示（テレメトリーダッシュボードのヘッダーではなくなりました）
-- 初回起動時に表示されるチュートリアルオーバーレイで、3つのトラックと3つのシステム障害段階を紹介
-- コーナーに装飾のあるタイル、システムカラーバンド、機関、ルート、税金、イベントスペースの区別されたデザイン
-- `CREDIT_CRISIS` / `DEFAULT` / `REBELLION`のログ行には、明瞭さを示す色、ボーダー、ラベルが適用されています（アクセシビリティに配慮）
-- トラックパネルには、信用危機の警告帯（1～4）と、デフォルトおよび反乱の状態を示すマーカーが表示されます
-- ゲーム終了レポートには、スコア列の上部に、信用状態/危機状態/反乱状態を示すチップが表示され、結果について明確に説明するナレーションが追加されます
-- バッチモードは、「バランスの証拠実行」として再構成されました
-- 768px以下の画面に対応するレスポンシブデザインと、印刷用のスタイルシート
-
-デザインシステムのリファレンスと、15フレームの視覚的な状態の監査は、`release/design-system/` ディレクトリに格納されています。これらは、v1.1.0 リリースのインターフェースがどのようなものかを示す、信頼できる記録です。
-
-### 動作の維持
-
-v0.18のプロモーション監査では、起源、実装、回帰テスト、バランス/失敗の証拠、およびドキュメントの準備に関する44件のチェック項目すべてに合格しました。100個のシードを使用した標準的なゲームの状態のハッシュ値は、v0.18のNodeシミュレーションと、改良されたHTML版（`3189375454`）で完全に一致しています。`SAVE_VERSION`は、改良プロセスにおいて動作に変更がなかったため、`'v0.18-candidate'`のままです。
+## 概要
 
-### 注意点
+Sovereignは、**アメリカの公的信用に関するハミルトン・システムを基にしたボードゲーム**であり、さらに、**完全なソロ/デジタル版**も付属しています。このデジタル版は、同じルールに基づいて、ブラウザ上で2つの決定的なAI対戦相手とローカルで動作します。
 
-v1.1.0の動作は、標準的なT/M/Mfgのトリプレット（400個のシード）と、MFG-MIRRORのバリアント（100個のシード）を使用してシミュレーションで検証されています。ただし、まだ人間によるプレイテストは実施されていません。
+- **ボードゲーム** — 印刷可能な34ページの版。40マス、22の物件、4つのルート、2つの機関、8つのカラーシステム、7つの議会法（歴史的な順序で固定）、4つのプレイヤー役割、3つの共通トラック（公的信用、公的抵抗、工業生産力）、12+12枚のイベントカード。財務省以外にも、商人道と製造業者道という2つの有効な経済的戦略が存在します。v0.2のバランス調整は完了しており、変更はありません。
+- **デジタルモード** — 単一の自己完結型のHTMLファイル。回路ベースの終了条件：あるプレイヤーが財務省のオープンを4回完了したときにゲームが終了します。平均ゲームラウンド数は約23ラウンド（67ターン）。最終的な集計で、最も高い影響力を持つプレイヤーが勝利します。*必ずしも最初にボードを一周したプレイヤーが勝利するわけではありません*。安全のために、ラウンド数の上限は30のままです（通常のプレイではこの上限に達することはありません）。決定論的な mulberry32 ランダムナンバージェネレーター、スクリプトによるAI対戦相手（財務省/金融、商人/インフラ、製造業者/産業）、ハッシュ整合性によるセーブ/ロード機能、リプレイの削除機能、開発者限定のバッチシミュレーションツール。
+- **バランスの基準** — 回路モデル（v1.1.2ベータ版）：財務省 56%、商人 19%、製造業者 25%（標準値 × 100）。3つの戦略すべてが意味のある勝利を得られるように設計されています。製造業者は、ゲームが長くなるほど、工業セットが成熟するにつれて、その割合が上昇します。ルートが優位な場合、商人の割合は低下します。v0.18の基本的なメカニズム（信用危機、キャッシュIPスコアリング、工業特許、セット完了ボーナス）は、1,000以上の決定論的なシミュレーションゲームによって駆動されたv0.3 → v0.10 → v0.18のデザインアーキテクチャから、バイト単位で完全に引き継がれています。
 
 ---
 
@@ -168,10 +127,12 @@ Sovereignは、スタンドアロンで動作するブラウザベースのボ
 
 ## 既知の制限事項
 
-- **生産能力の閾値は、通常のプレイでは稀です。** 最終的な平均生産能力は3.49。100回のゲームのうちわずか4回で6以上を達成します。ゲーム終了時の工業生産能力は、上限として存在しますが、一般的な進行ルートではありません。
-- **財務省/金融は、意図的に最も強力なプロファイルです。** これは、目標範囲内にあります。これは、歴史的な視点と一致しています。公的信用と連邦金融は、ハミルトンの主要な経済的手段でした。
-- **失敗イベントは、v0.10の検証テストで0/400回しか発生しませんでした。** デフォルト/反乱/破産のリスクは、現在では装飾的なものです。将来のバージョンで、ゲームオーバーにつながる可能性のあるリスクを再検討するかもしれません。
-- **シミュレーションでのみテスト済み。** バランスは、v0.3からv0.10までの1,000以上の決定論的なゲームで検証されています。まだ人間によるプレイテストは行われていません。戦略的な変更により、これらの数値が変動する可能性があります。
+- **v1.1.2はベータ版です。** 診断テストの結果（HTML内蔵のバッチシミュレーション、100ゲーム中100ゲームで回路終了、平均23ラウンド、勝者構成：財務省 56%、商人 19%、製造業者 25%）。まだ、熟練したプレイヤーによる完全なプレイテストは行われていません。ご自身の責任で試していただくものとして、まずは実際にプレイしていただくことをお勧めします。
+- **AIはまだ回路を巡回するための競争をしていません。** v0.18の意思決定関数を使用しており、影響力を獲得することに重点を置いています。実際のプレイヤーは、終了条件を理解すると、全く異なる行動をとる可能性があります。
+- **終了条件を満たしても、必ずしも勝利するわけではありません。** 4回目の回路を完了したプレイヤーが勝利するのは、ゲームの約3分の1の場合のみです。これは意図的なものであり、最終的な集計では、ボードを速く一周するだけでなく、経済的な深さが評価されます。エンドゲームの説明では、この点が明確に示されています。
+- **共和国内の後半部分は、新しい議会法がないため、長く感じられるかもしれません。** 議会法は、1〜7ラウンドで発動されます。平均的なプレイ時間は約23ラウンドで、共和国内の後半約16ラウンドは、新しい政治的な変化がありません。もし、この部分が単調に感じられる場合は、次回のアップデートで議会法の再配置を検討しますが、現在のシステムに戻すことはありません。
+- **財務省/金融は、意図的に最も強力なものとして設計されています。** これは、目標範囲内に収まるように調整されています。これは、歴史的な背景に基づいています。公的信用と連邦金融は、ハミルトンの経済政策における主要な要素でした。
+- **失敗イベント（デフォルト/反乱）は、主に装飾的なものです。** 信用危機は時々発生しますが、デフォルトと反乱はほとんど発生しません。エスカレーションシステムは、より多くの時間で影響を及ぼしますが、それでもほとんどの場合、上位の段階には到達しません。今後のバージョンでは、失敗状態の発生頻度について再検討する可能性があります。
 
 ---
 

package/README.md

@@ -19,58 +19,17 @@
 
 ---
 
-## What it is
-
-Sovereign is a **Hamilton-system Monopoly-grammar board game** about the founding of US public credit, plus a **complete solo / digital adaptation** that runs the same rules locally in a browser against two deterministic scripted opponents.
-
-- **Board game** — printable 34-sheet edition. 40-space board, 22 properties + 4 routes + 2 institutions, 8 color systems, 7 Acts of Congress in fixed historical order, 4 player roles, 3 shared tracks (Public Credit · Public Resistance · Industrial Capacity), 12+12 event cards. Two viable economic paths beyond Treasury: Merchant and Manufacturer.
-- **Digital mode** — single self-contained HTML file. Full state machine, deterministic mulberry32 RNG, scripted AI opponents (Treasury / Finance, Merchant / Infrastructure, Manufacturer / Industry), save / load with hash integrity, replay scrubber, batch simulation tool, local balance telemetry.
-- **Balance baseline** — v0.10 core balance preserved through v1.1.0. Treasury 60.0% · Merchant 23.5% · Manufacturer 16.5% (CANONICAL × 400 at the v0.18 polished foundation; target band met for all three profiles).
-- **Failure system** — three tiers: **Credit Crisis** (Public Credit ≤ 4, warning), **Rebellion** (Public Resistance 12, catastrophe), **Default** (Public Credit 0, catastrophe). New as of v1.1.0; the catastrophic endpoints are unchanged from v0.10.
+> **Status — v1.1.2 (beta).** v1.1.0 was withdrawn the same day it shipped (2026-05-20). v1.1.1 rebuilt the digital mode with playability fixes + 12-round pacing + a mandate victory threshold but still felt too short. v1.1.2 swaps the round-cap end condition for a **circuit-based** end condition: the game ends when one player has carried their faction around the Republic four times. Median game length ~23 rounds / 67 turns (1.9× v1.1.1). The player who triggers the end doesn't automatically win — the highest Influence at Final Accounting does. The printable board game remains stable at v0.2. See `CHANGELOG.md` for the full delta and beta caveats.
 
 ---
 
-## What's new in v1.1.0
-
-### Failure-system foundation
-
-v1.1.0 introduces a working three-tier failure hierarchy. Default at Public Credit 0 stays as the catastrophic financial collapse condition (50% cash + 1 upgrade lost per player). Rebellion at Public Resistance 12 stays as the catastrophic political collapse (Revenue upgrades destroyed). Between them, a new soft intermediate event — **Credit Crisis** — fires the first time Public Credit collapses to 4 or lower, ticks Resistance +1, and logs a System row in the ledger. It does not reset Credit, does not destroy assets, and does not end the game.
-
-To make the failure layer actually visible in play, four pressure cards now move Credit downward:
-
-| Card | Effect |
-|---|---|
-| Bank Run | Public Credit −1, Industrial Capacity −1 |
-| Speculation Fever (Credit ≥ 7) | Public Credit −1, Resistance +1, auction unowned Rev/State Debt |
-| Speculation Fever (Credit ≤ 6) | Public Credit −2, Resistance +1, auction unowned Rev/State Debt |
-| Anti-Federalist Pamphlet | Public Credit −1, Resistance +1, 30 TN per Revenue-System property |
-
-Funding Act at lap 1 still adds +2 Credit. The catastrophic Default endpoint is preserved as a dramatic limit, not a balance target — Credit Crisis carries the live signal.
-
-CANONICAL-400 evidence (seeds 2026 – 2425): Treasury 60.0% · Merchant 23.5% · Manufacturer 16.5%. Credit Crisis fires 2 / 400. Default fires 0 / 400. Rebellion fires 0 / 400. Resistance ≥ 8 holds at 0 / 400. Determinism PASS. Full evidence at `experiments/v0.18-failure-pressure-candidate/sovereign-v0.18-evidence-sweep.html`.
-
-### Whole-game visual polish
-
-Every player-facing surface art-directed as one coherent Federalist Treasury product:
-
-- Topbar wordmark + mode tag + a discreet version pill (no longer a telemetry dashboard header)
-- First-load orientation overlay introducing the three tracks and three failure tiers
-- Board tiles with corner crests, system color bands, distinct treatments for institutions, routes, taxes, and event spaces
-- Ledger rows for `CREDIT_CRISIS` / `DEFAULT` / `REBELLION` get distinct severity treatments (color + border + label — accessibility-aware)
-- Tracks panel marks the Credit Crisis warning band (1–4) and the Default and Rebellion endpoints
-- Endgame report shows posture chips (credit posture / Crisis state / Rebellion state) above the score columns, with narration that explicitly mentions Crisis / Default / Rebellion outcomes
-- Batch modal reframed as "Balance Evidence Run"
-- Responsive ≤ 768 px breakpoint and a print stylesheet
-
-The design-system reference and a fifteen-frame visual state audit live under `release/design-system/` — these are the durable record of how the v1.1.0 release surface looks.
-
-### Mechanics preserved
-
-The v0.18 promotion audit passed 44 / 44 checks across provenance, implementation, regression, balance/failure evidence, and documentation readiness. The 100-seed canonical game-state hash is byte-identical between the v0.18 Node simulation and the polished HTML (`3189375454`). `SAVE_VERSION` remains `'v0.18-candidate'` because no mechanic changed across the polish pass.
+## What it is
 
-### Caveat
+Sovereign is a **Hamilton-system Monopoly-grammar board game** about the founding of US public credit, plus a **solo / digital adaptation** that runs the same rules locally in a browser against two deterministic scripted opponents.
 
-v1.1.0 mechanics are simulation-verified across the canonical T/M/Mfg triplet (400 seeds) and the MFG-MIRROR variant (100 seeds). They are not yet human-playtested.
+- **Board game** — printable 34-sheet edition. 40-space board, 22 properties + 4 routes + 2 institutions, 8 color systems, 7 Acts of Congress in fixed historical order, 4 player roles, 3 shared tracks (Public Credit · Public Resistance · Industrial Capacity), 12+12 event cards. Two viable economic paths beyond Treasury: Merchant and Manufacturer. v0.2 balance, frozen.
+- **Digital mode** — single self-contained HTML file. Circuit-based end condition: the game ends when one player completes their fourth crossing of Treasury Opens. Median game length ~23 rounds (67 turns). At Final Accounting the highest-Influence player wins, *not necessarily the one who got around the board first*. Hard cap stays at round 30 as safety (never fires in CANONICAL × 100). Deterministic mulberry32 RNG, scripted AI opponents (Treasury / Finance, Merchant / Infrastructure, Manufacturer / Industry), save / load with hash integrity, replay scrubber, designer-gated batch simulation tool.
+- **Balance baseline** — circuit model (v1.1.2 beta): Treasury 56 % · Merchant 19 % · Manufacturer 25 % (CANONICAL × 100). All three profiles win meaningfully; Manufacturer rises in longer games as industrial sets have time to mature; Merchant share drops as routes are less dominant when cash gets spent on upgrades over a longer arc. The underlying v0.18 mechanics (Credit Crisis, cash IP scoring, Industrial Charter, set completion bonuses) are preserved byte-identical from the v0.3 → v0.10 → v0.18 design arc driven by 1,000+ deterministic simulation games.
 
 ---
 
@@ -142,7 +101,7 @@ See [`SECURITY.md`](./SECURITY.md) for vulnerability reporting and the full secu
 
 ## Features
 
-- **Solo 7-lap game** vs. two scripted opponents (Treasury / Finance and Merchant / Infrastructure by default; Manufacturer / Industry available for batch play).
+- **Solo circuit-victory game** vs. two scripted opponents (Treasury / Finance and Merchant / Infrastructure by default; Manufacturer / Industry available for batch play). The game ends when one player completes their fourth crossing of Treasury Opens. Median ~23 rounds / 67 turns. Highest Influence at Final Accounting wins.
 - **Deterministic AI** — every opponent decision is a pure function of visible state with a ledgered reason. No LLM, no opaque magic.
 - **8 game surfaces** — Board, Treasury Panel, Asset Inspector, Event Drawer, Acts of Congress, Shared Tracks, Turn Log / Ledger, Endgame Report.
 - **Auctions** — declined assets go to multi-player auction with profile-driven scripted bidding.
@@ -168,10 +127,12 @@ The fourth concept-doc profile (Opportunist / Cash) is deferred. The locked v0.1
 
 ## Known caveats
 
-- **Capacity thresholds remain rare in canonical play.** Avg final Capacity is 3.49; ≥ 6 reached in only 4 / 100 games. The endgame industrial scoring exists as a ceiling, not a regular path.
-- **Treasury / Finance remains intentionally strongest**, within the target band. This matches the historical thesis: public credit + federal finance were Hamilton's dominant economic lever.
-- **Failure events fired 0 / 400 times** in the v0.10 evidence pass. Default / Rebellion / Bankruptcy threats are currently decorative; a future version may revisit fail-state pressure.
-- **Simulation-tested only.** Balance is validated against 1,000+ deterministic games across the v0.3 → v0.10 arc. Not yet human-playtested; strategic deviation may shift these rates.
+- **v1.1.2 is a beta.** Numbers from the diagnostic held on the in-HTML batch sim (100 / 100 games trigger circuit-end, median 23 rounds, 56 / 19 / 25 winner split). It has **not** been cold-walked end-to-end by a fresh human player. Treat it as opt-in until you've walked it yourself.
+- **AI profiles do not yet race for circuits.** They run v0.18 decision functions — playing to accumulate Influence rather than racing to the fourth circuit. Real human players may behave very differently once they understand the end condition.
+- **Trigger ≠ winner.** The player who completes the fourth circuit only wins by Influence in about a third of games. This is intentional — Final Accounting rewards economic depth, not speed around the board. Endgame copy makes the distinction explicit.
+- **Late Republic stretch is long with no Acts.** Acts still fire in rounds 1-7. Median play is ~23 rounds, leaving ~16 rounds of Late Republic with no new political shocks. If this feels empty in cold play, the next fix is an Acts redistribution pass, not a return to mandate.
+- **Treasury / Finance remains intentionally strongest**, within the target band. Matches the historical thesis: public credit + federal finance were Hamilton's dominant economic lever.
+- **Failure events (Default / Rebellion) remain mostly decorative.** Credit Crisis fires occasionally; Default and Rebellion almost never. The escalation system has more time to compound but still rarely reaches the top tiers. Future versions may revisit fail-state pressure.
 
 ---
 

package/README.pt-BR.md

@@ -19,58 +19,17 @@
 
 ---
 
-## O que é
-
-Sovereign é um **jogo de tabuleiro do tipo Monopoly, inspirado no sistema Hamilton**, sobre a fundação do crédito público dos EUA, além de uma **adaptação completa para jogar sozinho / digitalmente** que executa as mesmas regras localmente em um navegador contra dois oponentes programados e determinísticos.
-
-- **Jogo de tabuleiro** — edição impressa com 34 páginas. Tabuleiro com 40 espaços, 22 propriedades + 4 rotas + 2 instituições, 8 sistemas de cores, 7 Atos do Congresso em ordem histórica fixa, 4 papéis de jogador, 3 trilhas compartilhadas (Crédito Público · Resistência Pública · Capacidade Industrial), 12+12 cartas de evento. Dois caminhos econômicos viáveis além do Tesouro: Comerciante e Fabricante.
-- **Modo digital** — um único arquivo HTML autônomo. Máquina de estados completa, gerador de números aleatórios mulberry32 determinístico, oponentes com IA programada (Tesouro / Finanças, Comerciante / Infraestrutura, Fabricante / Indústria), salvar / carregar com integridade por hash, ferramenta de análise de repetições, ferramenta de simulação em lote, telemetria de equilíbrio local.
-- **Equilíbrio base** — o equilíbrio central da versão 0.10 foi mantido até a versão 1.1.0. Tesouro: 60,0% · Comerciante: 23,5% · Fabricante: 16,5% (valor CANONICAL × 400 na base refinada da versão 0.18; a faixa alvo foi atingida para todos os três perfis).
-- **Sistema de falha** — três níveis: **Crise de Crédito** (Crédito Público ≤ 4, aviso), **Rebelião** (Resistência Pública 12, catástrofe), **Inadimplência** (Crédito Público 0, catástrofe). Novo na versão 1.1.0; os pontos finais catastróficos permanecem inalterados em relação à versão 0.10.
+**Status — v1.1.2 (beta).** A versão v1.1.0 foi descontinuada no mesmo dia do lançamento (20 de maio de 2026). A v1.1.1 reformulou o modo digital com correções de jogabilidade + ritmo de 12 rodadas + um limite de vitória por mandato, mas ainda parecia muito curto. A v1.1.2 substitui a condição de término por limite de rodadas por uma condição de término baseada em "circuitos": o jogo termina quando um jogador faz com que sua facção complete quatro voltas pela República. A duração média do jogo é de ~23 rodadas / 67 turnos (1,9 vezes a v1.1.1). O jogador que dispara o fim não vence automaticamente — vence quem tiver a maior influência no balanço final. A versão física do jogo permanece estável na versão v0.2. Consulte o arquivo `CHANGELOG.md` para obter todos os detalhes das alterações e as ressalvas da versão beta.
 
 ---
 
-## O que há de novo na versão 1.1.0
-
-### Fundação do sistema de falha
-
-A versão 1.1.0 introduz uma hierarquia de falha funcional de três níveis. A inadimplência em Crédito Público 0 permanece como a condição de colapso financeiro catastrófico (50% de dinheiro + 1 atualização perdida por jogador). A Rebelião em Resistência Pública 12 permanece como o colapso político catastrófico (atualizações de receita destruídas). Entre eles, um novo evento intermediário — **Crise de Crédito** — ocorre na primeira vez que o Crédito Público cai para 4 ou menos, aumenta a Resistência em +1 e registra uma linha no registro. Não redefine o Crédito, não destrói ativos e não termina o jogo.
-
-Para tornar a camada de falha realmente visível durante o jogo, quatro cartas de pressão agora reduzem o Crédito:
-
-| Carta | Efeito |
-|---|---|
-| Corrida aos Bancos | Crédito Público −1, Capacidade Industrial −1 |
-| Febre Especulativa (Crédito ≥ 7) | Crédito Público −1, Resistência +1, leilão de Receitas/Dívida Estadual não possuídas |
-| Febre Especulativa (Crédito ≤ 6) | Crédito Público −2, Resistência +1, leilão de Receitas/Dívida Estadual não possuídas |
-| Panfleto Antifederalista | Crédito Público −1, Resistência +1, 30 TN por propriedade do sistema de Receita |
-
-A Lei de Financiamento no turno 1 ainda adiciona +2 Crédito. O ponto final catastrófico de Inadimplência é mantido como um limite dramático, não como um alvo de equilíbrio — a Crise de Crédito transmite o sinal ativo.
-
-Evidência CANONICAL-400 (sementes 2026 – 2425): Tesouro 60,0% · Comerciante 23,5% · Fabricante 16,5%. A Crise de Crédito ocorre 2 / 400 vezes. A Inadimplência ocorre 0 / 400 vezes. A Rebelião ocorre 0 / 400 vezes. A Resistência ≥ 8 permanece em 0 / 400. Determinismo: APROVADO. Evidência completa em `experiments/v0.18-failure-pressure-candidate/sovereign-v0.18-evidence-sweep.html`.
-
-### Polimento visual do jogo completo
-
-Cada elemento visual voltado para o jogador foi projetado como um produto coerente do Tesouro Federalista:
-
-- Logotipo superior + tag de modo + uma versão discreta (não é mais um cabeçalho do painel de telemetria)
-- Sobreposição de orientação na primeira execução, apresentando as três trilhas e os três níveis de falha
-- Tiles do tabuleiro com cristas nos cantos, faixas de cores do sistema, tratamentos distintos para instituições, rotas, impostos e espaços de eventos
-- As linhas do registro para `CRISE_DE_CREDITO` / `INADIMPLÊNCIA` / `REBELIÃO` recebem tratamentos de severidade distintos (cor + borda + rótulo — com acessibilidade em mente)
-- O painel de trilhas marca a faixa de aviso da Crise de Crédito (1–4) e os pontos finais de Inadimplência e Rebelião
-- O relatório de fim de jogo mostra os chips de postura (postura do crédito / estado da Crise / estado da Rebelião) acima das colunas de pontuação, com uma narração que menciona explicitamente os resultados de Crise / Inadimplência / Rebelião
-- A janela de lote foi renomeada para "Execução de Evidência de Equilíbrio"
-- Quebra de responsividade ≤ 768 px e uma folha de estilo para impressão
-
-A referência do sistema de design e uma auditoria visual de quinze quadros estão localizadas em `release/design-system/` – estes são os registros permanentes de como a interface da versão v1.1.0 se apresenta.
-
-### Mecânicas preservadas
-
-A auditoria de promoção da versão v0.18 passou em 44 de 44 verificações, abrangendo a origem, a implementação, a regressão, as evidências de equilíbrio/falha e a prontidão da documentação. O hash do estado do jogo canônico, com 100 "sementes", é byte a byte idêntico entre a simulação Node da versão v0.18 e o HTML finalizado (`3189375454`). A variável `SAVE_VERSION` permanece como `'v0.18-candidate'` porque nenhuma mecânica foi alterada durante o processo de finalização.
+## O que é
 
-### Observação importante
+Sovereign é um **jogo de tabuleiro do tipo Monopoly, inspirado no sistema Hamilton**, sobre a fundação do crédito público dos EUA, além de uma **adaptação completa para jogar sozinho / digitalmente** que executa as mesmas regras localmente em um navegador contra dois oponentes programados e determinísticos.
 
-As mecânicas da versão v1.1.0 foram verificadas por simulação em todo o conjunto canônico de T/M/Mfg (400 sementes) e na variante MFG-MIRROR (100 sementes). Elas ainda não foram testadas por jogadores humanos.
+- **Jogo de tabuleiro** — edição impressa de 34 páginas. Tabuleiro com 40 espaços, 22 propriedades + 4 rotas + 2 instituições, 8 sistemas de cores, 7 Atos do Congresso em ordem histórica fixa, 4 papéis de jogador, 3 trilhas compartilhadas (Crédito Público · Resistência Pública · Capacidade Industrial), 12+12 cartas de evento. Dois caminhos econômicos viáveis além do Tesouro: Comerciante e Fabricante. Equilíbrio da versão v0.2, congelado.
+- **Modo digital** — um único arquivo HTML autônomo. Condição de término baseada em "circuitos": o jogo termina quando um jogador completa sua quarta passagem pelo Tesouro. A duração média do jogo é de ~23 rodadas (67 turnos). No balanço final, o jogador com a maior influência vence, *não necessariamente aquele que completou o tabuleiro primeiro*. O limite máximo permanece em 30 rodadas como medida de segurança (nunca é atingido no modo CANONICAL × 100). Gerador de números aleatórios mulberry32 determinístico, oponentes de IA programados (Tesouro / Finanças, Comerciante / Infraestrutura, Fabricante / Indústria), salvar / carregar com integridade de hash, ferramenta de simulação em lote para o designer.
+- **Ponto de referência de equilíbrio** — modelo de "circuitos" (versão beta v1.1.2): Tesouro 56% · Comerciante 19% · Fabricante 25% (CANONICAL × 100). Todos os três perfis têm chances significativas de vitória; o Fabricante se torna mais relevante em jogos mais longos, pois os conjuntos industriais têm tempo para se desenvolver; a participação do Comerciante diminui à medida que as rotas se tornam menos dominantes quando o dinheiro é gasto em melhorias ao longo de um período mais longo. As mecânicas subjacentes da versão v0.18 (Crise de Crédito, pontuação de propriedade de dinheiro, Carta Industrial, bônus de conclusão de conjuntos) são preservadas de forma idêntica em bytes, desde o ciclo de design v0.3 → v0.10 → v0.18, baseado em mais de 1.000 simulações determinísticas.
 
 ---
 
@@ -168,10 +127,12 @@ O quarto perfil conceitual (Oportunista / Caixa) está adiado. O conjunto compet
 
 ## Observações
 
-- **Os limites de capacidade permanecem raros no jogo padrão.** A capacidade final média é de 3,49; ≥ 6 é alcançada em apenas 4/100 jogos. A pontuação industrial no final do jogo existe como um limite máximo, não como um caminho comum.
-- **O perfil Treasury / Finance permanece intencionalmente o mais forte**, dentro da faixa alvo. Isso corresponde à tese histórica: o crédito público + as finanças federais eram o principal instrumento econômico de Hamilton.
-- **Eventos de falha ocorreram 0 / 400 vezes** na versão de teste da v0.10. As ameaças de falha (default, rebelião, falência) são atualmente decorativas; uma versão futura pode reavaliar a pressão para atingir um estado de falha.
-- **Testado apenas em simulação.** O equilíbrio é validado contra mais de 1.000 jogos determinísticos no intervalo v0.3 → v0.10. Ainda não foi testado com jogadores humanos; desvios estratégicos podem alterar essas taxas.
+- **A v1.1.2 é uma versão beta.** Os números vêm da simulação em lote dentro do HTML (100 / 100 jogos terminam pelo "circuito", duração média de 23 rodadas, divisão de vencedores 56 / 19 / 25). Ela **não** foi testada completamente por um jogador humano. Considere-a como uma versão opcional até que você a experimente você mesmo.
+- **Os perfis de IA ainda não competem pelos "circuitos".** Eles usam as funções de decisão da versão v0.18 — jogando para acumular influência, em vez de competir para completar o quarto "circuito". Jogadores humanos reais podem se comportar de maneira muito diferente assim que entenderem a condição de término.
+- **Acionar ≠ vencer.** O jogador que completa o quarto "circuito" só vence por influência em cerca de um terço dos jogos. Isso é intencional — o balanço final recompensa a profundidade econômica, não a velocidade ao redor do tabuleiro. A cópia do final do jogo torna essa distinção explícita.
+- **A parte final da República é longa e não tem Atos.** Os Atos ainda são ativados nas rodadas de 1 a 7. A duração média do jogo é de ~23 rodadas, deixando ~16 rodadas da parte final da República sem novos eventos políticos. Se isso parecer vazio em um jogo "frio", a próxima correção será uma redistribuição dos Atos, e não um retorno ao sistema de mandatos.
+- **O Tesouro / Finanças permanece intencionalmente o mais forte**, dentro da faixa alvo. Isso corresponde à tese histórica: o crédito público + as finanças federais eram o principal instrumento econômico de Hamilton.
+- **Os eventos de falha (Inadimplência / Rebelião) permanecem principalmente decorativos.** A Crise de Crédito ocorre ocasionalmente; a Inadimplência e a Rebelião quase nunca ocorrem. O sistema de escalada tem mais tempo para se desenvolver, mas ainda raramente atinge os níveis mais altos. Versões futuras podem reavaliar a pressão dos estados de falha.
 
 ---
 

package/README.zh.md

@@ -19,58 +19,17 @@
 
 ---
 
-## 简介
-
-Sovereign是一款**基于“汉密尔顿系统”的“大富翁”风格棋盘游戏**，讲述的是美国公共信贷的建立，同时还提供一个**完整的单人/数字版本**，它可以在浏览器中本地运行，与两个具有固定行为的AI对手对战。
-
-- **棋盘游戏** — 可打印的34页版本。 棋盘包含40个格子，22个地产 + 4条路线 + 2个机构，8种颜色系统，7部国会法案按照固定的历史顺序排列，4种玩家角色，3条共享进度条（公共信用、公共抵抗、工业能力），12+12张事件卡。 除了财政部之外，还有两种可行的经济发展路径：商人路径和制造商路径。
-- **数字模式** — 单个独立的HTML文件。 完整的状态机，使用deterministic mulberry32随机数生成器，带有脚本化的AI对手（财政部/金融、商人/基础设施、制造商/工业），支持保存/加载并具有哈希完整性，支持重播，包含批量模拟工具，以及本地平衡数据统计。
-- **平衡基准** — v0.10的核心平衡在v1.1.0版本中得到保留。 财政部：60.0%；商人：23.5%；制造商：16.5%（在v0.18版本中经过完善的基础版本，所有三种模式都达到了目标范围）。
-- **失败系统** — 分为三个等级：**信用危机**（公共信用≤4，警告），**叛乱**（公共抵抗12，灾难），**违约**（公共信用0，灾难）。 新增于v1.1.0版本；灾难性的结束条件与v0.10版本相同。
+**状态 — v1.1.2 (beta)。** v1.1.0 版本在发布当天就被撤回（2026年5月20日）。v1.1.1 版本对数字模式进行了改进，修复了游戏体验问题，增加了12轮的游戏流程，并设置了强制胜利的条件，但仍然感觉太短。v1.1.2 版本将回合限制的结束条件替换为基于“电路”的结束条件：当一名玩家使自己的派系在共和国境内完成了四次循环时，游戏结束。平均游戏时长约为23轮/67回合（比 v1.1.1 版本长 1.9 倍）。触发游戏结束的玩家不一定获胜，最终根据“最终结算”时的影响力值来决定胜负。可打印的桌面游戏版本仍然稳定在 v0.2 版本。请参阅 `CHANGELOG.md` 文件以获取完整的变更记录和 beta 版本的注意事项。
 
 ---
 
-## v1.1.0版本的新功能
-
-### 失败系统基础
-
-v1.1.0版本引入了一个可用的三级失败等级系统。 违约，即公共信用为0，仍然是导致金融崩溃的灾难性条件（每个玩家损失50%的现金 + 1个升级）。 叛乱，即公共抵抗为12，仍然是导致政治崩溃的灾难性条件（收入升级被摧毁）。 在这两者之间，新增了一个轻微的中间事件——**信用危机**。 当公共信用降至4或更低时，会触发该事件，增加公共抵抗1点，并在账本中记录一条“系统”行。 它不会重置信用，不会摧毁资产，也不会结束游戏。
-
-为了使失败层级在游戏中真正可见，现在有四张压力卡会降低信用值：
-
-| 卡牌 | 效果 |
-|---|---|
-| 银行挤兑 | 公共信用 -1，工业能力 -1 |
-| 投机狂热（信用≥7） | 公共信用 -1，抵抗 +1，拍卖未拥有收入/国家债务 |
-| 投机狂热（信用≤6） | 公共信用 -2，抵抗 +1，拍卖未拥有收入/国家债务 |
-| 反联邦党宣言 | 公共信用 -1，抵抗 +1，每个收入系统地产损失30个TN |
-
-在第一回合，资助法案仍然会增加+2个信用值。 违约，即公共信用为0，仍然被保留为一种戏剧性的限制，而不是平衡目标——信用危机会发出警报。
-
-CANONICAL-400实验数据（种子：2026 – 2425）：财政部：60.0%；商人：23.5%；制造商：16.5%。 信用危机触发2次/400次。 违约触发0次/400次。 叛乱触发0次/400次。 公共抵抗≥8的概率为0次/400次。 确定性测试通过。 完整的实验数据请参考：`experiments/v0.18-failure-pressure-candidate/sovereign-v0.18-evidence-sweep.html`。
-
-### 整体视觉美化
-
-所有面向玩家的界面都经过艺术设计，使其成为一个统一的联邦财政产品：
-
-- 顶部栏：包含游戏名称、模式标签以及一个不显眼的版本标识（不再是数据统计仪表盘的标题）
-- 首次加载时，会显示一个引导界面，介绍三个进度条和三个失败等级
-- 棋盘：包含角落的徽章，颜色分区分块，以及对机构、路线、税收和事件区域的独特处理
-- 账本行：`信用危机` / `违约` / `叛乱` 会有不同的严重程度标识（颜色 + 边框 + 标签，考虑可访问性）
-- 进度条面板：标记信用危机的警告区域（1–4）以及违约和叛乱的结束点
-- 游戏结束报告：显示姿态芯片（信用姿态/危机状态/叛乱状态），位于分数列上方，并带有明确提及危机/违约/叛乱结果的说明
-- 批量模式：重命名为“平衡数据运行”
-- 响应式设计，支持≤768像素的屏幕，并包含打印样式表。
-
-设计系统参考文档以及包含十五个画面的视觉状态审核文件位于 `release/design-system/` 目录下，这些文件记录了 v1.1.0 版本的界面外观。
-
-### 机制保持不变
-
-v0.18 版本的推广审核通过了所有 44 项检查，涵盖了来源、实现、回归测试、平衡/失败证据以及文档准备情况。 100 个种子值的标准游戏状态哈希值在 v0.18 版本的 Node 模拟和优化后的 HTML 版本之间完全一致（值为 `3189375454`）。 `SAVE_VERSION` 仍然是 `'v0.18-candidate'`，因为在优化过程中，没有任何机制发生改变。
+## 简介
 
-### 注意事项
+Sovereign是一款**基于“汉密尔顿系统”的“大富翁”风格棋盘游戏**，讲述的是美国公共信贷的建立，同时还提供一个**完整的单人/数字版本**，它可以在浏览器中本地运行，与两个具有固定行为的AI对手对战。
 
-v1.1.0 版本的机制已通过模拟验证，验证范围包括标准的 T/M/Mfg 三元组（400 个种子值）以及 MFG-MIRROR 变体（100 个种子值）。 但尚未经过人工游戏测试。
+- **桌面游戏** — 可打印的 34 页版本。游戏棋盘包含 40 个格子，22 处地产 + 4 条路线 + 2 个机构，8 个颜色系统，7 项国会法案按照固定的历史顺序排列，4 个玩家角色，3 条共享进度条（公共信用、公共抵抗、工业能力），12+12 张事件卡。除了财政部之外，还有两种可行的经济发展路径：商人路径和制造商路径。v0.2 版本的平衡性已确定，不再进行调整。
+- **数字模式** — 单个独立的 HTML 文件。基于“电路”的结束条件：当一名玩家完成第四次“财政部开放”时，游戏结束。平均游戏时长约为 23 轮（67 回合）。在“最终结算”时，拥有最高影响力的玩家获胜，*不一定是最先绕完棋盘的玩家*。为了安全起见，回合上限仍然设置为 30 回合（在标准模式下几乎不会达到）。使用确定性的 mulberry32 随机数生成器，包含脚本化的 AI 对手（财政部/金融、商人/基础设施、制造商/工业），支持保存/加载功能，并具有哈希完整性校验、回放清除功能以及设计师提供的批量模拟工具。
+- **平衡性基准** — 基于“电路”模型的 v1.1.2 beta 版本：财政部 56%，商人 19%，制造商 25%（标准模式下）。这三种发展模式都有可能获胜；在较长的游戏中，制造商的发展会更加突出，因为工业组合有时间成熟；随着路线的重要性下降，以及玩家将资金用于升级，商人的份额可能会下降。底层 v0.18 的机制（信用危机、现金 IP 得分、工业特许、完成组合的奖励）与 v0.3 → v0.10 → v0.18 版本设计保持完全一致，并通过 1000 多场确定性模拟游戏进行验证。
 
 ---
 
@@ -168,10 +127,12 @@ Sovereign 是一款独立的、基于浏览器的棋盘游戏。命令行界面
 
 ## 已知问题
 
-- **产能阈值在标准游戏中仍然很少出现。** 最终平均产能为 3.49；只有在 4/100 局游戏中达到 ≥ 6。游戏结束时的工业评分仅作为上限，而不是常规的路径。
-- **财政/金融配置仍然有意地是最强的，** 处于目标范围内。这符合历史论证：公共信用和联邦财政是汉密尔顿的主要经济杠杆。
-- **在 v0.10 版本的测试中，失败事件触发了 0/400 次。** 默认/叛乱/破产威胁目前仅为装饰；未来的版本可能会重新考虑失败状态的压力。
-- **仅进行模拟测试。** 平衡是通过对 v0.3 → v0.10 版本中的 1000+ 局确定性游戏进行的验证。尚未进行人类游戏测试；战略偏差可能会改变这些数值。
+- **v1.1.2 是一个 beta 版本。** 诊断数据来自 HTML 批量模拟（100/100 场游戏触发“电路”结束，平均 23 轮，获胜者比例为 56%/19%/25%）。它**尚未**由全新的人类玩家进行完整的体验测试。请将其视为可选版本，直到您亲自体验过它为止。
+- **AI 角色目前尚未争夺“电路”。** 它们运行的是 v0.18 的决策函数，目标是积累影响力，而不是争夺第四个“电路”。真正的玩家在理解结束条件后，可能会表现出不同的行为。
+- **触发 ≠ 获胜。** 只有大约三分之一的游戏中，完成第四个“电路”的玩家会获胜。这是有意为之——“最终结算”奖励的是经济深度，而不是在棋盘上的速度。游戏结束时的提示会明确这一点。
+- **共和国晚期阶段的棋盘很长，没有新的法案。** 法案仍然在 1-7 回合生效。平均游戏时长约为 23 轮，这意味着在共和国晚期阶段，有大约 16 轮没有新的政治事件发生。如果这部分内容在体验中感觉空洞，下一个改进方向是重新分配法案，而不是恢复强制胜利的条件。
+- **财政部/金融仍然是最强大的，** 并且在目标范围内。这符合历史背景：公共信用和联邦金融是汉密尔顿的主要经济杠杆。
+- **失败事件（违约/叛乱）仍然主要起到装饰作用。** 信用危机偶尔会触发；违约和叛乱几乎从不发生。升级系统有更多时间来累积，但仍然很少达到最高等级。未来的版本可能会重新考虑失败状态的压力。
 
 ---
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@mcptoolshop/sovereign",
-  "version": "1.1.0",
+  "version": "1.1.2",
   "description": "Sovereign — a Hamilton-system economic board game and its solo / digital adaptation. Browser-based, deterministic, self-contained.",
   "type": "module",
   "main": "bin/sovereign.js",

package/release/CHANGELOG.md

@@ -5,37 +5,106 @@
 
 ---
 
-## v1.1.0 — Failure-system foundation + whole-game visual polish — 2026-05-20
+## v1.1.2 — Circuit victory + human voice — 2026-05-20
 
-Promotes the v0.18 failure-pressure foundation and the whole-game visual polish pass to the canonical Sovereign Solo / Digital Mode release surface. The digital artifact ships at `release/digital-mode/sovereign-solo.html`; the prior v0.10 canonical is archived alongside it as `sovereign-solo-v0.10-baseline.html`. The in-HTML `SAVE_VERSION` remains `'v0.18-candidate'` — no mechanic changed across the polish pass.
+**Beta.** Same day as v1.1.1, but a meaningfully different shape. After the v1.1.1 cold playthrough the game still felt too short and the ending too arbitrary — it ended because the clock ran out, not because someone won the Republic. v1.1.2 replaces the round-cap-with-mandate end condition with a **circuit-based** end condition tied to physical movement around the board.
 
-**Three-tier failure system (new in v1.1.0):**
+**The new end condition:**
 
-| Tier | Trigger | Severity | Effect | Reset |
-|---|---|---|---|---|
-| Credit Crisis | Public Credit ≤ 4 (first time per game) | warning | Resistance +1, System log | none |
-| Rebellion | Public Resistance 12 | catastrophe | Revenue upgrades destroyed; Whiskey owner → Crisis | resistance → 6 |
-| Default | Public Credit 0 | catastrophe | 50% cash + 1 upgrade per player | credit → 3 |
+The game ends when one player has carried their faction around the Republic four times — i.e., completed their fourth crossing of Treasury Opens. At that moment, Final Accounting fires and the highest-Influence player wins, *not necessarily the player who got around first*. Hard cap stays at round 30 (a safety; essentially never fires — diagnostic showed 100 / 100 CANONICAL × 100 games complete by round 28).
 
-**Failure-pressure arc (v0.11 → v0.18) — what shipped:**
+**Why this is structurally different from v1.1.1:**
 
-- Bank Run drops Public Credit -1 + Industrial Capacity -1 (v0.11).
-- Speculation Fever drops Public Credit -1 + Resistance +1 + auctions an unowned Rev/State Debt property (v0.13).
-- Anti-Federalist Pamphlet adds Public Credit -1 to its existing Resistance +1 + 30 TN per Revenue-System property hit (v0.16).
-- Speculation Fever fragile-credit escalation: Credit -1 at Credit ≥ 7, Credit -2 at Credit ≤ 6 (v0.17).
-- Credit Crisis intermediate failure event at Credit ≤ 4 (v0.18).
+- v1.1.1: 12-round cap with mandate (15 IP + 5 lead) as early-trigger. Median 9 rounds, 36 turns. Ended by clock.
+- v1.1.2: no round cap that anyone hits in practice. Median **23 rounds / 67 turns** — about 1.9× the previous game length. Ends when somebody physically rounds the board for the fourth time.
 
-**Why v0.14 / v0.15 recovery gates were rejected:** mechanically correct but aggregate-inert. The v0.16 → v0.18 chain pivoted to pressure-side changes rather than recovery-side gates.
+**CANONICAL × 100 (in-HTML batch sim):**
 
-**Whole-game visual polish (presentation only, mechanics byte-identical):**
+| Metric | v1.1.1 (mandate) | v1.1.2 (circuit) |
+|---|---:|---:|
+| Triggered | 62 / 100 | 100 / 100 |
+| Median rounds | 9 | 23 |
+| Median turns | ~27 | 67 |
+| Treasury wins | 52 % | 56 % |
+| Merchant wins | 32 % | 19 % |
+| Manufacturer wins | 16 % | 25 % |
+| Avg winning IP | ~16 | 28.7 |
+| Avg lead margin | 5 | 7.9 |
+| Bankruptcy events | 7 | observed |
+| Safety-cap fires | n/a | 0 / 100 |
 
-All 16 player-facing surfaces art-directed as one Federalist Treasury system: topbar wordmark; first-load orientation overlay; board tile crests with corner glyphs and system color bands; ledger row severity treatments for `CREDIT_CRISIS` / `DEFAULT` / `REBELLION`; tracks panel warning bands; distinct Market Shock vs Republic Debate card chrome; endgame posture chips + drop-cap narration; "Balance Evidence Run" batch modal; responsive ≤ 768 px breakpoint; print stylesheet. Design-system reference under `release/design-system/`.
+Trade-off worth naming: Merchant share drops (32 → 19) while Manufacturer rises (16 → 25). Longer games give Treasury and Manufacturer time to compound industrial / financial scoring, where Merchant's route-rent advantage matures faster in shorter games. Treasury still doesn't run away — share stays in the band.
 
-**Mechanics-preservation evidence:** v0.18 promotion audit 44 / 44 PASS. CANONICAL-400 balance bands held: Treasury 60.0 % / Merchant 23.5 % / Manufacturer 16.5 %. 100-seed canonical state hash 3189375454 byte-identical between Node sim and polished HTML.
+**Trigger ≠ winner.** The player who completes the fourth circuit wins by Influence only 33 / 100 games. The other 67, somebody else holds the heavier ledger at termination. This is intentional — it preserves the "carry it across the line" feel while still rewarding economic depth. Endgame copy makes this explicit in both branches.
 
-**Caveat:** v0.18 mechanics are simulation-verified, not yet human-playtested.
+**Human voice copy pass.** All circuit / endgame narration rewritten away from gameplay-jargon ("Circuit Victory", "trigger player", "claimed the Republic") toward narrative-voice ("Hamilton has been around four times. The books close.", "Hamilton rounded the Republic four times — but the heaviest ledger belonged to Morris."). Per-circuit ledger entries vary by circuit number — first lap, two circuits in, three deep, books close.
 
-Translation re-run required (README.md changed). Translations run locally via TranslateGemma before tag push.
+**What's preserved from v1.1.1:**
+
+- All Pass 1 / Pass 2 chrome (action rail above board, positions strip, portfolio strip, collapsed sidebar, 26 px tokens, designer gate, dice overlay)
+- Pass 3 rent surfacing band
+- All v0.18 mechanics: card effects, Acts (still firing rounds 1-7), scoring math, rent math, profile decision functions, RNG, deck order, Credit Crisis / Default / Rebellion triggers
+- Designer URL flag: `?designer=1&circuit_target=N` for tuning the circuit threshold
+
+**What changed:**
+
+- `SAVE_VERSION = 'v0.21-circuit-candidate'`. v0.20 (v1.1.1) saves refuse to load with a friendly non-crash message.
+- Mandate threshold removed. No more 15-IP + 5-lead trigger; the game ends on the fourth circuit only.
+- Round counter on the chrome no longer reads "Round N of 12" — just "Round N" (and "· Late Republic" after round 7).
+- Acts still fire in rounds 1-7; rounds 8-onward are Late Republic with no new political shocks. **Open risk**: median play has ~16 rounds of Late Republic with no Acts firing. If cold play finds this stretch feels empty, the v1.1.3 fix is an Acts redistribution pass, not a return to mandate.
+
+**Beta caveats:**
+
+- Not yet cold-walked end-to-end by a fresh human player. Beta tag is the safety net.
+- Behavioral adaptation unmeasured: AI profiles still play current decision functions, not circuit-aware ones (e.g., Treasury doesn't yet race to hit the fourth circuit).
+- The Late Republic stretch (~16 rounds in median play with no Acts firing) is a known open risk. Worth flagging in any playtest notes.
+
+**npm history note:**
+
+v1.1.0 remains on npm as the historical withdrawn version. v1.1.1 (mandate beta) and v1.1.2 (circuit beta) are both published. `npm install @mcptoolshop/sovereign` resolves to v1.1.2 (`latest`).
+
+---
+
+## v1.1.1 — Digital mode beta: human playability rebuild + 12-round pacing + mandate victory — 2026-05-20
+
+**Beta release.** Second attempt at the digital mode after v1.1.0 was withdrawn the same day it shipped (2026-05-20). v1.1.0 simulation-verified but failed first cold human play; v1.1.1 fixes the structural human-playability gaps that audit could not catch. Not fully cold-validated end-to-end; shipping as beta so anyone who wants the digital mode can opt in while the player experience continues to settle.
+
+**What's new vs v1.0.2:**
+
+- **Human playability rebuild (Pass 1 + Pass 2).** Action rail moved above the board (visible without scrolling). Positions strip shows each player's location, cash, and Influence with active-player highlight. Portfolio strip near the action rail surfaces holdings without requiring the sidebar. Right sidebar collapsed by default with `▶ Panels` toggle so the board occupies more real estate. 26 px tokens with active-player halo and `YOU` pointer. Designer-only chrome (Balance Sweep, seed pill, version pill, phase pill, canonical/telemetry labels) hidden by default; visible only with `?designer=1`. Real dice-roll overlay (~3.3 s, deterministic outcome). Acts explain their effect before the YES / NO vote. First-round helping-hand footer fades after lap 1.
+- **Rent surfacing (Pass 3).** New landing-outcome band synthesized from existing ledger rows. Opponent property: "You landed on Hamilton's Customs House. Paid 40 TN rent." Opponent route: "Paid 50 TN route toll." Opponent institution: "Paid 70 TN institution payment." Own property: "No rent due" (plus "Upgrade available" when applicable). Tax space: "Paid 200 TN in tax." Sent-to-Crisis surfaces too. No reducer change; the band reads rows the ledger already wrote.
+- **12-round pacing.** Game length extended from 7 rounds (21 turns total) to 12 rounds (36 turns total). Acts still fire in rounds 1-7 as before; rounds 8-12 are Late Republic rounds with no new Acts. Diagnostic across CANONICAL × 100: average board circuits per player 1.24 → 2.12, average upgrades per game 3.3 → 11.0, average rent events 2.2 → 5.8. Treasury share drops 59 % → 52 % because cash gets spent on upgrades instead of hoarded for scoring. Merchant 25 % → 32 %. Manufacturer holds at 16 %. Player-facing language: "Lap" → "Round" throughout chrome.
+- **Mandate victory model.** From end of round 8, a player with at least 15 Influence and a 5-point lead over second place triggers Final Accounting and ends the game early. If no mandate triggers by end of round 12, the existing hard-cap scoring decides. Diagnostic across 100 games: 62 / 100 end by mandate (median round 9), 38 / 100 by hard cap. Treasury mandate share 53 % — no runaway. All three profiles win mandates. Designer-overridable via `?designer=1&mandate_threshold=N&mandate_lead=N&mandate_min_round=N&mandate_hard_cap=N`.
+- **Influence visibility.** Each player's current Influence shows in the positions strip from round 1. Starting round 8, a compact mandate-distance badge appears on the leader: "Hamilton · 14 IP · 1 from threshold, +2 lead needed."
+
+**Mechanics preservation:**
+
+| Item | Status |
+|---|---|
+| Card effects (Market Shock, Republic Debate) | Untouched from v0.18 |
+| Act effects | Untouched |
+| Reducer logic apart from round cap + mandate check | Untouched |
+| Scoring math (cashIP 1 per 400 TN, sets, upgrades, institutions, NF Credit endgame, Capacity bonuses) | Untouched |
+| Profile decision functions (Treasury, Merchant, Manufacturer) | Untouched |
+| Track thresholds (Credit Crisis ≤ 4, Default = 0, Rebellion = 12) | Untouched |
+| RNG (`mulberry32(state.rngSeed)`) and deck order | Untouched |
+| v0.18 Credit Crisis intermediate failure event | Preserved (now fires ~2 / 100 instead of 0 / 100 because pressure has more time to compound) |
+| Rent math | Untouched (the band surfaces existing reducer rent; doesn't change the math) |
+| Designer gate (`?designer=1`) | Preserved + extended for mandate URL flags |
+
+**Save format:**
+
+`SAVE_VERSION = 'v0.20-mandate-candidate'`. Saves from v0.18 (v1.1.0) and v0.19 (interim pacing candidate) refuse to load with a clear non-crashing message. v1.0.0 / v1.0.1 / v1.0.2 do not run the digital mode, so no save compatibility concern there.
+
+**Beta caveats:**
+
+- Numbers held to the diagnostic on the in-HTML batch simulation (62 / 100 mandate vs predicted 67 / 100; 51 / 33 / 16 winner split exactly matching predicted), but the candidate has not been cold-walked end-to-end by a fresh human player. Beta tag is the safety net.
+- Behavioral adaptation is unmeasured: the AI profiles do not yet "race for the mandate" — they play the same decision functions as v0.18, just with more rounds available. Real human players may behave differently; trigger rate may shift.
+- Bankruptcy emerges as a new soft-pressure dynamic in longer games (7 / 100 events at 12 rounds with mandate, vs 18 / 100 without mandate, vs 0 / 100 at 7 rounds). The mandate model partially defuses it because games end before late-game cash stress accumulates. Worth observing.
+
+**npm history note:**
+
+v1.1.0 remains on npm as an unrecommended historical version. `npm install @mcptoolshop/sovereign` resolves to v1.1.1 (`latest`).
 
 ---