Skip to content

A Civilization-style strategy game about governing Israel from 15 May 1948. Real ministers, real map, every fact sourced. Unity + deterministic C# sim.

40.8 KB
Raw Blame
# Game rules. Values are game balance, except where a comment cites a source.
start: 1948-05-15                 # first day covered by the CBS 1948 immigration series (aliyah.yaml)
start_rung: 5                     # the war is under way when the game opens
attention_per_turn: 4                    # kept at 4 for monthly turns (owner, 15 October 2026, Q-CUT-5)
# Attention each order costs. The GDD orders' costs are design values (the owner's balance, not history), docs/rules/mvp.md
# R-LOOP-005; a unit order will be free (D-A6).
attention_cost:
  set_policy: 1
  allocate_budget: 2
  choose_event_option: 1
  set_mobilization: 1
  security_order: 1
  appoint: 2
  map_action: 1
  assign_workers: 0.5                    # DESIGN VALUE: owner-set balance, not a historical figure (D-A6)
  train: 1                               # DESIGN VALUE: owner-set balance, not a historical figure (D-A6)
  cancel_training: 0.5                   # DESIGN VALUE: owner-set balance, not a historical figure (D-A6)
  return_craftsmen: 0.5                  # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: as assign_workers, the normal Attention of an order, #1511 owner ruling)
  reassign_craftsmen: 0.5                # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: as assign_workers, the normal Attention of an order, #1511 owner ruling)
  reassign_pioneers: 0                   # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: unit orders are free, the normal Attention of an order, #1511 owner ruling)
  reinforce: 1                           # DESIGN VALUE: owner-set balance, not a historical figure (D-A6; owner, 24 October 2026, #1510: "1, like Train")
  pioneer_work: 0                        # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: unit orders are free)
  stop_work: 0                           # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: unit orders are free)
  unit_order: 0                          # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: unit orders are free; R-ORDER-001, #746)
  start_project: 1                       # DESIGN VALUE: owner-set balance, not a historical figure (D-A6, D-C5, Q-PROJ-5: the map action's cost)
  cancel_project: 0.5                    # DESIGN VALUE: owner-set balance, not a historical figure (D-A6, D-C5, Q-PROJ-5: cancel_training's cost)
  move: 0                                # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: unit orders are free)
  buy_arms: 1                            # DESIGN VALUE: owner-set balance, not a historical figure (D-A6, D-G6, #781: one order, whatever the quantity)
  convert_dollars: 0                     # DESIGN VALUE: owner-set balance, not a historical figure (D-A6, D-C3, #813: the war months' cards already take all of Attention in 5 of 11 turns, so the lever is free; owner 2 November 2026 on #2204: "Keep 0", the recommended option)
  resolve_capture: 0                     # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: unit orders are free; Q-TECH-3: one key, so Raid, Occupy and Deport cost the same; proposed on #797)
  end_occupation: 0                      # DESIGN VALUE: owner-set balance, not a historical figure (D-A6: unit orders are free; R-CTRL-014: one key for Annex and Withdraw, as resolve_capture; proposed on #799)
# Turn length on each security rung, rungs 1 to 5 (R-TURN-001). The GDD MVP plays monthly turns on every rung (D-B1), so
# turns from 15 May 1948 end on the 15th of each month.
turn_lengths: [month, month, month, month, month]
# The GDD v0.1 rules (docs/rules/mvp.md). Every number here is a design value: the owner's balance, set or accepted in
# review (Q-CUT-1, owner, 15 October 2026: "Labelled design values"), never a historical claim. DesignValuesTests lists
# them all (R-LOOP-005); changing one is a balance edit that re-pins the goldens.
mvp:
  # The game ends on 15 January 1952, after 44 monthly turns (R-LOOP-004; owner, 15 October 2026, D-E5: "End on 15 Jan
  # 1952"). The founding era still runs to 1957 (G2), so its later content stays and loads, but play stops here.
  end: 1952-01-15
  people_per_unit: 1000                  # DESIGN VALUE: owner-set balance, not a historical figure (D-A1: people in one population unit)
  income_per_job_per_year: 0.06          # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: IL millions per employed unit a year)
  craftsman_income_multiple: 2           # DESIGN VALUE: owner-set balance, not a historical figure (D-D4, Q-CRAFT-2: an employed craftsman earns this times income_per_job_per_year; R-CRAFT-002)
  # R-MONEY-002, option 3 (owner, 18 October 2026, #884): ordinary spending already represents aggregate public expense,
  # so no civilian place pays upkeep in the MVP. These are design values, not claims that historical upkeep was zero.
  city_upkeep_per_year: 0.0              # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: IL millions per city a year)
  housing_upkeep_per_year: 0.0           # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: IL millions per permanent-housing project a year)
  industry_upkeep_per_year: 0.0          # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: IL millions per industry project a year)
  transit_camp_upkeep_per_year: 0.0      # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: IL millions per transit camp a year)
  military_camp_upkeep_per_year: 0.0     # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: IL millions per military camp a year)
  hakhshara_upkeep_per_year: 0.0         # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: IL millions per Hakhshara a year)

  # R-MIN-001 (D-H3; Q-MIN-1, owner, 17 October 2026, #784): the office whose multiplier (competence times budget
  # factor, EffectsSystem.Multiplier) scales each GDD system. Q-MIN-1's mapping, "each with its office's sourced remit
  # in a comment": the Ministry of Labour ran housing and public construction; the Ministry of Agriculture, the fields
  # and the new settlements; the Ministry of Immigration, the immigrants' housing (the offices and their terms are the
  # sourced rows of offices.yaml; the remits are as the decision names them). Transportation scales corridor work
  # (Q-MIN-1); military training (Defense) joins the map with #742.
  minister_systems:
    housing_construction: labour_minister
    industry_construction: labour_minister
    field_work: agriculture_minister
    yishuv_work: agriculture_minister
    absorption: immigration_minister
    corridors: transportation_minister
  # The day the cities start taking arrivals (R-CITY-003; Q-ARRV-3, owner, 15 October 2026, #375): the census day of their
  # start values, 8 November 1948 (CBS Statistical Abstract of Israel No. 14, Table 11, pp. 30-31). The census already
  # counts the people who arrived from 15 May, so the cities take none before it; the national head count takes them all.
  city_arrivals_from: 1948-11-08         # https://www.cbs.gov.il/he/publications/DocLib/shnaton_saruk/shnaton1963_num14.pdf, checked 2026-10-15
  build_yishuv_turns: 6                  # DESIGN VALUE: owner-set balance, not a historical figure (D-D6: turns a pioneer build takes)
  build_yishuv_cost: 2                   # DESIGN VALUE: owner-set balance, not a historical figure (D-D6: IL millions, paid at the start)
  # Build a training camp (R-CAMP-001; D-D5, Q-CAMP-3). Proposed by an agent under "Agents propose design values" (owner,
  # 15 October 2026, #375): the yishuv build's pair, since no source sets a camp's cost in game units. The owner accepted
  # them ("Accept 6 turns, IL 2m", 17 October 2026, #927); the Phase 7 balance pass (#813) may tune them.
  build_camp_turns: 6                    # DESIGN VALUE: owner-set balance, not a historical figure (D-D5, Q-CAMP-3: turns a camp build takes)
  build_camp_cost: 2                     # DESIGN VALUE: owner-set balance, not a historical figure (D-D5, Q-CAMP-3: IL millions, paid at the start)
  # Plow fields (R-FIELD-001, R-FIELD-002; D-D10). Proposed by an agent and taken by the owner with the plan's other
  # recommendations (Q-FIELD-3, owner, 17 October 2026, #782: "1 Food a turn"): the smallest whole yield, about 0.1% of a
  # month's national need; plowing takes the recruit-training pair, the smallest project-sized values in the code.
  plow_fields_turns: 3                   # DESIGN VALUE: owner-set balance, not a historical figure (D-D10, Q-FIELD-3: turns plowing takes)
  plow_fields_cost: 0.5                  # DESIGN VALUE: owner-set balance, not a historical figure (D-D10, Q-FIELD-3: IL millions, paid at the start)
  # Dry swamp (R-SWAMP-001; D-D6, Q-SWAMP-1). Proposed by an agent under "Agents propose design values" (owner,
  # 15 October 2026, #375) and taken by the owner with the plan's other recommendations (17 October 2026, #782): twice
  # the yishuv build's pair, for the largest land work in the MVP. The owner may override them.
  dry_swamp_turns: 12                    # DESIGN VALUE: owner-set balance, not a historical figure (D-D6, Q-SWAMP-1: turns draining the Hula takes)
  dry_swamp_cost: 4                      # DESIGN VALUE: owner-set balance, not a historical figure (D-D6, Q-SWAMP-1: IL millions, paid at the start)
  field_food_per_turn: 1                 # DESIGN VALUE: owner-set balance, not a historical figure (D-D10, Q-FIELD-3: Food units a plowed field adds each turn)
  # Q-COR-6 (agents propose design values, owner, 15 October 2026, #375; taken with the plan on 17 October 2026, #782): a
  # road segment as the Mefalsim build, a rail segment twice its Money at the same pace. Not historical figures.
  road_segment_turns: 6                  # DESIGN VALUE: owner-set balance, not a historical figure (D-D7, Q-COR-6: turns a road segment takes)
  road_segment_cost: 2                   # DESIGN VALUE: owner-set balance, not a historical figure (D-D7, Q-COR-6: IL millions, paid at the start)
  rail_segment_turns: 6                  # DESIGN VALUE: owner-set balance, not a historical figure (D-D7, Q-COR-6: turns a rail segment takes)
  rail_segment_cost: 4                   # DESIGN VALUE: owner-set balance, not a historical figure (D-D7, Q-COR-6: IL millions, paid at the start)
  objective_points: 100                  # DESIGN VALUE: owner-set balance, not a historical figure (D-B6: points per objective)
  objective_bonus_per_month: 5           # DESIGN VALUE: owner-set balance, not a historical figure (D-B2: points per month early)
  objective_bonus_cap: 60                # DESIGN VALUE: owner-set balance, not a historical figure (D-B2: the most bonus points)
  # The national Food stock (R-FOOD-001 to R-FOOD-003; D-C2; maintainer decision of 16 October 2026, #382, option (b)). One
  # unit of Food is one population unit's food for a month. The stock starts at, and the monthly import keeps it at,
  # cover_months of the nation's need; the foreign currency crisis cuts the import. food_supply carries the term
  # k_food * (stock - full) / full, minus the unmet share: an empty stock costs k_food, which with food_supply's base of
  # 0.40 is the food_crisis line (0.30, model.yaml), so an empty larder brings the country to the edge of a food crisis
  # and any unmet need tips it over. Three months' cover makes one month of the 1951 crisis cost a third of k_food,
  # about what the legacy direct cut (fx_crisis_food_per_year) took over the same weeks. Neither is a historical figure.
  k_food: 0.1                            # DESIGN VALUE: owner-set balance, not a historical figure (D-C2: food_supply lost with an empty Food stock)
  cover_months: 3                        # DESIGN VALUE: owner-set balance, not a historical figure (D-C2: months of need a full Food stock holds)
  # The national Arms pool (R-ARMS-001 to R-ARMS-003; D-G6). The owner approved these values on 18 October 2026 (#781:
  # "Approve as drafted"; the balance pass, #813, may tune them). One Arms is the equipment package for one recruit group of
  # people_per_unit (1,000) people, so training takes 1 a group. The reasoning, and the AI-drafted historical cross-check a
  # person still has to check, are in research_notes/GDD phase 2/arms-values-draft.md.
  arms_start: 12                         # DESIGN VALUE: owner-set balance, not a historical figure (D-G6, #781: Arms on 15 May 1948)
  arms_price: 1                          # DESIGN VALUE: owner-set balance, not a historical figure (D-G6, #781: IL millions per Arms)
  arms_per_recruit_group: 1              # DESIGN VALUE: owner-set balance, not a historical figure (D-G6, Q-UNIT-3: Arms per recruit group trained)
  # R-MONEY-004, "Spend the dollars" (#813; owner, 2 November 2026: "Spend the dollars (Recommended)",
  # https://github.com/eyalgolan/statecraft-israel/issues/813#issuecomment-6564883926): during the war the player may
  # convert up to $4m a turn of foreign currency into Money, $100m in all; the reserves run out sooner, and the sourced
  # 1951 currency crisis (fx_start, capital_inflow_per_year below) and its Food-import cut come closer. $100m is the
  # Export-Import Bank's 1949 credit (FRUS 1949 vol. VI, editorial note d407,
  # https://history.state.gov/historicaldocuments/frus1949v06/d407).
  dollars_per_turn: 4                    # DESIGN VALUE: owner-set balance, not a historical figure (D-C3, #813: US$ millions a turn)
  dollars_total: 100                     # DESIGN VALUE: owner-set balance, not a historical figure (D-C3, #813: US$ millions in all)
  # The rate, IL millions per US$ million, each from its day on (sourced on #813, comment 6565039439). To 18 September
  # 1949 the effective dollar rate was IL 0.333 per $ ($3.00): the official sterling parity gave IL 0.248 per $, but most
  # dollar transactions ran at 0.333. The devaluation of 19 September 1949 set IL 0.357 per $ ($2.80), in force to 14
  # February 1952. Sources: IMF Annual Report 1950, Appendix I, pp. 88-89; Federal Reserve Bulletin, November 1949, p. 1336;
  # Michaely, Israel's Foreign Exchange Rate System (1971), p. 10 and Table 2-1; Patinkin, The Israel Economy: The First
  # Decade, Table 43, p. 121. The economy comment below rounds the second rate to 0.36; the lever uses 0.357.
  il_per_dollar:
    - { from: 1948-05-15, rate: 0.333, status: sourced }
    - { from: 1949-09-19, rate: 0.357, status: sourced }
  # Agent-proposed values accepted by the owner on 19 October 2026 (Q-GEN-1, Q-UNIT-2, Q-UNIT-3):
  # https://github.com/eyalgolan/statecraft-israel/pull/1130#issuecomment-6269103829
  # Upkeep is per formation, not strength:
  # IL 0.05m is one tenth of a single group's IL 0.5m training cost per turn. Six service turns match the veteran pioneer
  # threshold as an accepted initial design value; neither number asserts a historical cost or period.
  military_upkeep_per_turn: 0.05         # DESIGN VALUE: owner-set balance, not a historical figure (D-C6, Q-UNIT-3: IL millions per formation per turn)
  veteran_soldier_service: 6             # DESIGN VALUE: owner-set balance, not a historical figure (D-D9, Q-UNIT-2: service turns, excluding training)
  # The veteran pioneer (R-VET-001, R-VET-002; D-D9). Agent-proposed design values the owner took on 17 October 2026 (#782:
  # "take all of the plan's recommendations"; Phase 4 plan Q-VET-1, Q-VET-2). Service is build_yishuv_turns, so one
  # completed yishuv makes a veteran pioneer.
  veteran_halutz_service: 6              # DESIGN VALUE: owner-set balance, not a historical figure (D-D9, Q-VET-1: turns of work to become a veteran)
  veteran_progress: 2                    # DESIGN VALUE: owner-set balance, not a historical figure (D-D9, Q-VET-2: a veteran's turns of progress a turn)
  # Housing condition (R-HOUSE-001, R-HOUSE-002; D-D1, D-D2): each city's average condition, 0 to 1. Every city starts at
  # model.yaml's housing base, so the cities start where the national level does (Q-START-3), and each turn a city loses
  # overflow_condition_loss times its overflow share (people beyond its housing over its people). 0.15 is model.yaml's
  # homes_lift_morale k, borrowed: no closer comparable exists (Q-HOUSE-1). Agents proposed both numbers ("Agents propose
  # design values", owner, 15 October 2026, #375); the owner may override them. Neither is a historical figure.
  housing_start_condition: 0.35          # DESIGN VALUE: owner-set balance, not a historical figure (D-D2, Q-START-3: every city's start condition)
  overflow_condition_loss: 0.15          # DESIGN VALUE: owner-set balance, not a historical figure (D-D1, Q-HOUSE-1: condition lost a turn at a full overflow share)
  # The economic defeat (R-END-001, R-END-002; D-B5). Money is exhausted when the debt is at or above debt_ceiling (Q-DEF-1,
  # owner, 17 October 2026, #784: "a debt ceiling", first value IL 250m, as recommended), Food when the national stock is
  # 0; defeat_turns turn ends in a row with both end the game (Q-DEF-2, owner, 18 October 2026, #784: "3 turns"). Neither
  # is a historical figure; the Phase 7 balance pass (PR 12) sets the final ceiling.
  defeat_turns: 3                        # DESIGN VALUE: owner-set balance, not a historical figure (D-B5, Q-DEF-2: turn ends in a row with Money and Food exhausted)
  debt_ceiling: 250                      # DESIGN VALUE: owner-set balance, not a historical figure (D-B5, Q-DEF-1: IL millions of debt at which Money is exhausted)
  # Standing orders (R-ORDER-002, R-ORDER-003; Q-HOLD-1 "(a) 1 turn, free, lost on move" and Q-HOLD-2 "(a) Own hex, no
  # interception", owner, 15 October 2026, #375). A unit's defence in percent of its strength. Proposed in Phase 5 PR 7
  # (#746): fortify 125 was decided by the owner on 19 October 2026 (PR #1071, https://github.com/eyalgolan/statecraft-israel/pull/1071#issuecomment-6259021297); hold stays at 100
  # (it adds nothing, it keeps the position). Not play-tested (combat is PR 8); the balance pass (#813) may tune both.
  fortify_defence_percent: 125           # DESIGN VALUE: owner-set balance, not a historical figure (D-G4, Q-HOLD-1: a fortified unit's defence, percent; owner-decided 19 Oct 2026, #1071)
  hold_defence_percent: 100              # DESIGN VALUE: owner-set balance, not a historical figure (D-G4, Q-HOLD-2: a holding unit's defence, percent)
  # Delegation and loyalty (R-EVT-003; #107, owner, 18 October 2026: "Yes: minister loyalty changes a delegated event's
  # outcome, with labelled values in the PR"). A minister below the floor still picks as competence says, but carries the
  # decision out half-heartedly: only the share of each numeric effect lands. The floor is 5 because people.yaml rates
  # the deputies 5 as a neutral rating, so neutral is not disloyal. Agent-proposed; the owner may override them.
  delegation_loyalty_floor: 5            # DESIGN VALUE: owner-set balance, not a historical figure (D-H1, #107: loyalty below which a delegated decision is half-done)
  operation_lapse_security: -0.01        # DESIGN VALUE: owner-set balance, not a historical figure (D-H1, R-ENEMY-007, #2543: the Security a lapsed operation costs, Horev's Hold; the owner chose it on 7 November 2026, quoted in docs/rules/mvp.md)
  operation_truce_break_standing: -0.10  # DESIGN VALUE: owner-set balance, not a historical figure (D-H1, R-ENEMY-007, #2543: the International Support a late operation launched inside an accepted truce costs, as Latrun's offensive; the owner chose it on 7 November 2026, quoted in docs/rules/mvp.md)
  operation_wait_turns: 6                # DESIGN VALUE: owner-set balance, not a historical figure (D-H1, R-ENEMY-007, #2543: turns a postponed operation waits for its units; the owner chose 6 on 7 November 2026, quoted in docs/rules/mvp.md)
  disloyal_effect_share: 0.5             # DESIGN VALUE: owner-set balance, not a historical figure (D-H1, #107: share of each numeric effect a disloyal minister delivers)
  # Movement (R-MOVE-001, R-MOVE-003; Q-MOVE-1, Q-MOVE-3, owner, 15 October 2026, #375: "Take all recommendations").
  # The numbers below were proposed in PR #1073 (#744) and accepted by the owner as design values on 19 October 2026:
  # https://github.com/eyalgolan/statecraft-israel/pull/1073#issuecomment-6259021508
  # The smallest integer spread that separates the plain from the hills from the mountains; lake and sea have no cost,
  # so they are impassable. Three units of a side per hex: a brigade of battalion-size formations.
  move_cost:
    coastal_plain: 1                     # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-1: movement points to enter the hex)
    valley: 1                            # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-1: movement points to enter the hex)
    lowland: 1                           # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-1: movement points to enter the hex)
    hills: 2                             # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-1: movement points to enter the hex)
    desert: 2                            # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-1: movement points to enter the hex)
    mountains: 3                         # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-1: movement points to enter the hex)
  stack_limit: 3                         # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-3: units of one side a hex holds)
  # Roads and rail (R-MOVE-007; Q-MOVE-2, owner, 15 October 2026, #375: option (a), only the corridors Phase 4 builds).
  # Proposed in Phase 5 PR 6 (#745) under "Agents propose design values"; the owner may override them. 1 is the lowest
  # whole cost, so a built segment halves a hills step and thirds a mountain one, but on the coastal plain, valley and
  # lowland (already 1) it saves nothing. Not historical figures.
  road_move_cost: 1                      # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-2: movement points for a step along a built road segment)
  rail_move_cost: 1                      # DESIGN VALUE: owner-set balance, not a historical figure (D-G2, Q-MOVE-2: movement points for a step along a built rail segment)
  # R-MOVE-005 (Q-CIV-1, owner, 15 October 2026: "No entry until Phase 6"): the settlements no player unit may enter or
  # cross. DRAFT, UNCHECKED: the owner accepted this list as a draft on 19 October 2026 (https://github.com/eyalgolan/statecraft-israel/pull/1073#issuecomment-6259083180).
  # No person has yet checked each place's control on 15 May 1948; research task #1096 will check it. Do not treat it as history.
  enemy_held_localities: [acre, nazareth, lod, ramla, gaza, beersheba]
  # Combat (R-COMBAT-001 to R-COMBAT-006; Q-COMBAT-1 to Q-COMBAT-4, owner, 15 October 2026, #375: "Take all
  # recommendations"). PROPOSED in Phase 5 PR 8 (#747) and accepted by the owner as design values on 20 October 2026
  # (all six defaults, asked at
  # https://github.com/eyalgolan/statecraft-israel/issues/747#issuecomment-6297295556, ruling at
  # https://github.com/eyalgolan/statecraft-israel/issues/747#issuecomment-6299184133). They stay labelled design values
  # for the balance pass (#813). None is a historical figure. Each engagement draws once from the RNG: a face 1 to 6, the six
  # entries of its row. A row applies from its min_ratio_percent (the attackers' summed strength over the defenders'
  # strength after fortify and terrain, in percent) up to the next row's: below 1:2, 1:2, 1:1, 2:1 and 3:1 and up.
  # Losses are population units; defender_retreats 1 sends the defenders that are left one hex back.
  arms_per_attack: 1                     # DESIGN VALUE: owner-set balance, not a historical figure (D-G6, Q-COMBAT-4: Arms each Israeli unit spends on an attack)
  combat_terrain_percent:
    coastal_plain: 100                   # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-2: a defender's strength on this biome, percent)
    valley: 100                          # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-2: a defender's strength on this biome, percent)
    lowland: 100                         # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-2: a defender's strength on this biome, percent)
    hills: 125                           # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-2: a defender's strength on this biome, percent)
    desert: 100                          # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-2: a defender's strength on this biome, percent)
    mountains: 150                       # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-2: a defender's strength on this biome, percent)
  combat_table:
    - min_ratio_percent: 0               # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-1: the attack : defence ratio the row starts at, percent)
      attacker_loss: [2, 2, 1, 1, 1, 1]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the attackers lose, per face)
      defender_loss: [0, 0, 0, 0, 0, 1]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the defenders lose, per face)
      defender_retreats: [0, 0, 0, 0, 0, 0]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: 1 where the defenders retreat, per face)
    - min_ratio_percent: 50              # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-1: the attack : defence ratio the row starts at, percent)
      attacker_loss: [2, 1, 1, 1, 1, 0]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the attackers lose, per face)
      defender_loss: [0, 0, 0, 1, 1, 1]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the defenders lose, per face)
      defender_retreats: [0, 0, 0, 0, 0, 0]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: 1 where the defenders retreat, per face)
    - min_ratio_percent: 100             # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-1: the attack : defence ratio the row starts at, percent)
      attacker_loss: [1, 1, 1, 1, 0, 0]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the attackers lose, per face)
      defender_loss: [0, 0, 1, 1, 1, 1]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the defenders lose, per face)
      defender_retreats: [0, 0, 0, 0, 0, 1]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: 1 where the defenders retreat, per face)
    - min_ratio_percent: 200             # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-1: the attack : defence ratio the row starts at, percent)
      attacker_loss: [1, 1, 0, 0, 0, 0]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the attackers lose, per face)
      defender_loss: [0, 1, 1, 1, 1, 2]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the defenders lose, per face)
      defender_retreats: [0, 0, 0, 1, 1, 1]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: 1 where the defenders retreat, per face)
    - min_ratio_percent: 300             # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-1: the attack : defence ratio the row starts at, percent)
      attacker_loss: [1, 0, 0, 0, 0, 0]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the attackers lose, per face)
      defender_loss: [1, 1, 1, 2, 2, 2]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: population units the defenders lose, per face)
      defender_retreats: [0, 1, 1, 1, 1, 1]  # DESIGN VALUE: owner-set balance, not a historical figure (D-G3, Q-COMBAT-3: 1 where the defenders retreat, per face)
  # The enemy's simple goals (R-ENEMY-003; Q-ENEMY-2, owner, 15 October 2026, #375: "Take all recommendations"). PROPOSED in
  # Phase 5 PR 10 (#749) under "Agents propose design values"; the owner may override it. An enemy unit off its script
  # attacks an adjacent player unit when the attack preview's ratio reaches this, else it holds its hex fortified. 200 is
  # the 2:1 row of combat_table, the first where the defenders can be made to retreat. Not a historical figure.
  enemy_attack_ratio_percent: 200        # DESIGN VALUE: owner-set balance, not a historical figure (D-G7, Q-ENEMY-2: the least attack : defence ratio an enemy unit attacks at, percent)
  # Prisoners (R-POW-002, R-POW-004, R-POW-005; Q-POW-2, Q-POW-4, Q-POW-5, owner, 15 October 2026, #375). The amounts are
  # agent defaults posted on #751 (https://github.com/eyalgolan/statecraft-israel/issues/751#issuecomment-6318787961)
  # and accepted by the owner as design values (owner, 21 October 2026, #751:
  # https://github.com/eyalgolan/statecraft-israel/issues/751#issuecomment-6321366548); none is a historical figure. A prisoner
  # unit eats the ration any population unit eats (one Food is one unit's food for a month, R-FOOD-001). A return scores a
  # tenth of an objective (D-B6) per unit. The player's soldiers held cost one month of income_per_job_per_year (0.06 /
  # 12) a unit, in Money, so Q-CIV-3 keeps the Food shortfall as its one cohesion exemption.
  food_per_prisoner_unit: 1              # DESIGN VALUE: owner-set balance, not a historical figure (D-C2, Q-POW-5: Food a population unit of prisoners eats a turn)
  prisoner_shortfall_cohesion: 0.01      # DESIGN VALUE: owner-set balance, not a historical figure (D-C1, Q-POW-5: Social cohesion lost per unit of the prisoners' Food not given)
  points_per_prisoner_unit_returned: 10  # DESIGN VALUE: owner-set balance, not a historical figure (D-B6, Q-POW-2: points per population unit of prisoners returned)
  held_soldier_cost_per_turn: 0.005      # DESIGN VALUE: owner-set balance, not a historical figure (D-C6, Q-POW-4: IL millions a turn per population unit of our soldiers held)
  # Occupation (R-CTRL-011, R-CTRL-012; Q-OCC-1 (a) and Q-OCC-3 (a), owner, 18 October 2026, #783 batches 3 and 4: one unit
  # stays with Money upkeep per turn, one X for every place, both "labelled design values proposed by agents for owner
  # override"). Accepted by the owner on 22 October 2026 (https://github.com/eyalgolan/statecraft-israel/pull/1324#issuecomment-6338594839).
  # X is 3 monthly turns: the owner ruled "3 turns, keep upkeep" on 27 October 2026 (#1712,
  # https://github.com/eyalgolan/statecraft-israel/issues/1712#issuecomment-6461246243) and confirmed "3 turns (27 Oct)" on
  # 1 November 2026, replacing the 6 turns accepted on 22 October. The upkeep
  # is derived as held_soldier_cost_per_turn is (#751): one month of income_per_job_per_year (0.06 / 12) for the one
  # population unit that stays. Neither is a historical figure.
  occupation_turns: 3                    # DESIGN VALUE: owner-set balance, not a historical figure (D-F4, Q-OCC-3: X, the turns every occupation lasts)
  occupation_upkeep_per_turn: 0.005      # DESIGN VALUE: owner-set balance, not a historical figure (D-F4, Q-OCC-1: IL millions a turn per occupied place)
  # Deport (R-CTRL-018; owner ruling 1, "honest cost, no reward"; Q-DEP-5 (c), owner, 18 October 2026, #783 "Accept all 16
  # drafts": fixed costs to Social cohesion and International Support, a fixed International Support drag a turn and a lower
  # International Support ceiling for the rest of the game; "sizes are labelled design values the balance pass sets;
  # DeportNeverDominates must hold with a margin"). Proposed on #800 and accepted by the owner on 22 October 2026, per
  # place, with the Money cost below (#1327: https://github.com/eyalgolan/statecraft-israel/pull/1327#issuecomment-6350226175);
  # design values the balance pass (#813) may revisit. "Large" is read as 1.5
  # times the largest single event move on either parameter in content (0.10). The drag and the ceiling drop count per
  # place deported (DisplacedByPlayer counts places, Q-CAP-2 (b)). None is a historical figure.
  deport_cohesion_cost: 0.15             # DESIGN VALUE: owner-set balance, not a historical figure (D-C1, Q-DEP-5: Social cohesion lost at once)
  deport_support_cost: 0.15              # DESIGN VALUE: owner-set balance, not a historical figure (D-C1, Q-DEP-5: International Support lost at once)
  deport_support_drag_per_turn: 0.005    # DESIGN VALUE: owner-set balance, not a historical figure (D-C1, Q-DEP-5: International Support lost each turn after, per place)
  deport_support_ceiling_drop: 0.25      # DESIGN VALUE: owner-set balance, not a historical figure (D-C1, Q-DEP-5: International Support ceiling lowered for good, per place)
  # Deport's Money cost (owner decision, 22 October 2026: https://github.com/eyalgolan/statecraft-israel/pull/1327#issuecomment-6337946480):
  # set so that Deport's Money is at or below Occupy's and Raid's in every state. Occupy pays at most
  # occupation_upkeep_per_turn for all 44 turns (0.22); a depopulated hex shared with a city takes no new project, so the
  # city may spend less, and the dearest project is housing (6, projects.yaml). 0.22 + 6, rounded up to a quarter million.
  deport_money_cost: 6.25                # DESIGN VALUE: owner-set balance, not a historical figure (D-C6, #1327: IL millions paid at once for a Deport)
  # Deport stays out of play until the owner has reviewed its cards (Phase 6 PR 9) and its dialog (PR 10): with this off,
  # ResolveCapture(..., deport) is refused as not built yet (R-CTRL-007; plan, PR 8 sensitivity gate). PR 10 lifts it.
  deport_in_play: false
# Coalition stability lost by replacing a minister (Cabinet.Appoint): a seat taken from the minister's party, or a
# reshuffle within it.
appoint_stability_cost_other_party: 0.06
appoint_stability_cost_same_party: 0.02
# Map actions (MapActions): from 1950, on land within one hex of a settlement. Transit camps also need the ma'abarot
# policy (the first ma'abara, Kesalon, opened in May 1950: https://en.wikipedia.org/wiki/Ma'abarot, checked 2026-10-10).
# Costs (IL millions) and effects are game balance: a camp is cheap and grim, permanent housing dear and welcome.
map_actions_from: 1950-01-01
map_action_max_distance: 1
map_actions_per_turn: 2
# name: the phrase the player reads in a sentence ("a transit camp"); required.
map_actions:
  transit_camp: { cost: 1.5, housing: 0.02, morale: -0.005, name: a transit camp }
  housing: { cost: 6, housing: 0.03, morale: 0.01, name: a housing project }
rung_attention_step: 0.25
rung_min_mobilization: [0.02, 0.05, 0.15, 0.4, 0.8]
ceasefire_weeks: 52               # long enough that the historical armistice events end the war first
# ---- Economy (calibrated in M3; checked by sim/Statecraft.Sim.Tests/EconomyCalibrationTests.cs) ----
# Units: treasury and debt in IL millions; fx_start, exports, donations/loans and imports in US$ millions.
# Formal rate 1950-51: IL 0.36 per US$ (IL 1 = $2.80), Michaely, Foreign Trade Regimes and Economic Development: Israel
# (NBER 1975) Table 5-1, https://www.nber.org/system/files/chapters/c4560/c4560.pdf (checked 2026-10-10).
treasury_start: 20
fx_start: 40
# Taxes finance the ordinary budget: FY 1950/51 IL 60 million, approved 29 June 1950
# (https://en.wikipedia.org/wiki/1950_in_Israel, checked 2026-10-10). 60m / ~1.10m Jews (CBS average 1950) = 0.0000545.
tax_per_capita_per_year: 0.0000545       # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: derived ordinary-revenue abstraction retained by #884)
# Spending above taxes is borrowed. The gap is set so 1950-51 borrowing matches the rise in internal debt, IL 65.4m (1950)
# + 33.0m (1951) = 98.4m (Michaely Table 5-2, same chapter). The 1951 fiscal tightening is not modelled, so the model
# matches the two-year total, not each year. Job income (R-CITY-002) also reaches the Treasury, so the gap must cover it
# too (owner, 16 October 2026, #821, option (a)): the twelve cities' 211 employed units (cities.yaml; HistoricalAI assigns
# no workers) x mvp.income_per_job_per_year 0.06 x 2 years = 25.32m, and (98.4m - 6m reserve call-up + 25.32m) / ~2.40m
# person-years = 0.00004905 per person. Spending = tax 0.0000545 + 0.00004905 = 0.00010355, rounded to 0.0001036.
# (M3's value was 0.0000545 + 0.0000385 = 0.000093, derived before job income existed.)
spending_per_capita_per_year: 0.0001036  # DESIGN VALUE: owner-set balance, not a historical figure (D-C6: derived debt-path calibration retained by #884)
# Exports cover under a third of imports (spec section 3; EconomyCalibrationTests checks it every turn of 1950-51).
exports_per_year: 40
imports_per_capita_per_year: 0.00025
# Donations, loans and transfers, set so reserves run out in the last quarter of 1951: the October 1951 crisis
# (FRUS 1951 v.V doc 513, https://history.state.gov/historicaldocuments/frus1951v05/d513, checked 2026-10-10).
# History of this value: 192 (first fit, brief) -> 193.8 (re-fitted after Task 4 moved the Israel Bonds inflow to its
# 10 May 1951 launch; consistency anchor: first fx crisis in H2 1951, FRUS d513, 19 Oct 1951) -> 200. The 193.8 fit was
# knife-edged (valid range about 0.12 wide) because the model left out the documented drain of Q4 1951: about $45m of
# short-term credit matured Oct-Dec 1951 (FRUS d513), now applied by both options of forex_crisis_1951. With that drain the
# anchor (reserves not exhausted before the turn covering 19 Oct 1951; at its end below $45m; exhausted by the turn that
# covers the end of 1951) held for 193.78 to 206.62 (measured by bisection); 200 is the centre. Re-measured after the 1948
# war-period events (playtest 1 E, which moved the turn grid by a day): the range is 194.00 to 206.83 (bisection), and the
# margins at 200 are reserves $18.2m at the turn ending 1951-05-27 and $20.9m at the turn covering 19 Oct (against $45m
# falling due); first crisis in the turn ending 1952-02-27 (the one covering 31 Dec 1951); reserves -36.8 at the end of
# the run. The value was not retuned. Re-measured at the GDD cutover (monthly turns, play ending 15 January 1952): the
# anchor holds for 190.09 to 202.81 (bisection); at 200, reserves are $29.7m at the turn ending 1951-05-15 and $35.2m at
# the turn covering 19 Oct (ending 1951-11-15), against $45m falling due; the first crisis is in the turn ending
# 1951-12-15; reserves -13.0 at the game's end. The value was not retuned.
capital_inflow_per_year: 200
# While reserves are exhausted, food imports are cut: food_supply falls by this much per year. The legacy loop only: in the
# GDD loop the crisis cuts the Food stock's imports instead (R-FOOD-002, Q-POOL-4), and food_supply follows the stock.
fx_crisis_food_per_year: -0.2
mobilization_cost_per_year: 60
# Derived from CBS table 2.1 (1949): (end-1949 1,013.9k - end-1948 758.7k - 1949 immigrants 239,954) / average 1949 population 901.0k.
natural_increase_per_year: 0.0169
# Derived: CBS end-1948 Jewish population 758,700 minus 101,828 immigrants (15 May-31 Dec 1948) minus natural increase over
# 7.5 months at the rate above. Checked against the 8 Nov 1948 census by the calibration test.
# Arithmetic: 758,700 - 101,828 = 656,872. Natural-increase base = average of 656,872 and 758,700 = 707,786.
# Natural increase = 707,786 x 0.0169 x 7.5/12 = 7,476. 656,872 - 7,476 = 649,396, rounded to the nearest hundred = 649,400.
initial_population: { veterans: 649400 }
budget_start:
  defense_minister: 0.45
  finance_minister: 0.05
  # The Immigration line. Its id stays `prime_minister` (ids are stable: a rename would move every golden hash and break
  # the replay of saved budget orders), but it was the Ministry of Immigration's line until 8 October 1951:
  # budget_owners below says which office owns it on each date, and the Budget screen shows that office.
  prime_minister: 0.15
  labour_minister: 0.10
  health_minister: 0.07
  education_minister: 0.06
  agriculture_minister: 0.07
  interior_minister: 0.05
# Who owns a budget line, by date (issue #108). A line with no entry here belongs to the office with its own id for the
# whole era. The first owner starts on the era's first day; each later owner starts on its `from`; every owner's office
# must exist on its dates (the loader checks). The Ministry of Aliyah (immigration_minister) ends on 8 October 1951: the
# Knesset's government-positions table (KNS_PersonToPosition, GovernmentNum 2, read 2026-10-12) lists "Minister of
# Aliyah" (GovMinistryID 784) for Haim-Moshe Shapira (PersonID 2661) from 1 November 1950 to 8 October 1951, and
# GovernmentNum 3 (from 8 October 1951) lists no such ministry. The same table lists no immigration or absorption
# portfolio for the Prime Minister (Ben-Gurion, PersonID 2198) in the Third Government. That the line then sits with the
# Prime Minister is an assumption (research notes M3 1950-51, Addendum (a), unverified), as in M3: absorption is
# assumed to stay with the Jewish Agency and the Prime Minister. Re-point the second owner if a source shows otherwise.
# Source: https://knesset.gov.il/Odata/ParliamentInfo.svc/KNS_PersonToPosition?$filter=GovernmentNum%20eq%202 (and eq 3).
budget_owners:
  prime_minister:
    - { from: 1948-05-15, office: immigration_minister }
    - { from: 1951-10-08, office: prime_minister }