





{"success":true,"content":"# Mechanics learning index\n\nTemplate revision: 7. These are optional lessons for the decision in front of you. Follow the [operations cycle](MCP_Agent_Operations.md); do not reread every lesson each session. Keep campaign-specific findings in your private [learning file](MCP_Agent_Learning.md).\n\nGame rules and authenticated observations outrank remembered recipes. Examples below are **hypothetical**, with stated inputs using current default rules; they are not observations, guarantees, reservations or instructions to spend. Refresh the player's class, technology, occupation, taxes and quotes before using them. Never add Credits to Exium or assume an exchange rate.\n\n| Decision | Lesson | Useful player observation |\n| --- | --- | --- |\n| Find a constraint | Population, mines and production | `query.operations_snapshot`, colony observations |\n| Price a next step | Costs and class effects | `query.build_cost`, production/research quotes |\n| Support another world | Colony command and penalties | Support state, building effects |\n| Avoid stalled wealth | Storage and accrual | `query.economic_forecast`, local headroom |\n| Make output usable | Logistics | Cargo, route, pickup and delivery history |\n| Reach a legal target | Movement and protection | `query.move_preview`, visible map/intelligence |\n| Commit military resources | Combat command and conquest | Designs, full reports, [combat playbook](MCP_Agent_Combat.md) |\n| Buy a new capability | Research | Designs, compatibility, queues and complete quotes |\n\n## Population, mines and production\n\n**Prerequisites:** an owned colony's housing/development, mines, concentration, class, occupation and storage observations.\n\nHousing/population produces Credits. Mines produce Exium once the development requirement is met. Barracks, factories and orbital facilities reduce appropriate unit-production costs; a factory is not passive Credit income. Laboratory efficiency concerns research. Class bonuses and technology can change the comparison.\n\n**Worked example — hypothetical:** an unoccupied ordinary colony with population 1,600 and no city multiplier produces 100 Credits over one complete default tick (`1600 / 16`). Population 2,000 would produce 125, a marginal 25. At sufficient development, 100 mines with concentration 5 produce 10 Exium per default tick; a base Miner multiplier of 1.5 gives 15. These are separate yields, before storage limitations. Obtain the actual price and effects of adding housing or mines rather than assuming the extra income pays for its own construction.\n\n**Common mistake:** treating facilities or public empire totals as spendable local income. **Prediction exercise:** identify which resource reaches the next funded goal first and what changes if the world is occupied. **Decision implication:** buy the missing useful output, not the building with the most impressive level.\n\n## Specialized production and research worlds\n\n**Prerequisites:** owned eligible colonies, their development capacity and relevant local facility counts, occupation, current blueprint and quantity, funding and delivery observations.\n\nUse the [specialization procedure](MCP_Agent_Economy.md#specialized-production-and-research-worlds) to dedicate worlds selected for facility construction to their chosen facility type and choose the matching site. Very Rich worlds serving other purposes need no facility specialization. Barracks train soldiers, factories manufacture ground vehicles, orbital facilities manufacture ships and nuke plants manufacture nuclear units. Laboratories discount research, including ship/nuclear blueprints and TECH upgrades. Facilities on different planets do not pool. Very Rich gives development room rather than an extra price multiplier.\n\n**Worked example — controlled calculator inputs:** the same space design has hull 100, engine 100 and stored build points 1,000. Manufacture three units with ordinary class effects, no occupation, no tax and no support penalty. At 1,000 orbitals the ordinary calculator charges 1,500 Credits and 699 Exium; at 5,000 it charges 500 Credits and the same 699 Exium. More barracks or higher Exium concentration does not change this ship-production discount. Research-site laboratories are a separate comparison. These are calculator examples, not a reservation or live quotes.\n\n**Common mistake:** building on whichever world was last inspected, adding every facility everywhere, assuming all empire facilities pool, or expecting orbital facilities to lower manufacturing Exium. **Prediction exercise:** compare the same order at the current orbital hub and a candidate; then re-quote after occupation or a delivery deadline changes. **Decision implication:** concentrate useful facility investment and select the best feasible quoted site for the actual job, including transport, funding and defense.\n\n## Costs and the four classes\n\n**Prerequisites:** current class, unlocked technology, tax/penalty and an exact quote for the intended quantity.\n\nExpander improves colony command scaling; Miner improves mining/storage opportunities; Technologist improves research efficiency; Warmonger improves unit-production costs. Their unlocked branches matter. None replaces the user's chosen identity or mandates one universal strategy. Production facilities, laboratory reductions, class effects, taxes and rounding act on specific calculations; do not spread one discount to every action.\n\n**Worked example — hypothetical:** a space design with hull 100, engine 100 and build points 1,000, built in quantity three with 1,000 orbital facilities, 25% penalty and 20% tax, costs an ordinary player 2,175 Credits and 699 Exium. Base Warmonger production costs are 1,849 and 594 after the 15% discount and rounding; unlocked War Economy gives 1,631 and 524 after the 25% discount. Both currencies are reduced. This example is a production calculation, not the cost of researching that design.\n\n**Common mistake:** using the upgraded discount before it is unlocked, or treating cheaper production as free logistics. **Prediction exercise:** compare the exact quote with an available alternative design and reserve requirements. **Decision implication:** compare complete feasible packages in both currencies, including the option to keep reserves.\n\n## Colony command and the rounded penalty\n\n**Prerequisites:** owned-colony count, stored colony command, class, supported count and current penalty.\n\n[Colony command centres](Buildings_Guide.md) are separate from a unit's [combat command](Command.md). Support uses the stored command value established by the strongest command-centre site, not the sum of centres across worlds. The current owned-colony counter includes colony records; do not assume a scanner outpost is excluded. Inspect actual support before claiming or donating a world.\n\n**Worked example — hypothetical:** for a non-Expander with effective command 100, default support is `round(5 + 100 / 100 × 3) = 8` worlds. With 11 owned worlds, excess is three and the Credit penalty is `floor(3² / 2) / 100 = 4%`. Excess counts zero through five give 0%, 0%, 2%, 4%, 8%, 12%; it is not a flat charge per colony. The support penalty and tax are added in the production price factor before later rounding/discounts.\n\n**Common mistake:** constructing many smaller command sites to add their levels together, or interpreting one zero-penalty excess world as unlimited free expansion. **Prediction exercise:** compare support and penalty before and after two proposed acquisitions. **Decision implication:** include empire-wide cost changes and support investment in the expansion package.\n\n## Storage, accrual and forecasts\n\n**Prerequisites:** actual local balances, capacity for each resource, last accrual time, class/technology and competing spending/automation.\n\nBase capacity is 150,000 per resource, increased by silos and applicable effects. The same nominal silo count can have different capacity under occupation or technology. Normal growth clamps to capacity. Above-capacity stocks can decay after the accrual grace interval; an income display at capacity is not a substitute for actual stored-balance rules. Credits use integer arithmetic over elapsed complete ticks; Exium is rounded over the elapsed batch. Repeatedly adding a rounded one-tick display can give the wrong forecast. Read-only inspection does not persist a new production interval. A forecast that assumes intervening commands persist balances is a different scenario from simply observing more often.\n\n**Worked example — hypothetical:** at initial technology, no special capacity multiplier and no occupation, two silos give `150000 + 2 × 3000 = 156000` capacity. With occupation, the silo contribution is halved, giving 153,000. A separate ordinary world at capacity 150,000 with balance 149,990 and 100 arriving from one production tick stores only 10 more Credits. It ends at 150,000; the other 90 cannot accumulate. At the default population divisor, population 31 accrued over two ticks gives three Credits, while summing two rounded one-tick displays would suggest only two.\n\nUse `query.economic_forecast` with an owned `pla_id` and a required `horizon_seconds` from 60 to 86,400. Read assumptions, tick timing and coverage gaps. Automation, spending, combat and uncertain deliveries can invalidate a conditional projection. `query.build_cost` adds effects to a quote; it reserves nothing.\n\n**Common mistake:** assuming that visiting a full world creates room, or forecasting unmodelled deliveries as certain. **Prediction exercise:** predict the first capacity constraint and compare with actual balances over a controlled interval. **Decision implication:** storage, spending or delivery may matter more than another production building.\n\n## Production is not delivered throughput\n\n**Prerequisites:** authorized source and destination, pickup settings, resource eligibility, real free cargo space, route time, destination headroom and observed deliveries.\n\nTrace production → eligible pickup → loading → loaded travel → receiving capacity → usable funds. Check each resource separately, including competing automatic routes. A full cycle is verified pickup, delivery and repeat; a moving empty fleet proves none of those outcomes. Include fleet construction, fuel/resource costs where applicable, travel and reserves in the package.\n\n**Worked example — hypothetical:** a running route delivers 60,000 Exium every 30 ticks: delivered throughput is 2,000 per tick. Its source produces 3,000 per tick, so adding mines does not fix the demonstrated delivery bottleneck. A receiver with only 40,000 headroom cannot accept that whole automatic load; expand room, spend/redistribute stocks or choose another valid receiver before predicting another complete delivery.\n\n**Common mistake:** summing nominal mining income and calling it military purchasing power. **Prediction exercise:** identify whether the next failure is pickup eligibility, loading, travel or receiving space. **Decision implication:** compare more cargo, shorter routes, more headroom and reduced consumption before buying more source production.\n\n## Movement, protection and visible uncertainty\n\n**Prerequisites:** visible target, fleet composition/location, legal route, current memberships, account state and applicable protection.\n\nUse `query.move_preview` with the same selection and destination as `move_fleets`. It assesses current eligibility and ETA without launching, replacing mission drafts or revealing hidden defenders. A complete assessment returns `assessment_complete: true` with `eligible: true` or a known visible blocker (`eligible: false`). An unseen/missing target or unobserved wormhole entry can return `assessment_complete: false` and `eligible: null`, with an assessment issue or coverage gap and no hidden target reference or ETA. This means eligibility is unknown, not that normal movement is prohibited. If the campaign’s risk policy permits, the ordinary move command can still be attempted and will revalidate actual rules. Refresh before execution; accepted movement still does not prove arrival, orbital clearance or ground conquest.\n\n[Points protection](Points_Protection.md) uses effective empire scores, founder age/status and relevant war exceptions, not a remembered fixed score ratio. The ordinary age-based threshold drops by ten percentage points per complete day after its initial guard, with a 20% floor. The initial age guard, one-million-point rule and existing war relationships must be considered separately.\n\n**Worked example — hypothetical:** outside other exceptions, an original-founder target four complete days old has an 80% threshold. Against an attacker at 100,000 effective points, a target at 79,999 is below that threshold; 80,000 meets this size comparison. That arithmetic alone does not grant attack permission or establish safe defenses.\n\n**Common mistake:** treating a visible route or empty scanner view as combat permission or safety. **Prediction exercise:** preview a route and name the visible blocker or missing evidence that could change the plan. **Decision implication:** reconnaissance, diplomacy, a route upgrade or postponement may dominate an immediate launch.\n\n## Combat command and complete conquest\n\n**Prerequisites:** compatible designs, installed combat command, participating unit counts, reports and the [complete operation checklist](MCP_Agent_Combat.md).\n\nCurrent shared command is `round((floor(installed_command / 500))^0.8 × installed_command)` per contributing unit. The side's empire band averages the pooled contribution over all its participating units, rounds and caps at 50 command points. Defenders receive doubled pooled command. Command losses and added units can change that average; there is no private 10%-retained/90%-shared rule. This bonus is only one combat input.\n\n**Worked example — hypothetical:** installed command 499 contributes zero; 500 contributes 500; 1,000 contributes 1,741. An attacking band of 100 units, one with command 1,000 and the others zero, gets rounded average 17 (17%). With the defender multiplier and the same composition, it gets 35. Neither figure predicts victory without range, weapons, durability, targeting and the rest of the situation.\n\n**Common mistake:** maximizing a component or fleet count while omitting transports, ground delivery, reinforcement or garrison. **Prediction exercise:** explain how losing the sole commander changes the prediction and your abort condition. **Decision implication:** fund reconnaissance, orbital clearance, ground delivery, garrison and commissioning as one operation.\n\n## Research versus available designs\n\nUse the [canonical research playbook](MCP_Agent_Research.md) for the full decision, queue, unlock and deployment procedure.\n\n**Prerequisites:** available legal designs, compatibility, full research and production quotes, queues and a concrete capability gap.\n\nCompare immediate procurement, research plus production, trade/sharing, another target and postponement to fix an economic bottleneck. Include laboratory/facility requirements, queue delay, both currencies, transport and later field evidence. A purchased or researched blueprint is not a deployed capability.\n\n**Propulsion example — hypothetical:** a space design with hull 100 and raw engine 1 has zero effective propulsion: the space calculation first rounds engine divided by ten to zero. A hyperunit requires positive effective propulsion, so a positive raw engine field alone does not satisfy that prerequisite. Hull 50 with engine 100 also remains immobile because its effective hull is below the minimum. Hull 100 with engine 100 has positive effective propulsion, but still needs the ordinary component, technology and cost checks. Compare the quoted effective capability and movement preview, rather than assuming small positive fields make a usable scout.\n\n**Worked example — hypothetical prices, not game quotes:** an existing design costs 800 Credits and 90 Exium per unit; an alternative would cost 500 and 90 after research costing 5,000 Credits and 600 Exium. Ten units save only 3,000 Credits while consuming the extra research Exium and delay. Credit-only break-even is 17 units, ignoring other costs and tactical differences. Do not convert the 600 Exium into an invented Credit amount; a necessary new counter may still justify the investment for a different reason.\n\n**Common mistake:** buying a cheaper blueprint without enough remaining funds or compatibility to field it. **Prediction exercise:** state the observed deficiency, expected improvement, complete allocation and evidence that would reject the research hypothesis. **Decision implication:** buy the capability that advances the goal under current constraints, then revise from actual deployment and reports.\n\n## Empire TECH, fitting capacity and personal unlocks\n\n**Prerequisites:** authenticated empire TECH, own research sites and local balances, current pricing mode, personal class/Components/nodes, useful capability target and existing design quotes.\n\nEmpire TECH, design fitting/build points and personal Components are different quantities. The [technology tree](Technology_Tree.md) requires empire TECH tiers plus that player's Components and prerequisite nodes. Merely reaching 2,000/3,000/4,000 or receiving shared TECH does not unlock them automatically. Preserve class-branch progress; changing class is not an automatic funding strategy.\n\nSpace equipment capacity is rounded from `(hull - hangar) × (TECH / 100)`. More TECH permits more equipment in the same physical hull; it does not itself enlarge that hull. Higher TECH can reduce the build points stored for a newly researched equivalent design, affecting its later production cost. Existing blueprints retain their stored build points and existing units do not retrofit themselves. Compare exact research and production quotes before choosing richer designs or claiming a saving.\n\nUse `ge_research_read` / `query.technology_upgrade_quote` with owned `pla_id` and positive `quantity`. The price uses the ordinary upgrade calculation, laboratories, Technologist effectiveness, occupation and the current round's pricing mode. The active exponential curve adds a surcharge strictly above TECH 2,000 and doubles that multiplier every additional 1,000; legacy rounds do not. Bulk pricing averages endpoint costs before discounts/rounding, so separate increments can cost a different total. Affordability and eligibility are separate; the normal minimum researching-player score is 50,000. Quotes reserve nothing.\n\n**Worked example — hypothetical:** hull 1,000 with hangar 100 permits 9,000 fitting points at TECH 1,000 and 18,000 at TECH 2,000. Those are equipment capacity, not salvage Components. If an owned site quotes an increment costing 91,000 Credits and 7,000 Exium, has 120,000/8,000 local funds and must preserve 50,000/2,000 for an accepted operation, the available 70,000/6,000 leaves a funding gap of 21,000 Credits and 1,000 Exium before any transfer fee. A promised contribution does not close it. These illustrative prices are not game quotes.\n\n**Common mistake:** treating TECH 2,000 as universally unaffordable, assuming an alliance grants free automatic TECH/tree unlocks, or expecting old units to improve. **Prediction exercise:** name the next useful milestone, quote the intended batch at a real site, identify the complete funding/unlock/design/deployment gap, and compare independent funding with an explicit eligible sharing project. **Decision implication:** follow the [balanced roadmap](MCP_Agent_Strategy.md#empire-position-and-technology-roadmap); fund the useful capability while retaining survival, logistics and agreements, then verify actual TECH, personal unlocks and deployed benefit separately.\n"}

