Research and scientific personnel

Technology knowledge and availability

A technology definition is a row of FCT_TechSystem. It is not a race’s research progress. It holds a whole-number DevelopCost in research points (RP), a type (TechTypeID) and a category, two prerequisite IDs (Prerequisite1 and Prerequisite2), an owning race (RaceID), the flags AutomaticResearch, RuinOnly and ConventionalSystem, and the numerical AdditionalInfo values. A race knows a technology when it holds its own knowledge entry for it (FCT_RaceTech), and each entry carries an Obsolete flag. A race counts as knowing a technology when that entry exists. A separate check asks whether it is known and not obsolete. Marking a technology obsolete therefore does not erase the knowledge.

The available-project list on the Research tab of the Economics window is a specific eligibility query, not every unknown technology. For the selected research field it requires all of the following:

RuinOnly is false.
The technology is generic (RaceID = 0), belongs to this race, or is on the race’s list of foreign technologies made eligible for study.
Each nonzero prerequisite is already known to the race when the list is built.
The race has no knowledge entry for the technology, and no active project or queued entry for it at any of its colonies.

Prerequisite knowledge includes obsolete entries. Queued technologies do not satisfy prerequisites. The list looks only at technologies the race has already researched, so this window cannot build a chain of unknown prerequisites. The list does not filter on AutomaticResearch, ConventionalSystem or NoTechScan. Those flags serve other rules.

One deliberate exception lets a known technology reappear as a project. A component whose prototype state is research prototype can be listed when its technology is known, belongs to the selected field, and is neither active nor queued. The list adds “(P)” to its name. When it completes at a real colony, the component’s prototype state changes to not a prototype. A completion with no colony attached does not perform this conversion. The component design page covers component creation, prototype variants and development-cost recipes. Those definitions are blueprints, not physical component stock.

The Completed list shows known technologies in the selected field, except those with ConventionalSystem set. It does not hide obsolete knowledge. Absence from this list therefore does not prove that the race lacks the technology. When the game looks for the most advanced researched technology of a type, it normally picks the known one with the highest development cost. The numerical AdditionalInfo is not a universal technology ordering.

Ordinary research uses the definition’s stored DevelopCost. Starting a project does not recompute base-technology costs from game speed. Some player-designed definitions apply research speed when their cost is created, and others do not. The shared development-cost rule, cost = truncate(sqrt(build cost x 5000) x Research Speed / 100), applies it. See the component cost conventions. The Research Speed option of the Game Information window (FCT_Game.ResearchSpeed, 100 is normal) also multiplies colony RP output separately, as described below.

Laboratories output and wealth

A colony’s research capacity is sum over its installations of DIM_PlanetaryInstallation.ResearchValue x installation amount. The Research Facility (ID 8) has ResearchValue 1 in the game database as shipped. Lab assignment and the research step truncate this capacity to a whole number. It counts assignable facilities, not efficiency-adjusted RP. A worker shortage lowers the output of each assigned facility through colony efficiency. It does not shrink this lab count.

A separate count of effective research facilities takes the colony’s first Research Facility entry and multiplies its amount by the race economic production modifier, efficiency, the radiation production modifier, stability and the political status production modifier. Downloading technical data from a ship uses that count. A colony can receive the data only when the count is at least 1.

Annual base RP per facility is

B = species research rate modifier x race research rate x race economic production modifier x efficiency x radiation production modifier x stability x political status production modifier x (Research Speed / 100).

The species modifier is FCT_Species.ResearchRateModifier. The race research rate (FCT_Race.Research) starts at 200. Completing an Improved Research Rate technology (DIM_TechType ID 29) replaces it with that technology’s AdditionalInfo truncated to a whole number. All factors are exact decimals except the Research Speed conversion, which passes through a floating-point step. RP are not rounded to whole numbers.

The governor’s Production bonus and the sector governor’s Production bonus at quarter strength (1 + (bonus - 1) / 4) belong to the general production output modifier. That modifier multiplies those two bonuses by the species production rate, the economic production modifier, efficiency, the radiation production modifier, stability and the political status production modifier. It is absent from the research formula. Scientists supply their own research multiplier. Governors can still affect colony conditions through other systems, but the research formula gives them no direct laboratory Production bonus.

Efficiency starts at 1. Required manufacturing workers are shipyard workers + sum over installations of amount x DIM_PlanetaryInstallation.Workers. When required workers exceed the manufacturing population, efficiency becomes manufacturing population / required workers. A negative result is clamped to zero. With zero required workers, efficiency stays 1. See colony and population lifecycle for workforce, radiation, political status and stability. Production mechanics covers the other industrial outputs.

Let s be the scientist’s Research bonus as a multiplier, not a percentage. On a project outside the scientist’s specialty the scientist factor is s. In the specialty it is 4s - 3. A scientist displayed as 20% has s = 1.20, which gives 1.20 outside the specialty and 1.80 inside it. No penalty for the number of assigned labs appears in this formula. The Research Admin bonus is a lab limit, not another RP multiplier.

The project multiplier also includes and the race’s research-field bonus. The research step uses any construct on the colony’s body and does not check whether it is active. Race-wide field bonuses start at 1. Each active construct of the matching field at an owned colony with total population of at least 1 (population counts in millions) adds (construct ResearchBonus - 1) / 10. These contributions add up, and the resulting race factor then multiplies the local and scientist factors. The project list shows the local multiplier only for a construct marked Active, so an inactive local construct can make the displayed rate differ from the rate the research step applies.

For L assigned labs and y elapsed years, generated RP = B x L x effective project multiplier x y.

The wealth amount charged for the main project is B x L x y x (100 / Research Speed) x (1 / species research rate modifier). Ordinary positive research speed and species modifiers therefore cancel out of the base wealth cost, apart from rounding in the floating-point conversion. Scientist, construct and field multipliers improve RP per unit of wealth because they are absent from this cost. The other colony modifiers remain. A partial cycle records the full amount. A completing main project records only remaining RP x (wealth amount / generated RP). The research step does not require enough wealth in the treasury, does not impose a minimum RP spend, and does not round each cycle to whole RP. It records no research expense for NPR races. See wealth, income and treasury for the treasury consequences.

Assignments pause cancellation and queues

The scientist list on the Research tab shows the race’s active scientists with no current assignment, ordered by Research bonus and then by Research Admin. The Matching Scientists Only box narrows it to specialists in the chosen field. Refreshing this list also frees any scientist whose assignment says research project but who has no project attached.

The Create Project button requires a selected colony, technology and scientist, a colony whose total population is not 0, a positive whole number of labs, and no more labs than the stored count of available research facilities. Selecting a scientist sets the requested lab count to the lesser of the available labs and the scientist’s Research Admin, whether the box held a smaller or a larger number. Leaving the box clamps it to the Research Admin value. The Add RL button is disabled when the project’s labs are at or above Research Admin, or when the stored available count is 0. Remove RL is disabled at exactly one lab. The buttons themselves add or remove a lab without repeating these limits. These are window constraints, not rules the game enforces everywhere.

Starting a project does not itself check prerequisites, scientist availability, capacity or a positive remaining cost. It subtracts the first matching saved progress (FCT_PausedResearch.PointsAccumulated) from the whole-number development cost, removes that saved entry, then subtracts any decimal carry-over RP. It creates a new project ID, sets pause off, assigns the scientist to the project and the project to the scientist, moves the scientist to the colony and clears any transport ship. The resulting remaining RP (ResearchPointsRequired) is not clamped at zero.

Two different states are called “paused” in everyday use:

Action or state	Active project and labs	Progress	Scientist and queue
Pause Project toggle (`FCT_ResearchProject.Pause`)	Project stays and its assigned labs stay reserved	Decimal remainder unchanged. The research step skips its spending	Scientist and follow-on queue remain
Cancel Project or another assignment clear	Project removed	A separate race-level saved-progress entry (`FCT_PausedResearch`) may be created	Scientist freed by the assignment clear. Every queued entry attached to that project removed

Cancellation stores DevelopCost - truncate(remaining RP), and only when the result is positive. Truncation toward zero can credit part of a point on a later restart. A cost of 100 with 73.8 remaining stores 27, so a restart has 73 remaining. This is not a preserved decimal RP balance. Pause in place does not perform that conversion. The Cancel Project confirmation warns that it also cancels “any queued projects for this scientist”.

Assign New means that the project takes newly free labs automatically, up to the scientist’s whole-number Research Admin. It is not an automatic choice of new technologies. It survives normal queue succession. Paused projects take part in lab allocation and shortage adjustment before the research step skips their RP generation.

Queues live in one race-wide list, but each entry points to a colony and a particular current project. The Add to Queue button adds an already available technology behind a selected active project. A new entry gets order 1000, then that project’s entries are renumbered 1, 2, 3 and so on. Up Queue and Down Queue first space the orders by 10, offset the chosen entry by -15 or +15, then renumber. They do not reorder other colonies’ projects or assign another scientist. Remove Queue removes the entry and renumbers its siblings. There is no separate independent global first-in first-out queue.

Cycle execution completion and overflow

At each research step the game lists the colony’s projects by assigned labs, largest first, then totals the assigned labs and the available whole-number labs. If capacity exceeds assignments, Assign New projects take the free labs in descending order of their current labs, each up to its scientist’s Research Admin. Leftover capacity logs an inactive research facilities event. Zero-lab and paused projects are not excluded from these allocation lists.

If assignments exceed capacity, the game logs an overallocation event and reduces projects in descending lab order. It first selects only projects without Assign New. Only if that list is empty does it select all projects. It does not fall back to Assign New projects after exhausting a nonempty list of other projects. A project with fewer labs than the remaining shortage is reduced to zero. Otherwise it loses the shortage and the reduction stops. This can leave total assignments above capacity. Ordinary generation does not also clamp labs to the scientist’s Research Admin.

For each unpaused project at the colony, generated RP strictly below the remaining RP is a partial cycle. The game records the RP generated and subtracts it. Otherwise the project completes, including at exact equality. The main project is removed, its scientist’s assignment is cleared, the technology is completed, and saved-progress and queued entries for the completed technology are removed. Main completion records the actual remaining RP and one completed project as scientist measurements. An overshoot sets overflow RP = generated - remaining and overflow fraction = overflow RP / generated. Equality leaves both zero.

Follow-on queue entries for this colony and current project are taken in ascending order. The same scientist and assigned labs continue. Several rules differ from a simple conserved RP bucket:

For each queued entry the game recomputes the specialty and local construct factors for the new field, but multiplies by the race bonus of the original completed project’s field.
It calculates new generated RP = B x original labs x new multiplier x y x overflow fraction.
The full-completion test is queued DevelopCost at or below the previous overflow RP. It does not use the new generated RP, and it does not use the queued cost after saved progress.
If the test passes, the game completes the queued technology directly, sets overflow RP = new generated RP - queued DevelopCost, and multiplies the overflow fraction by overflow RP / new generated RP.
Otherwise it starts a project for the queued technology with the new generated RP as carry-over, copies Assign New, removes that queue entry, moves the remaining follow-on entries to the new project, renumbers them, and stops.

These formulas are not clamped. A change of specialty can therefore make the direct-completion threshold disagree with the recomputed RP, or start a new project with a negative remainder. The follow-on path calculates a wealth amount but never records it. Directly completed queued entries do not receive the RP-generated and completed-project measurements that the main project records. Direct queued completion normally removes saved progress through the completion cleanup described in the next section. Starting the partial follow-on project consumes it instead.

At exact main completion, a positive-cost queue entry starts with zero carry-over. When no queue exists, the scientist becomes free and unused RP is not stored for a later manual assignment. No separate minimum spending rule or general overflow conservation rule applies.

Completion cascades and external research

Acquiring a technology normally runs through one common completion step. Creating an independent race and loading a game add knowledge entries directly. When no research prototype is being converted at a colony, completing an already-known technology does nothing. A new acquisition adds a knowledge entry that is not obsolete. With cascades enabled (every acquisition except the starting technologies a race is given at creation and the key technologies granted for automated designs), active projects for this exact technology are removed and their scientists’ assignments cleared, matching queued and saved entries are removed, and an event is logged. A direct acquisition therefore differs from a normal project that finishes and moves along its queue.

The completion step applies supported technology effects to the race. Examples include research rate, mining, construction, ordnance, fighter and fuel output, shipbuilding, terraforming, maintenance, transfer rates, training-related capability and genetic conversion. Improved Research Rate truncates AdditionalInfo to a whole number. Completion does not immediately recalculate every colony’s per-lab RP. Generic technology can also make shipping-line designs obsolete. Other rules read known technology directly, so acquisition is more than an event-log entry.

For generic technology (RaceID = 0) with cascades enabled, races with a technology treaty receive the same technology through the same completion step. If the technology type allows it (DIM_TechType.DistributeLowerTech = 1), other Allied races receive their cheapest unknown non-automatic technology of that type, only when its cost is strictly below the completed cost. Trade-treaty partners below Allied, and non-hostile contacted races with neither treaty whose capital sits on the same body as this race’s capital, use strictly below completed cost / 3. That division is whole-number division because development costs are whole numbers. These selections do not apply the available-list prerequisite query. Known-technology checks stop ordinary back-and-forth propagation between races.

Finally, every automatic technology whose Prerequisite1 equals the newly completed technology ID is completed in turn through the same step. This last pass does not test the second prerequisite, race ownership or RuinOnly, and it runs even for those starting-technology and automated-design grants. Automatic completion is therefore a first-prerequisite cascade, not the same operation as checking ordinary project eligibility.

Salvage, component disassembly, conquest, intelligence breakthroughs and treaties are further sources of knowledge or RP. Intelligence mechanics covers their selection filters and random rewards, including captured race-specific eligibility. Salvage and archaeology covers recovered technical data. Production mechanics covers consuming physical components for disassembly. Making a foreign component eligible for study does not by itself create knowledge. It only adds a race-specific technology the race does not know to the race’s list of technologies eligible for study.

External whole-number RP, such as downloaded technical data, first go to the cheapest active project of the same technology type, not necessarily the exact technology supplied. If the award is below that project’s remaining RP, it is simply subtracted. Otherwise the game completes that active technology and runs the cleanup for the supplied technology, and any excess award is discarded. Without an active project, the game selects the cheapest saved-progress entry of the same type. Command and Control technology (DIM_TechType ID 87) requires the exact technology instead. The comparison uses the supplied technology’s development cost minus that entry’s points. Below that, the award adds to the saved entry. Otherwise the supplied technology completes and the cleanup runs. If no saved entry exists, the game stores a new saved entry for the supplied technology holding the award, without checking whether the award already covers the cost. These rules are distinct from the time-based queue overflow of the research step.

The cleanup that also removes active projects has its own succession path. It removes matching active projects and tries to start the first queued replacement for each, without checking that a replacement exists. When none exists, the game shows error Function #2449 and the cleanup ends there. It does not apply the research step’s full queue retarget and overflow routine. External awards therefore do not share the succession behaviour of every other acquisition.

NPR research is separate. The production step sends player colonies to the project-based research step, but sends AI-enabled NPR colonies to design-philosophy research. That path adds the colony’s output to the race research total (FCT_Race.CurrentResearchTotal). The amount is per-lab output x labs x multiplier x elapsed years, where multiplier = max(best specialist's 4s - 3, best scientist's nonspecialist s), with the specialist taken in the next progression’s field. It can complete one next design-philosophy progression per colony step when the total reaches the target (FCT_Race.ResearchTargetCost), and it keeps the remainder by subtracting the target. It does not run the player’s scientists, projects or queues, and it does not charge the per-project wealth. The conventional NPR population AI runs separately, chosen by starting level. Player conventional technologies still follow the ordinary known and available distinctions above.

Scientific personnel

The shared commander record, creation, initial bonus rolls, personality traits, rank themes and the assignment model for all four commander types are on the commander lifecycle page. The subsections below keep the scientist-specific rules.

Recruitment and academy commandants

Military academies produce a stream of generic personnel, not ten scientists each year. The academy step adds academy count x 10 x elapsed years to the colony’s decimal stock of academy officers (FCT_Population.AcademyOfficers). While the stock is at least one, it subtracts one and creates a commander, with the colony’s current academy commandant as instructor. Fractions carry over to the next cycle. No efficiency or workforce multiplier and no random graduation check apply. The academy step runs before the positive-population check that gates research in the production step. Crew recruitment is a separate stock (FCT_Race.AcademyCrewmen) that grows by academy count x floor(1000 / race training level) x elapsed years.

Creating a commander allocates an ID and picks a profession. If an academy commandant is present, a first d100 copies the commandant’s profession on 8 or less for an administrator, 50 or less for a ground force commander, 80 or less for a naval officer, or 8 or less for a scientist (10 or less under Limited Research Administration, the Game Information window option stored as FCT_Game.LimitedLabs = 1). If that roll fails, or there is no commandant, the game rolls r = d100 + limited penalty, where the penalty is 1 under Limited Research Administration and 0 otherwise. An r below 65 gives a naval officer, below 93 a ground force commander, below 97 an administrator, and anything higher a scientist. Without a commandant the scientist chance is therefore 4% normally and 5% under the limited option. A scientist commandant gives 0.08 + 0.92 x 0.04 = 11.68% normally, or 0.10 + 0.90 x 0.05 = 14.5% under the limited option. Other commandant types use up part of the fallback chance with their own inherited profession.

Every recruit gets a research specialty, including nonscientists. With an instructor, a 25% chance copies the instructor’s field. Otherwise a d9 picks field ID 1 to 9 of DIM_ResearchField. Without an instructor the d9 is used directly. The range is fixed at IDs 1 to 9 and does not follow the size of the research field catalogue. A scientist instructor therefore passes on their own field with probability 0.25 + 0.75 / 9 = 1/3. Profession inheritance, specialty inheritance and bonus influence are separate rolls.

Recruits receive the graduating colony’s species, education colony, homeworld and location. Career start = current game time - d5 years. The assignment and promotion times start at the same value. Even a new academy graduate therefore starts with 1 to 5 years of service. The first rank is the most junior rank of the profession’s ladder where one exists, and a scientist may have none. A promotion history entry is added only when a rank is set. Loyalty is d30 + 70. The academy’s name theme names the recruit when one is set, otherwise the race naming theme supplies the name. The new commander joins the active commanders after bonus, health and trait setup. New-game setup creates max(academy count x 10 x 20, 300) generic commanders at the homeworld with no instructor, not that many scientists. Independence also creates new generic personnel, floor(old race's active commander count x population share) of them, rather than moving existing scientists.

The academy list in the Commanders window offers a scientist an academy only when Research Admin / threshold is at least that colony’s academy count. The threshold is 5 normally or 1 under Limited Research Administration. A scientist with Research Admin 10 can command two academies in a normal game. Research Admin 3 can command three in a limited game. The list shows owned colonies with academies, and only those without a commandant when Available Only is checked. Assigning clears the scientist’s old assignment, which cancels any research, relieves any incumbent, sets the academy command, location and assignment time, and adds an assignment history entry. A commandant does not multiply laboratory output. The commandant’s influence is on recruits. Eligibility is checked only by the list. The assignment itself trusts the selection.

Initial research and administration bonuses

Scientist creation rolls d100 to set an absolute administration capacity, the Research Admin bonus:

d100	Normal max labs	Limited administration max labs
1 to 20	5	2
21 to 50	10	3
51 to 70	15	4
71 to 85	20	5
86 to 95	25	6
96 to 100	30	7

Limited Research Administration turns the normal value into normal / 5 + 1. The instructor’s administration does not improve this roll. Research is rolled separately. A d100 of 1 to 15 creates no bonus entry, 16 to 30 gives 1.05, 31 to 60 gives 1.10, 61 to 75 gives 1.15, 76 to 90 gives 1.20, 91 to 95 gives 1.25 and 96 to 100 gives 1.30. If the instructor’s Research multiplier is at least 1.20, the roll is the higher of two independent d100 rolls, not a copy of the instructor’s bonus.

A missing bonus reads as 1 for bonuses marked Percentage in DIM_CommanderBonusType and as 0 for absolute bonuses. The displayed research percentage is (multiplier - 1) x 100. Maximum bonuses come from DIM_CommanderBonusType.MaximumBonus. The game database as shipped gives Research (ID 3) a maximum of 1.50 (50%) and Research Admin (ID 27) a maximum of 50. Limited Research Administration changes recruitment values and gain increments. It does not change that cap in the experience gain.

Experience and changing specialty

Research skill grows through random experience over elapsed time, not through RP earned or projects completed. The experience check covers every active commander . One d10000 success check passes on elapsed years x 500 or less for unassigned personnel and for leaders of research projects with zero labs, and on elapsed years x 2500 or less otherwise. A paused project with labs still gets the assigned threshold because pause is not checked. One lab and fifty labs give the same chance. An academy commandant also gets the assigned threshold. Dice give whole numbers, so a fractional limit such as 12.5 acts as 12. Nothing caps the threshold except the size of the die, but each check gives only one gain opportunity.

On success, a d3 of 1 sets the gain to 0.10, otherwise 0.05. A scientist then rolls another d3. A result below 3 raises Research by that amount. Otherwise Research Admin rises by d10 normally or by d2 under Limited Research Administration. These are additive research increases of 10 or 5 percentage points, not relative growth. A missing bonus entry is created at 1 for a percentage bonus or 0 for an absolute one, the gain is added, and the result is clamped to the catalogue maximum. A bonus already exactly at its maximum is left unchanged. A gain aimed at a capped bonus is not redirected. With the Commander Political Bonuses option of the Game Information window (FCT_Game.PoliticalAdmirals = 1), an independent d10000 at or below elapsed years x 1000 first raises Political Reliability by 0.05, before the ordinary experience check. No species, training, academy or augmentation multiplier appears in the scientist experience formula.

For example, at 0.02 years an unassigned scientist succeeds on a d10000 of 10 or less (0.1%), while a scientist leading a project with labs succeeds on 50 or less (0.5%), even when the project is paused. After success, Research receives the gain on two of three outcomes. A 1.48 multiplier receiving +0.05 clamps to 1.50. A cycle in which labs produce RP records RP and completion measurements separately. Those measurements add to the scientist’s statistics and are checked for medals, and NPR races keep none. They never raise Research or Research Admin. New graduates appear after the cycle’s experience pass, so they miss that cycle’s opportunity.

The Commanders window shows a Change Field button for a selected scientist, and it needs no Spacemaster mode. After the confirmation, which warns that the research bonus “will be reduced by 75%”, it offers the displayed fields other than the current one (DIM_ResearchField rows with DoNotDisplay false). It sets the chosen specialty. If a Research bonus entry exists, it replaces it with (old multiplier - 1) x 0.25 + 1, rounded to two decimals with halves rounded away from zero. 20% becomes 5%, 30% becomes 8%, and 50% becomes 13%. A missing Research bonus stays missing. Research Admin, career start, health, current assignment and RP are unchanged. The specialty multiplier of a running project can change at once.

Age health retirement and assignment loss

Service years = (current game time - career start) / one year, rounded down. Age is service years plus the species GraduationAge (default 21). No separate birth date enters this calculation. Initial health risk uses a d100. A roll of 80 or less gives 1, 81 to 95 gives 2, 96 to 99 gives 3, and 100 gives 5. The species GraduationAge affects derived age. The species research rate modifier belongs to colony output, not to scientist experience or administration.

The health and retirement check covers active commanders who are not story characters (FCT_Commander.StoryCharacter), in this order:

Accident. A d1000000 below 5000 x elapsed years kills the commander in an accident, and the check moves on to the next person. At exactly one year the strict comparison gives 4999 successful die values, not 5000.
Health. Let h = health risk + max(age - 60, 0) / 2.0. The test is a d100000 below h x elapsed years x 1000. On success, d10 + previous health risk gives a risk increase of 1 at 6 or less, 2 at 7 to 8, 3 at 9 to 10, medical retirement at 11 to 12, and natural death above 12. A terminal outcome moves on to the next person. A nonterminal worsening can be followed by a career check.
Career. For scientists the minimum service is 39 years. With service years of 39 or more, the test is a d10000 below (service years - 39) x 20 x elapsed years x 100, and the 20 doubles only for a commander with no assignment. At 39 the threshold is zero, so the first possible career retirement comes at 40 service years (age 61 by default). At 40 years over one year, an assigned scientist retires on 1999 of 10000 die values and an unassigned one on 3999 of 10000. Nothing else caps the probability, and no repeated catch-up roll applies.

Paused and zero-lab projects count as assignments for career retirement. The zero-lab rule of the experience check does not apply here. Story scientists skip these health, accident and career checks but still gain experience. The health label is a separate coarse display. It adds truncate((age - 60) / 2) only above age 61, while the check uses the fractional half-age term. At age 61 the check adds 0.5 and the label adds zero. A label total of 1 shows Excellent, 2 Good, below 5 Fair, below 8 Poor, and anything higher Very Poor.

Retirement, whether by death, medical or career retirement or the Retire button, clears the assignment, moves the commander to the retired list and removes them from the active list. Clearing a research assignment cancels the project and stores its progress as described above, even when the project is paused, then clears the project link, location and transport ship. The whole attached queue disappears. No replacement scientist is assigned automatically. Saved progress from earlier cancellations belongs to the race and technology, so unrelated saved entries survive. Education colony, species, career start and bonuses stay on the retired commander.

Scientists do not take part in ordinary naval or ground promotion, and the Promote and Demote buttons reject scientists and administrators (“Only naval or ground commanders can be promoted”). Promotions made with the time requirement waived, such as those at game setup and independence, move a promoted naval or ground commander’s career start back by 2 + d5 years, as described in object identity and reload. That change is therefore not ordinary scientist ageing. SM: Add Extra creates generic recruits at the capital. SM: Replace All clears the selected race’s active personnel, cancelling assigned research and queues, and creates new generic recruits. These are personnel administration tools, not a species augmentation or rejuvenation multiplier. See Spacemaster interventions.

The Spacemaster Restore button resets health risk to 1, sets the status to active, removes the retired entry and updates history, display and seniority. It does not restore cancelled assignments, reset career start or reroll bonuses. Restore does not put the commander back in the active commander list. The status change alone therefore does not return the scientist to the available-scientist list, the experience check or the save, which read that list.

Personnel persistence

A save stores the active commanders, plus retired commanders only when RetainRetired or Prisoner is set in FCT_Commander. It stores specialty (ResSpecID), assignment type and target (CommandType and CommandID), species, health risk, career start, assignment and promotion times, the story character flag, and separate bonus, measurement and history tables. Loading builds the commanders without rerunning recruitment rolls. It restores career, health and specialty, links each project ID to both the scientist and the project, and places each commander in the active or retired list by saved status. A missing project clears the assignment type. After commanders load, projects with no commander are cancelled and their progress stored. Loading bonuses from FCT_CommanderBonuses rounds saved values to three decimals, where an exact half goes to the even neighbour, and only for known bonus IDs. Experience gains during play have no such rounding step.

Timing previews and persistence

Research advances in the strategic construction cycle. The elapsed years are the cycle’s elapsed seconds / one year, converted through floating point at the cycle level and back to decimal for population production. The cycle first refreshes colony environment and workforce, then commander health and retirement, then bonus experience, then the research field bonuses and economy of standard races, then population production. Assignment loss or scientist improvement can therefore affect the research step in the same cycle. Population production recalculates per-lab output before research, and requires total population above 0 and a whole-number research capacity above 0 before research runs. Research does not advance on every movement sub-pulse or when the Research tab opens. Game time and state refresh covers the cadence and the cycle boundaries.

Opening colony details can recalculate workforce, efficiency and research output values without advancing project RP. Refreshing the research project list rebuilds race field bonuses, project multipliers, annual RP and estimated completion times. These are display values, not saved RP changes. The refresh also stores the available research facility count that Create Project uses. With the Empire view selected, it sums the labs of every displayed project against the selected colony’s capacity, and calculates annual RP with the selected colony’s per-lab output even for projects at other colonies. Each other colony’s own construct is used for its rows, but the per-lab base can still be wrong for them.

An active row estimates remaining RP / annual RP x one year. Zero annual RP substitutes 1000 years. Pause only shows “Yes” in the Pause column. The estimate does not become infinite and does not drop the paused project’s labs. Queued rows subtract matching saved progress, recompute the new field’s scientist, construct and race bonuses, and show their own remaining RP / annual RP duration in years, not a cumulative date after every predecessor. The list uses the queued field’s race bonus, while the overflow in the research step uses the old field’s race bonus. The refresh removes queued entries whose current project has no commander.

Stored state	What the game stores and restores
Active project	`FCT_ResearchProject` holds ID, technology, race, colony, labs, research field, decimal remaining RP, `Pause` and `AssignNew`. Loading links the existing race, colony, technology and field. A row with a missing required object is skipped.
Scientist assignment	The commander save writes the project ID as `CommandID` with the research project assignment type. Commander loading resolves that ID after projects exist and sets both links. A missing ID clears the assignment type to none.
Queue	`FCT_ResearchQueue` holds colony, technology, current project ID and order. Loading attaches an entry only when the referenced project exists at that colony.
Cancelled progress	`FCT_PausedResearch` holds technology, race and whole-number accumulated points. Loading restores race-level entries for valid technologies.

Loading creates active, queued and saved research before commanders rebuild their project assignments. New project IDs continue from the highest saved FCT_ResearchProject.ProjectID. Cached annual RP, multiplier, completion estimate, available facilities and pause text are recalculated rather than saved with the project. Reload therefore keeps decimal active progress while rebuilding display values and links. Object identity and reload covers broader ID reuse and old-snapshot caveats. A displayed technology cost, completion estimate, unassigned-lab count or NPR research total is not interchangeable with a player’s saved project RP balance.

Worked examples

Specialty and wealth. Let the race research rate be 200 RP per lab per year, all colony factors 1, the species modifier 1.25, Research Speed 150%, ten labs and a 20% scientist. Base output is B = 375. A matching specialty gives multiplier 1.8, so annual output is 6750 RP, and 0.1 years produce 675 RP. Wealth before completion is 375 x 10 x 0.1 x (100 / 150) x (1 / 1.25) = 200, apart from floating-point rounding. A main project needing only 270 RP finishes and is charged 80 wealth. A nonmatching specialty would generate 450 RP for the same nominal full-cycle wealth.

Shortage and auto-assignment. Suppose capacity falls to five labs. Project A has eight labs and Assign New on. B has two labs and Assign New off. The assigned total is ten, a shortage of five. The reduction list contains only B, which is reduced to zero and leaves a shortage of three. A remains at eight. No second pass reduces A. If capacity instead rises to twelve and A’s scientist has Research Admin 10, its eight labs can grow by two, and any remaining free labs go to another Assign New project or to an inactive facilities event. Paused status does not exclude either project from these lists.

Overflow with unchanged multiplier. A main project generates 100 RP and needs 60. Main wealth is charged for 60% of the cycle. Overflow is 40 and the fraction 0.4. With the same multiplier, a queued 25-RP technology completes directly and leaves 15 RP and a fraction of 0.15. A following 50-RP technology starts with 15 carry-over and needs 35. The research step records no additional wealth for that follow-on spending. Only the main completion recorded an expense.

Overflow across specialties. If the first queued technology instead has half the previous effective multiplier, the recomputed generation for it is 20 RP, but the full-completion test still compares its cost against 40. A cost-30 entry is therefore completed directly, and the game computes overflow RP = 20 - 30 = -10, then overflow fraction = 0.4 x (-10 / 20) = -0.2. The example shows the inconsistency as the game applies it. It is not a proposed intended rule.

Assignment loss and resumption. A cost-1000 project has 600.4 RP remaining and three queued technologies. Cancel Project, retirement or any other assignment clear cancels it. Cancellation stores 1000 - 600 = 400 accumulated RP (the 600.4 remaining is truncated to 600), removes the active project and removes all three queued entries. Starting it later with a different scientist consumes that saved entry and starts at 600 remaining, with a new project ID. Toggling Pause Project instead would keep 600.4, the scientist and the queue, and keep the assigned labs reserved.

Concept map

Technology knowledge and availability defines technologies, knowledge entries, obsolescence and the available and completed lists.
Laboratories output and wealth turns lab capacity, colony factors and scientist bonuses into RP and wealth cost.
Assignments pause cancellation and queues covers project start, pause, cancellation, saved progress and queue order.
Cycle execution completion and overflow covers lab allocation, shortage cuts, completion and queue overflow.
Completion cascades and external research covers the common acquisition step, treaty sharing, automatic technologies and external RP awards.
Recruitment and academy commandants explains how academies and their commandants produce scientists.
Initial research and administration bonuses gives the starting Research and Research Admin rolls.
Experience and changing specialty covers skill growth and the Change Field penalty.
Age health retirement and assignment loss covers how scientists leave and what happens to their projects.
Personnel persistence lists what a save keeps for scientists.
Timing previews and persistence covers research cadence, displayed estimates and saved research tables.
Commander lifecycle holds the shared commander record and rules for all four professions.
Commander experience and bonus growth applies the same gain check to the other professions.
Component design owns blueprint formulas, prototypes and development costs.
Ship and station design uses researched components to build classes.
Technology to deployment follows knowledge from research to deployed ships.
Production mechanics owns manufacturing and component disassembly.
Species engineering follows its own research and deployment path.
NPR race generation and automated design holds the design-philosophy progression NPR research completes.
Intelligence mechanics owns captured knowledge, treaties and recovered-data rewards.
Salvage and archaeology supplies technical data downloads and .
Wealth, income and treasury receives the research wealth expense.
Colony and population lifecycle supplies efficiency, radiation, stability and political status.
Colony output and income cycle places research among the other colony outputs.
Game time and state refresh sets the construction cycle that advances research.
Game rule options and effects holds Research Speed, Limited Research Administration and Commander Political Bonuses.
Medals and achievement conditions uses the research measurements scientists accumulate.
History records stores scientist assignment and retirement history.
Civilian governance and sectors covers the governor bonuses that research output does not use.
Naval administration and command effectiveness covers the promotion rules scientists do not enter.
Ground force organisation covers the other rank ladder scientists do not climb.
Ship readiness and recovery owns deployed crew and officer readiness.
Spacemaster interventions lists the personnel tools Restore, Add Extra and Replace All.
Object identity and reload covers project ID reuse and reload caveats.

Resources

Related community documentation and developer posts. These may describe a different game version.