sots-engine/tests/game_sim
alex d3ee45364b game/sim + app: ComputeOutput on the turn path, and P01/P02 committed
`ComputeBudget` takes a system's money from two different functions. Projected mode
calls `ComputeMaxIncome`, which lane E1 closed 25/25 against the BnkEl oracle. The
TURN calls `ComputeOutput` with the system's own rate sliders, where the build queue,
the ship-repair pass and the infrastructure -> terraform -> money cascade are all
live and E1's proof that the cascades are zero does not apply.

Read from the instruction stream, both ranges disassembled to the next function start:

* `sim::ComputeSystemOutput` -- the channel algebra of `ComputeOutputFromRates`, with
  every rounding site (round-half-even per channel, truncating for the construction
  and money slots) and the association of every x87 sum as the original has them.
* `sim::IdealSuitability` -- the owner's own field, the server's species baseline for
  an independent colony, and the per-system `dsu` override.
* `sim::RepairShipsInOrbit` -- the round robin, which is provably equivalent to
  `points - min(points, demand)`: the per-pass share is at least 1, so the only early
  exit needs every remaining cost to be zero.
* two corrections to `ConstructionPoints` and `SplitLeftover`: the station bonus is
  ignored unless strictly positive and its association is `k x (b x cons) + cons`, and
  the leftover weights sum as `wi + (wf + wt)`.

The load-bearing fact: the leftover construction points come back to the TRADE
channel, so a colony with an empty build queue earns the same money whichever way its
sliders point. The engine now runs both paths on every load and reports the
difference; on the 11-save corpus every delta decomposes to the unit into the build
queue's points priced through the money chain.

P01/P02 move from blocked to partial and are committed:

    turn1-state -> turn2-state    209 -> 157   closed 52  regressed 0   (was 51 / 0)
    turn2-state -> turn3-state    108 ->  86   closed 22  regressed 0   (was 21 / 0)

One leaf per pair, and it is the easy one: the independent colony, whose population
does not grow and whose orders the turn does not change. The human's savings are
still short by the civilian growth `S11` does not commit, and the AI's by its own
orders. The ship-repair demand is taken as 0 because `Ship::RepairCost` is unread.

sots-re: findings/subsystems/output-turn-path.md, ghidra/addresses.d/lane-c3.json
2026-09-08 14:50:18 -04:00
..
build_and_run.sh game/sim: strategic formulas as pure functions (economy, research, colony, movement), 356 checks 2026-09-07 17:28:45 -04:00
check.h b4: colony + movement hooks; 22 formula corrections (growth curve has no capacity term, range margin +0.05f, ties-to-even rounding); 3 verified signatures 2026-09-08 00:57:55 -04:00
CMakeLists.txt lane E3: the per-system visibility record, the explored sweep and the system observed-stamp 2026-09-08 13:38:12 -04:00
mini_json.h game/sim: strategic formulas as pure functions (economy, research, colony, movement), 356 checks 2026-09-07 17:28:45 -04:00
smoke_real_save.cpp game/sim: strategic formulas as pure functions (economy, research, colony, movement), 356 checks 2026-09-07 17:28:45 -04:00
test_colony.cpp game/sim + app: ComputeOutput on the turn path, and P01/P02 committed 2026-09-08 14:50:18 -04:00
test_economy.cpp lane N: the population -> base-output term, live-verified 2026-09-08 12:11:36 -04:00
test_movement.cpp movefleet: reproduce the original's float32 normalise; 8 of 45 live divergences -> 0 2026-09-08 03:50:04 -04:00
test_research.cpp b3: research hook + RNG-snapshot compare design; fix draw mapping to y/(2^32-1), NextInt inclusive, float32 odds, spend floor, decay constants 2026-09-07 23:14:08 -04:00
test_techgraph.cpp U: the SetResearched unlock cascade, and the prediction for the run that checks it 2026-09-08 06:26:00 -04:00
test_visibility.cpp lane E3: the per-system visibility record, the explored sweep and the system observed-stamp 2026-09-08 13:38:12 -04:00