Balance and scale the plan
Open Factory to enter the whole-plan workspace. Its three views use the same enabled blocks, but answer different questions.
| View | Question |
|---|---|
| Overview | What does the entire plan produce, consume, and still need? |
| Connections | Do the intended block-to-block supplies match their consumers? |
| Scenario | What block goals satisfy the factory pins, and what remains external? |
Start with Overview, then use Connections when totals can hide a connection problem and Scenario when planning a scale change.
Find wiring problems with Connections
Factory nets all production and consumption of a good. That can hide a real problem: one block's surplus can numerically cancel another block's shortage even when the intended supply path is wrong. Connections keeps the producer and consumer blocks visible on each edge.
Connections groups goods as:
- Short: linked consumers require more than their producer blocks provide.
- Overproduced: linked producers make more than their internal consumers use.
- Balanced: the producer and consumer rates agree; collapsed by default.
- Unsourced imports: blocks consume the good, but no enabled block supplies it.
- Surplus / outputs: blocks produce the good, but no enabled block consumes it.
On a short good, use scale up to select a producer block, enter a new target, and preview the machines, power, and inputs before applying it. When a good touches many blocks, select the displayed block count to expand the individual connections.
Recompute stale block flows
Use Recompute all blocks when the page reports stale results after a data sync, TURD change, or solver update. This re-solves saved blocks; it does not invent new recipes.
Preview a target with Scenario
Select Scenario in the Factory workspace. It solves the enabled blocks as one speculative system without saving changes.
- Under Factory pins, add the goods that define the factory's desired outputs or consumption. Positive rates request production; negative rates request consumption.
- Edit a pin's rate to preview another target. Stock goals appear as derived stock targets and keep their amount and replenishment window.
- Read Goal changes for each affected good's current and next block goal, whole-factory use, actual block output, and factory surplus. Differences within 1% are treated as balanced so rounding-scale corrections do not keep reappearing. Select a good to open the block that owns its goal.
- Check Raw inputs for the projected demand from outside the planned factory.
- Check Overproduced for goods that would accumulate without another consumer.
- With the saved pin targets, select Balance factory to apply every listed goal change as one undoable action. After editing a final-product target, the same action is labelled Apply scenario.
Only factory pins are fixed. Every other block goal is a value for Scenario to calculate, including additional goals in a multi-goal block. On first use, PyOps proposes current terminal positive goals as initial pins; after you save the list, it uses that explicit list instead. Stock targets are always included.
A produce goal is a minimum. When another recipe in the same block already makes a needed good as a coproduct, Scenario keeps the amount the factory uses as that block's goal without requesting extra dedicated production. Block output shows everything the block will actually export and Surplus shows the amount left after projected use. Recovered supply calls out the portion of a goal covered this way. If factory use is genuinely zero, the next goal can still be 0/s while unavoidable output remains visible as surplus. To process that excess, open the block and select the exported good to add a consuming recipe, or route it to a dedicated consume block.
Scenario measures how each configured goal changes its complete multi-goal block, then builds one factory-wide material model from those local responses. Starting at the pins, it follows required ingredients to configured positive producers. A reached producer's natural byproduct can drive a configured consume goal for that same good. PyOps caps the source at the amount needed for its declared product, so the consumer cannot pull extra source production merely to obtain one of its own outputs. This allows Coal's natural Tar to feed a Tar consumer without enlarging Coal to manufacture Ash or Iron plate.
An unpinned consumption goal with no reached byproduct settles at zero, but it remains a consume goal: its icon still opens consuming recipes and a later factory solve still probes its configured sink. Pin the good with a negative rate when it should consume a fixed boundary amount even without an in-factory source. A valid configured producer currently set to 0/s can likewise be started because its response is probed in its saved produce or consume direction.
If no enabled block can supply a required ingredient, Scenario keeps the solve usable and lists the shortfall under Raw inputs as an external import.
If the factory material model is infeasible, Scenario shows the conflicting material balances before the work list. Each conflict states the shortage or excess, the total rate required and available, and the blocks consuming or supplying that good. A configured producer is called out when its current multi-goal response cannot scale that output any further. Follow the block links to change its goals or recipes, add another producer, or reduce the downstream demand. Nothing is saved from an infeasible preview.
If the factory model solves but the proposed rates fail a full block re-solve, Scenario shows a Scenario validation failed card before the work list. It links each affected block and lists the exact proposed goals and block-solver status. Material-flow mismatches show the expected and actual rates; a solve that does not settle shows the goals still changing between passes. Scenario does not save any of these proposed rates.
Use reset to current to discard the speculative target.
To inspect an unexpected result, enable Capture structured solver traces under Settings → Advanced, reproduce the Scenario calculation, then return there and select Refresh. The trace contains the pins, required-good closure, block-response columns, generated LP model, validation passes, imports, surplus, and final result as JSON.
Supply priority
When several blocks can supply the same good, Scenario uses their supply priority:
- Preferred suppliers are used first.
- Normal suppliers follow.
- Fallback suppliers fill the remaining demand.
Set priority beside the Goal heading in a block. Priority chooses among competing suppliers; it does not scale a block solely to obtain an incidental byproduct.
Scenario is not an optimizer for new designs
Scenario balances the blocks and recipes already in the plan. It can start an idle configured producer, but it does not add recipes or blocks. A required good with no reached configured producer remains a Raw input; add and configure a producer when it should be made inside the factory.