Check whether an Ark pantry loop can run alone
In Savebot: The Last Garden, test the pantry as a closed cycle. The developer describes the Ark as a system where nothing comes in or leaves. Leave the robots to work for a complete cycle and look for any stock that shrinks without being replaced.
Full Steam release unless a source specifies the demo. Evidence checked . Guide updated .
Steps
- Write down the jobs that fill the pantry and what each job consumes. Start from a small loop you can observe, with at least one way to replace each material used by the routine.
- Watch a normal run without adding resources manually. Separate temporary buffers from replacement: a full pantry at the start can hide a missing production step.
- Check what changes over more than one cycle. Fix the first observed shortage or stalled task before expanding the loop or assuming it will hold on its own.
How to tell it worked
Pantry inputs are replenished within the loop and the robots keep working without your direct intervention during the test.
If you are still stuck
The announcement supplies the closed-loop goal, with no universal Ark recipe, thresholds or tested program. Use the first failing input as the next debugging target.
Make an inventory of the consumables involved in one cycle
Choose the smallest pantry routine you can leave running by itself. Identify what its robots consume, what they produce and which visible supply or intermediate item connects those actions. Note the starting amounts or levels the game shows. The Ark is described as a closed system. Starting stock can let the robot complete a series of actions even when the program has no way to replace an input.
Choose two matching points in the routine for comparison. For example, if a robot fills a supply before the next round of work, compare the amount at that same stage of the next round. A reading taken immediately after consumption and compared with another taken after production tells little about long-term balance. Note which stage you are observing so the next value is meaningful.
Follow the first supply that fails to recover
Let one complete cycle run without manually topping up the pantry. If a supply ends a cycle lower than it began, observe one more comparable point when you can. The first item that keeps shrinking gives you the best place to start. Find the robot or action that should replace it and check whether it ran, whether it produced anything and whether the output reached the place where it is needed.
For example, a producer can finish an action while the consuming robot still sees an empty supply. Follow the movement from producer to pantry before editing how the consumer checks for it. If no robot initiates replacement, inspect the decision that should respond as the supply falls. If replacement happens and the amount keeps shrinking, compare what is produced with what the cycle uses using the values visible in your current setup.
Retest under the ordinary operating load
After correcting one observed gap, begin from a readable stock state and watch the same stages again. Let the jobs overlap as they do during normal operation. If you only observed replenishment while the other robots were idle, the next pass with them running can show a different rate of usage. A pantry that runs without you should recover its required supplies and keep the dependent tasks moving.
Keep notes for the exact loop tested: the measurement point, the inputs that recovered, the first one that did not and the changed action. The developer supplies a closed-system objective without a universal recipe or stock threshold. Your useful result is a cycle you can repeat under those observed conditions without relying on an unnoticed temporary reserve.