# Judging evidence map — 7GC Mission Control 1.1.0

The [2026 NASA Space Apps Judging & Awards Guide](https://www.spaceappschallenge.org/resources/judging-awards-guide/) identifies Impact, Creativity, Validity, Relevance and Presentation. This map links claims to implemented behavior; it does not predict a score or award.

| Criterion | Defensible contribution | Inspectable evidence | Boundary |
| --- | --- | --- | --- |
| Impact | A browser mission intended to teach resource interdependence, with friendly language, a dictionary and progressively deeper views. The same mission is usable without dependable connectivity. | [Junior view](./?view=junior), dictionary button, [standalone offline app](offline.html), mobile/offline checks in [test results](test-results.txt). | Intended educational benefit; no classroom learning study, measured reach or accessibility certification. |
| Creativity | Compare scientific work, greenhouse care and reserves; the most science is not automatically the preferred plan. Distinguish a surviving mission from one meeting extra reserve requirements. | [Interactive planner](./?view=explorer#planner), computed nondominated choices and preference selector; [all policy results](planning-evidence.json). | Exhaustive within a bounded phase family; not a new global optimization theorem or a universally optimal plan. |
| Validity | One deterministic state transition drives play, policy evaluation, sensitivity and replay. Resource and queue ledgers expose actual flows, losses, unmet demand and assumptions. | [Model/equations](MODEL.md), [source metadata](data-provenance.json), [41 model/planner tests and browser evidence](test-results.txt), [balanced run](balanced-flight-record.json), [robust run](robust-flight-record.json), source ZIP. | Conceptual NASA inputs combined with labeled teaching hardware, risk and communications parameters; no flight/medical validation. |
| Relevance | A complete lunar-outpost mission addresses the registered Junior Astronaut challenge through power, air, water, food stocks, greenhouse demand, shielding and scientific work. NASA oxygen/water/dry-food rates directly determine resource depletion. | [Playable mission](./), [NASA NTRS reference](https://ntrs.nasa.gov/citations/20230012635), [NASA ECLSS](https://www.nasa.gov/reference/environmental-control-and-life-support-systems-eclss/). | No real lunar location, live weather feed, crop harvest or operational communications network. |
| Presentation | Genuine success/failure walkthroughs, a clear strategy comparison, explanatory traces, seven slides and a timed four-minute pitch. | [Judge walkthrough](demo), [seven-slide PDF](7gc-mission-control-slides.pdf), [editable PPTX](7gc-mission-control-slides.pptx), [pitch and Q&A](PITCH.md). | The deck summarizes verified behavior; a live demo does not substitute for an official judging submission. |

## Reference-condition results

The phase family contains 567 specifications. Of these, 199 satisfy the authored reserve checks and 18 satisfy those checks in all five sensitivity cases. The science/crop-care frontier contains five distinct objective pairs. Different policy specifications may share outcomes.

| Policy | Science h | Crop-care h | Minimum battery kWh | Final battery kWh | Final water kg | All five sensitivity cases meet margins? |
| --- | ---: | ---: | ---: | ---: | ---: | --- |
| Science priority | 12 | 0 | 0.90 | 2.70 | 3.76 | No |
| Balanced frontier choice | 9 | 12 | 0.75 | 1.20 | 3.52 | No |
| Greenhouse priority | 4 | 24 | 0.90 | 1.50 | 3.28 | No |
| Tested-condition robust choice | 2 | 4 | 5.50 | 10.30 | 3.68 | Yes |

All four complete 24 hours in reference conditions, end with 1.53 kg usable oxygen and about 1.04 kg dry food, and accumulate 2 illustrative exposure points. The robust choice retains at least 1.035 kWh in the combined perturbation, with the same decisions and no retuning. These exact results are reproducible from the exported specifications and engine.

The unchanged initial controls fail at hour 9. The reference recovery completes with 14 science hours, but its battery reaches zero at a boundary while generation still serves the next step; it misses the planner’s 0.50 kWh extra margin. This distinction is an intended engineering lesson.

## Observable learning task

After a run, ask the learner to explain one sacrifice: why did pausing science protect life support, why did greenhouse care reduce another opportunity, or how did a local buffer preserve data during an outage? The interface supplies the underlying resource changes and explanations. Whether learners retain or transfer those ideas remains to be tested.

## Verification identity

Baseline model 1.0.0: source commit `7f2bd298ea6a0c99ad6110fb29cd3cf9e169babd`, 14 tests rerun before behavior changes. NASA baseline judging submission confirmed November 15 at 4:09 AM America/Chicago. Candidate model 1.1.0: 41 model/planning tests plus real-browser checks. Final deployment commit and resubmission receipt live in the release handoff outside the immutable source package to avoid self-referential commit hashes.