# External Field Effect (EFE) test
RRT gold standard: same radial "solar system," two runs. Run A = isolated (low global noise $N$). Run B = external field (high $N$). Prediction: internal attraction in B **weaker** than in A (external noise consumes budget → MOND regime sooner).
## Implementation
Entrypoint: `rst_reality_engine_test_suite.run_efe_test`. The test builds one radial graph, runs the same MOND scan twice with different `initial_H`, and compares mean |acceleration| at probes:
```python
# rst_reality_engine_test_suite.py (CLI: --run efe)
def run_efe_test(n_nodes, steps, H_isolated=1.25e-8, H_external=5.0e-8, ...):
nodes_spec, edges_spec, _, deficit_name = build_radial_spec(n_nodes=n_nodes, probe_step=probe_step)
run_a = run_mond_scan(config_path=tmp_cfg, initial_H=H_isolated, ...) # isolated
run_b = run_mond_scan(config_path=tmp_cfg, initial_H=H_external, ...) # external field
mean_abs_a = np.mean(np.abs(run_a["acceleration"]))
mean_abs_b = np.mean(np.abs(run_b["acceleration"]))
efe_detected = mean_abs_b < mean_abs_a # PASS if B weaker than A
# writes efe_report.txt, efe_comparison.png
```
**Run:** From Reality Engine directory:
```bash
python rst_reality_engine_test_suite.py --run efe --micro
```
Or without `--micro` for long run: **200 nodes, 80 steps** (quick: 30 nodes, 60 steps).
**Outputs:** `tests/results/efe_report.txt`, `tests/results/efe_comparison.png`.
**Interpretation:** See report. PASS if mean |acceleration| at probes is lower in Run B than in Run A.