Residual attribution
§11-§42 — which observed roads produce the TomTom-vs-SUMO gap, and what explains them
Loading corrections
Global residual on the frozen external support
§8-§13 — exact decomposition, reconciled against the oracle
Loading decomposition
Simulated loading — what the model can possibly explain
the precondition every interaction mechanism depends on
Loading loading
Residual concentration
§15-§17 — a few bad roads, or a property of the whole run?
Loading concentration
Mechanism attribution
§23-§26/§39 — what the measured state is consistent with
Loading mechanisms
External residual clusters
§18-§22
Loading clusters
Residual map
RESIDUAL = SUMO − TomTom · positive = SUMO faster · frozen support only
Loading map segments
Ranked segments
sorted by the selector below; opens the segment explorer
Segment contributions
§46 — sort by contribution, not only by the largest speed gap
Loading segments
Next-mechanism decision
§41/§42 — four criteria, reported separately, no weighted score
Loading decision table
Follow-up audits
§28-§38 — what each candidate mechanism was tested against
Loading audits