Benchmarks

Regression Test Suite

The project includes 98 regression tests covering 5 verdict families.

Run the full suite:

python -m pytest tests/ -q

Verdict Family Coverage

Family	Description	Test Count
supported	Claim clearly backed by sources	25
likely_supported	More supporting than conflicting evidence	20
contested	Significant supporting and conflicting evidence	18
likely_false	Conflicting evidence outweighs supporting	17
insufficient_evidence	Not enough information	18
Total		98

Running Benchmarks

# Full regression suite
python -m pytest tests/ -q

# Verbose output
python -m pytest tests/ -v

# Specific test module
python -m pytest tests/test_verify_claim.py -v
python -m pytest tests/test_core.py -v

Calibration Note

Confidence thresholds (support >= 1.35 for 'supported', etc.) are heuristic and have not been calibrated against a gold-standard dataset. Confidence levels (HIGH/MEDIUM/LOW) reflect relative signal strength, not probabilistic accuracy. See docs/trust-model.md for the semantic oracle boundary.