quick start: use the runnable harness; drop duplicated tier prose

README.md

@@ -45,13 +45,25 @@ ds = load_dataset("modulate/entity-transcription-benchmark", split="test")
 print(ds[0])
 ```
 
-Score with the shipped matcher:
+Or benchmark a system end to end. The harness transcribes every clip through
+your endpoint and scores the result — no transcript file needed up front:
 
 ```bash
-python scripts/matcher_v5.py --selftest        # expect 49/49
-python scripts/score_named_entities.py --hypotheses my_transcripts.csv
+huggingface-cli download modulate/entity-transcription-benchmark \
+  --repo-type dataset --include "scripts/*" --local-dir .
+cd scripts && pip install -r requirements.txt
+
+export MY_STT_KEY=...
+python run_benchmark.py --endpoint https://api.example.com/v1/transcribe \
+  --api-key-env MY_STT_KEY --response-path result.transcript \
+  --system-name my-system --out runs/mine.jsonl
+
+python score_benchmark.py --hyps runs/mine.jsonl
 ```
 
+Full instructions, vendor shortcuts and the scoring protocol are in
+[Reproducing our numbers](#reproducing-our-numbers) below.
+
 **Use the shipped scorer.** Entity accuracy is highly sensitive to
 normalization — casing, punctuation, spoken numbers — and independently written
 scorers do not produce comparable numbers. `matcher_v5` is the definition of
@@ -316,10 +328,9 @@ comparable to ours:
 | WER | computed over every transcribed clip |
 
 `entity_tiers` is an array parallel to `entities`, so a single clip can hold both
-tier A and tier B spans. Tier A spans are discriminative — multi-token names,
-rare surnames, uncommon organizations and places. Tier B spans are trivial for
-any English recognizer and inflate a score without separating one system from
-another.
+tier A and tier B spans; the scorer filters spans, not clips. See
+[Metric and difficulty tiers](#metric-and-difficulty-tiers) for what each tier
+contains and why the headline uses tier A.
 
 `matcher_v5` does considerably more than substring matching: Unicode
 canonicalization, abbreviation aliases, spoken-form reconstruction for numbers,