 table tokens    share   hit@1 shipped   hit@1 no-table          p  outcome
            0     0.00         30 /30           30 /30      1.0000  inconclusive
           50     0.11         30 /30           30 /30      1.0000  inconclusive
          100     0.20         30 /30           30 /30      1.0000  inconclusive
          140     0.26         30 /30           30 /30      1.0000  inconclusive
          160     0.29          0 /30           30 /30      0.0000  no-table better
          200     0.33          0 /30           30 /30      0.0000  no-table better
          400     0.50          0 /30           30 /30      0.0000  no-table better
          700     0.64          0 /30           30 /30      0.0000  no-table better
         1200     0.75          0 /30           30 /30      0.0000  no-table better
         2000     0.83          0 /30           30 /30      0.0000  no-table better

The share at which the shipped arm starts losing is the number to read, not the 30-0 at one size. Compare it against the corpus you care about.

sweep rows (10) -> work/regression/2026-09-12-reaim-and-instruments/evidence/w144-dose-response.jsonl
