Observed
?view=radar pair rows (e.g. pair 01) show a filled primary Triage with fast scouts next to Run deep pi, beside a green POSITIVE GROSS HINT / large bps figure. Nearby copy says proposal only · no auto spend.
Triage starts a bounded model/heuristic run (docs: radar triage spends the cheap scout pool). The button was not clicked in the playtest.
Why it matters
Distinct from #8 (Refresh live radar wording) and #9 (summary “positive gross hints”). This is the per-pair spend control styled as the thing to click, while “no auto spend” can be read as “clicking is free.”
Same pattern as #2 on Findings (green Explore next while blocked).
Proposed direction (not a spec)
Do not auto-start triage. Demote or reword the spend buttons so they look like an explicit budgeted action, and make “no auto spend” mean the page will not fire without a click — not that the click is free. Do not enable execution.
Out of scope
Observed
?view=radarpair rows (e.g. pair 01) show a filled primary Triage with fast scouts next to Run deep pi, beside a green POSITIVE GROSS HINT / large bps figure. Nearby copy says proposal only · no auto spend.Triage starts a bounded model/heuristic run (docs: radar triage spends the cheap scout pool). The button was not clicked in the playtest.
Why it matters
Distinct from #8 (Refresh live radar wording) and #9 (summary “positive gross hints”). This is the per-pair spend control styled as the thing to click, while “no auto spend” can be read as “clicking is free.”
Same pattern as #2 on Findings (green Explore next while blocked).
Proposed direction (not a spec)
Do not auto-start triage. Demote or reword the spend buttons so they look like an explicit budgeted action, and make “no auto spend” mean the page will not fire without a click — not that the click is free. Do not enable execution.
Out of scope