Sandbox

## The gap The Compare screen currently promises an 'average IAMS' figure that is not rendered (P0 bug #17, per `docs/SESSION_AGENT_INDEX.md`). Even once that number lands, it tells a sporting…

**Problem:** `docs/camelot_deep_dive.md §3.1` currently assigns the A/B letter purely via `flair = (Technique + Pace) / 2` vs `structure = (Decisions + Intelligence) / 2`. This binary split ignores the **Mental axis**, which `docs/camelot_deep_dive.md §3.0` lists as one of the four scoring axes (Role / Geo / League / Mental) but which `docs/SESSION_AGENT_INDEX.md P0 bug #18` flags as inconsistently labelled ('Mental' on Dashboard vs 'Age' on player profiles) — signalling it is under-engineered. **Proposed change:** Add a third attribute pair, `mental_load = (Decisions + Composure) / 2`, and introduce a **three-band letter rule**: - `flair > structure + 1.5` → **B** - `mental_load > flair AND mental_load > structure` → **A+** (new sub-letter, rendered as a small superscript badge, not a new key) - Otherwise → **A** This keeps the 24-key vocabulary intact (preserving the DJ-wheel metaphor from `docs/camelot_deep_dive.md §2`) while surfacing high-composure structural players (e.g. Regista, Libero) as meaningfully distinct from low-flair defensive players. **Measurable A/B:** Ship behind `?camelot_v2=1`. Compare shortlist conversion rate (shortlist_add events / camelot_wheel_open events) for keys 2A, 9A, and 11A — the three structural anchor slots most affected — over a 30-day window. Hypothesis: A+ badge increases click-through on structural roles by ≥10% vs control.

**Problem:** Clubs can find individual players whose Camelot key is harmonic with the club profile, but there is no surface that shows the *squad as a whole* — whether the 25-man roster is…