#39 MULTI-STRATEGY ENSEMBLE: grade 2..8 strategies and blend them. Body = members: [{"strategy_id": 5} | {"strategy_spec": {...}}, ...] plus weights: "equal" | [0.5, 0.3, 0.2] (one positive number per member, normalised for you).

Every member is graded by the SAME run_backtest a single run uses — there is no second grading path — and each keeps its OWN honest verdict. What the ensemble adds is the blended equity curve on a common date axis, the CORRELATION matrix between the members (measured on per-step CHANGES over the shared window, never on the cumulative curves) and the diversification benefit: blended volatility against the weighted mean of the members' own. "ENSEMBLE" is the row label, like "PORTFOLIO" — never a judgement.

Auth: same resolution as POST /v1/backtests. Polled via the SAME GET /v1/backtests/{run_id}; the row's kind is 'ensemble'.