Diagnostics

Autoresearch Ledger

Append-only mutation-evaluation log. Surfaces whether autoresearch is producing accepted strategy improvements, and why mutations are being reverted.

live
Total mutations
—
recent: —
Accepted (kept)
—
—
Reverted
—
—
Last successful mutation
None yet
autoresearch is not yet producing accepted strategy improvements.
Backend verdict —
—
—
source: /api/autoresearch/diagnostics
Is autoresearch improving these metrics?
Trend pills (accepted mutations only)
Computed from accepted-mutation deltas. Reverted mutations never ran in production and contribute zero signal here.
CLV (mean %)
—
no signal yet
All mutations · avg Δ = —
De-duped P&L
—
no signal yet
All mutations · avg Δ = —
De-duped outcomes
—
no signal yet
All mutations · avg Δ = —
Scale-in integrity
—
no signal yet
All mutations · avg Δ = —
Diagnostic

Why mutations revert

Top blocking reasons + warnings. computing…

—
Blocking reasons
loading…
Top warnings
no warnings recorded — all gate decisions are clean revert/keep, no warnings being raised
Recent trend (last 7d vs prior)
computing…
Tier ladder

Football → tier_1a · what's blocking

Failed criteria from /api/tier/promotion-progress. Football is the priority capital path — these gates must pass before real-money trade-enable.

—
loading…
Source: /api/tier/promotion-progress · refreshed every 60s · live readiness checks (can_go_live) run via /api/tier/upgrade only.
Operator focus

What improves next?

Ranked operational levers. Each is currently sub-target; closing each one moves football closer to tier_1a.

loading…
Breakdown
Per-arena mutation activity
Arena Total Accepted Reverted Accept rate Most recent
loading…
Arena
loading arenas…
Decision
Window
Agent
Limit
Showing 0 mutation cycles
newest first
loading mutation cycles…
Append-only ledger at data/mutation_evaluation_ledger.jsonl. This page is read-only — promotion / acceptance decisions live in autoresearch.py + promotion_policy.py.