commodity futures.
16 configurations.
live diagnostic grid for S4. each cell encodes one component's verdict on one config — hover for context, click the filter buttons to surface only the rows the agents need to address.
calibration basis — each configuration is backtested on its own $10,000 of starting capital over a 1-year window, net of modeled commissions and slippage. Aggregate figures sum across 16 configurations (16 × $10K = $160K deployed). Past backtested performance does not guarantee future results.
package state · honest-unknown
No published package state for S4. The live repos publish a signed slice per algorithm; S4 has none in the snapshot this page was built from. That is a gap in what has been published — not a statement that S4 is healthy, and not a statement that it is broken.
dating this lane against your clock…
- newest bar
- HONEST-UNKNOWN
- observed age
- HONEST-UNKNOWN
- ceiling
- HONEST-UNKNOWN
- source
- HONEST-UNKNOWN
no live_state_EdgeX.json in the store — this lane has published no live state
snapshot built 2026-10-21T21:32:47.524Z · graded against your clock, not the build’s
dating this lane against your clock…
- newest bar
- HONEST-UNKNOWN
- observed age
- HONEST-UNKNOWN
- ceiling
- HONEST-UNKNOWN
- source
- HONEST-UNKNOWN
no live_state_gTrade.json in the store — this lane has published no live state
snapshot built 2026-10-21T21:32:47.524Z · graded against your clock, not the build’s
dating this lane against your clock…
- newest bar
- HONEST-UNKNOWN
- observed age
- HONEST-UNKNOWN
- ceiling
- HONEST-UNKNOWN
- source
- HONEST-UNKNOWN
no live_state_S5.json in the store — this lane has published no live state
snapshot built 2026-10-21T21:32:47.524Z · graded against your clock, not the build’s
calibration heatmap · pass / fail per (asset × timeframe) cell. tooltips show headline metrics only — calibrated parameter values are subscriber-tier.
- MGC · 5MPASSWATCH169d agoSD-07MFE ratchet firing too early — capturing under 40% of MFE
- MEDIUMS4MGC · 5MSD-07 RatchetPrematureplanned169d ago
MFE ratchet firing on 78% of trades · avg MFE captured 0.31 vs 0.62 baseline · ratchet threshold drift
PF 1.21 → 1.06 (1.9σ)WR 56% → 52% (1.1σ)window 28 trades · 4drecipe SD-07-A · widen ratchet tier thresholds + leg-1 SL multoptuna 14 minopen S4 heatmapticket TKT-S4-MGC-5M-20260505-002
drift watch + repair triage · tier 6 of the AEGIS Adaptive Repair Loop. cells re-evaluated on every trade close; verdict transitions emit drift events; DEGRADED / FAIL cells auto-open repair tickets that route to a narrowed-budget Optuna run. spec · docs/AEGIS_ADAPTIVE_REPAIR_LOOP.md
pipeline diagrams + calibrated algo bundle
trade-flow PDF, HAI architecture PDF, Pine v6 source, per-config calibrated parameters, and the Optuna refinement trace. published per cycle to subscribers + institutional tiers.
access to calibrated parameters, source code, and pipeline diagrams is restricted to subscriber + institutional tiers — this is the engineering moat.