Genesis Numerical Diagnostics & Integrity Dashboard
Math-first observability — why aren't we trading? — Genesis Framework S39
Status: SPEC_ONLY — This page contains panel specifications from Genesis Framework S39. The GNDID will be built when HIRO construction reaches the observability layer. All panels below are specifications, not live dashboards.
GNDID Preflight
Review all formulas, thresholds, risk limits, and numerical parameters extracted from the Genesis Framework before building
Diagnostic Layers
4 diagnostic layers + fine-tuning interface + AI diagnostics — all currently in SPEC_ONLY state
Layer 1
Discovery Funnel (Throughput Math)
"Is the organism generating enough alpha candidates to sustain trading?"
8 metric gauges: events/second, signal-to-noise ratio, hypothesis p-value, analogue match score, validation pass rate, backtest throughput, operators pending deployment. Composite Funnel Pressure score (0.0–1.0). Funnel visualization showing candidate flow from discovery through validation to deployment.
Layer 2
Infrastructure Heartbeat (Sync Math)
"Is the multi-server infrastructure synchronized and healthy?"
4 server health cards (A1 Osaka, A2 Singapore, B Helsinki, C Frankfurt). Per server: CPU%, memory%, disk%, status (READY/DEGRADED/OFFLINE). Latency gauges: Oracle Lag (S3E), Cross-Domain A1↔A2 (target: 70ms p50), Helsinki↔Frankfurt (target: 25ms p50). Packet integrity ratio. Clock skew detection.
S39A — Global Pulse Pane (GPP)
Server Mesh Topological Map
Dark-mode world map (SVG/Canvas) showing 4 server nodes at geographic positions. 6 corridor lines connecting all pairs, colour-coded by percentage-of-baseline RTT: Green (≤105%), Amber (105–125%), Red (>125%), Gray dashed (DOWN). Each node: pulsing status dot, 60s CPU sparkline, memory bar, operator count.
| Corridor | Baseline RTT | Role |
|---|---|---|
| A1 ↔ A2 (Osaka ↔ Singapore) | 70ms | Cross-venue CEX arb, signal relay to DeFi |
| A1 ↔ B (Osaka ↔ Helsinki) | 230ms | Signal: research to Binance execution |
| A2 ↔ B (Singapore ↔ Helsinki) | 180ms | Signal: research to Bybit/DeFi execution |
| B ↔ C (Helsinki ↔ Frankfurt) | 25ms | European DeFi MEV relay |
| A1 ↔ C (Osaka ↔ Frankfurt) | 255ms | Fallback path |
| A2 ↔ C (Singapore ↔ Frankfurt) | 160ms | Singapore to EU DeFi |
Telemetry: RTT (1s), Jitter σ (1s, 60-sample), Packet Loss % (1s, 60-sample), Clock Skew (5s). Auto-stall: RED >30s, DOWN >10s, loss >1% >60s, skew >5ms. Jitter ≥8ms force-stalls transient operators.
Corridor Stall: click corridor → Stall button (pauses ONLY operators on that route). Auto-releases when condition clears unless Director locks.
SPEC_ONLY Refresh: WebSocket 1s Source: corridor_telemetry
Layer 3
Execution & Risk Gates (The "Kills")
"How many trades are being rejected and why?"
8 gate status rows: Slippage (S31), Liquidity (S30), Correlation (S32), Guardian (S33), Capacity (S28), Regime (S04), Decay (S10), Gas Budget (S31C). Each row: gate name, source section, current value, threshold, status (OPEN/CLOSED). Kill Rate percentage. Kill Feed: last 100 rejections, filterable by gate, showing metric vs threshold with margin of failure.
Layer 4
Portfolio & Attribution
"What is the organism actually producing?"
Real-time metrics: NAV (large number with trend), Daily P&L (with sparkline), RAR gauge (risk-adjusted return), Active Operators (count with status breakdown), Extraction Density (alpha per unit capital), Operator Diversity (C-Space spread metric from S06), Attribution bar chart (alpha / execution / beta / residual split).
Controls
Fine-Tuning Interface
"What can the operator adjust?"
5 parameter sliders: α Aggression (0.5–2.0), β Friction Tolerance (0.5–2.0), γ Decay Sensitivity (0.3–3.0), δ SNR Floor (1.0–3.0), ε Correlation Cap (0.3–0.9). Each shows current value, default, and logs all changes. Emergency Stop: large red button, requires confirmation. Suspends all new order submission (does NOT close positions — that is Guardian’s job via S33).
AI
AI Diagnostics
"What does the AI think is wrong?"
Latest Claude + GPT-4o diagnostic analysis displayed side by side. Trigger info (manual or stall-detected). Structured recommendations with "Apply Recommendation" button. History of previous analyses. Budget: $2.40/day (max 6 analyses at ~$0.40 each). Separate from verification budget.
Full specification: Architecture, stall detection, cross-section integration, refresh rates, failure modes, and SBS in Genesis Framework S39 →