Loading Morning Brief...
Top RV Trades
Alerts & Movers
Composite RV Signals
Idea Board
Unified ranked trade ideas. Conviction = |z-signal| × cluster/regime multiplier (from convergence_survivorship).| Kind | Trade | Z signal | Regime mult | Weighted | Rationale |
|---|
Color by:
Country Detail
Loading insights...
Morning Pulse
Select an issuer to view its spread curve
Choose from 777 fitted curves (80 countries) in the sidebar, or search by name/ticker.
Spread Curve
NIP Estimator
Historical Curve Shape
Bond-Level Detail
| Bond | Rating | TTM | OAS | Fitted | Residual |
|---|
Bond History
Computing regime detection, clustering & basis signals...
Running HMM regime detection, GMM clustering, and basis z-scores
Forward-Stress Forecaster
Loading…
Supervised LightGBM forecaster: P(stress in next N days). Replaces HMM's failed forward role.
Top features: GPR z60 lag-20d, SnP500 z60 lag-20d, internal_score. CV AUC: h=5 → 0.65, h=10 → 0.62.
Market Regime Timeline
HMM Regime Probability Bands
Macro Stress Decomposition
Behavioral Clusters
| Ticker | Country | Cluster | Prob | Beta | VIX β | Skew | Spread | Stress β | Vol Asym | Tail Risk | R² | Migr σ | Vol 3M | Vol Ratio | Trend | Drift |
|---|
Sov-Quasi-Corp Basis Signals
| Other Ticker | vs Ref | Country | Sector | Tenor | Ref Spd | Other Spd | Basis | Z-6m | Signal |
|---|
Basis cache not yet built. Run
python analytics/basis_monitor.py first.
Computing multi-dimensional RV metrics...
Multi-Lens Signal Concordance
| Ticker | Country | Sector | Rating | Spread | vs Sov | vs Rating | vs Cluster | Curve Shape | # Agree | Direction |
|---|
Quasi/Corp vs Sovereign Basis
| Quasi/Corp | vs Sov | Country | Sector | Tenor | Sov Spd | Q/C Spd | Basis | Z-Score | Signal | Raw Ref | Raw Q/C | Raw Basis |
|---|
Rating-Tenor Benchmarks (bps)
| Rating | 2Y | 5Y | 10Y | Slope | N |
|---|
Cluster Benchmarks (bps)
| Cluster | 2Y | 5Y | 10Y | Slope | N |
|---|
Spread vs Rating-Tenor Bucket
| Ticker | Country | Sector | Rating | Tenor | Spread | Bucket Avg | Diff | Z-Score | Signal |
|---|
Raw OAS vs Rating-Tenor Bucket (Bond-Level)
| Ticker | Bond | Country | Rating | TTM | Bucket | OAS | Bucket Avg | Diff | Z-Score | Signal |
|---|
Spread vs Behavioral Cluster
| Ticker | Country | Cluster | Tenor | Spread | Cluster Avg | Diff | Z-Score | Signal |
|---|
Raw OAS vs Cluster Bucket (Bond-Level)
| Ticker | Bond | Country | Cluster | TTM | Bucket | OAS | Cluster Avg | Diff | Z-Score | Signal |
|---|
Curve Shape Analysis (Current Slopes, bps)
| Ticker | Country | Sector | 3s5s | 5s10s | 10s30s | 5Y Fly |
|---|
Raw Bond Slopes (Nearest Actual Bonds, bps)
| Ticker | Country | Sector | 3Y OAS | 5Y OAS | 10Y OAS | 30Y OAS | 3s5s | 5s10s | 10s30s | 5Y Fly |
|---|
Computing macro factor sensitivities...
Building weekly spread changes and estimating per-issuer betas
Macro Factor Sensitivity Map
Bubble size = R² | How each issuer's spread responds to Oil and USD movesScenario Impact Tool
| Ticker | Country | Sector | Oil Grp | Cur 5Y | Δ Spread | New 5Y |
|---|
Issuer Macro Betas
| Ticker | Country | Sector | Rating | Oil Grp | Oil β | DXY β | US Real β | MOVE β | VIX β | S&P β | R² | N |
|---|
▶ Macro Sensitivity Heatmap (click to expand)
Computing composite RV scores across all lenses...
Composite RV Leaderboard
| Ticker | Country | Sector | Rating | Score | Direction | Concordance | L1 Curve | L2 Basis | L3 Rating | L4 Cluster | L5 Shape | Liq | CONV | P(Conv) | Carry | Roll | Total |
|---|
Computing cross-country spread convergence pairs...
Cross-Country Dislocations
| Trade | Wide Leg | Spread | Tight Leg | Spread | Z-Score | OU Z | HL | Corr 52w | Conviction | Rating | Region |
|---|
Regional Country Pairs (No Corr Gate)
Same-subregion sovereign pairs (GCC/SSA/MENA/CEE/CIS/ASIA/LATAM). Corr & β shown as quality flags, not gates — surfaces frontier pairs the corr>0.5 filter rejects.| Region | Tenor | Wide Leg | Spread | Tight Leg | Spread | Z 6m | OU Z | HL | Corr | β | Flags |
|---|
Loading novel signal backtests...
Alpha Lab — Novel Signals
Computing bond liquidity scores...
Bond Liquidity Scores
| Ticker | Country | Bond ID | Score | Tier | Size ($M) | DTC | Activity | Min Trade |
|---|
Scanning bond pairs for dislocations...
Computing spread differentials and z-scores across within-issuer and cross-issuer pairs
Within-Issuer Curve Pairs
| Ticker | Country | Short Leg | TTM | Long Leg | TTM | Diff (bps) | Z-Score | Signal |
|---|
Cross-Issuer RV Pairs
| Issuer A | Country | Issuer B | Country | Band | Diff (bps) | Z-Score | Signal |
|---|
Regional Cross-Issuer Pairs
Grouped by subregion (GCC/SSA/MENA/CEE/CIS/ASIA/LATAM) × sector × tenor band; allows ±1 rating-bucket span. Surfaces NGA-CIV, GHA-SEN, KEN-AGO and similar cross-rating frontier pairs the strict cross-issuer table misses.| Subregion | Issuer A | Issuer B | Sector | Rating | Band | Diff (bps) | Z-Score | Signal |
|---|
Pair Detail
Computing corp-sov basis z-scores...
Matching each corp/quasi bond to nearest-maturity sovereign and computing basis history
Basis Wideners (Corp Getting Cheaper vs Sov)
| Ticker | Country | Sector | Rating | TTM | OAS | Sov Match | Sov TTM | Sov OAS | TTM Diff | Basis | Z-Score | Signal |
|---|
Basis Compressors (Corp Getting Richer vs Sov)
| Ticker | Country | Sector | Rating | TTM | OAS | Sov Match | Sov TTM | Sov OAS | TTM Diff | Basis | Z-Score | Signal |
|---|
Trading Through Sovereign (Negative Basis)
| Ticker | Country | Sector | Rating | TTM | OAS | Sov Match | Sov TTM | Sov OAS | TTM Diff | Basis | Z-Score | Signal |
|---|
Bond Detail
Computing convergence analysis...
Loading residuals and running backtest across all windows
Convergence Funnel — cone of outcomes after entry
Every historical dislocation marked-to-market over a fixed forward horizon, aligned at t=0. Bold line = median P&L path; bands = 25/75 and 10/90 percentiles. The median peak ≈ optimal exit day; the lower band staying negative = the trap trades. Fixed horizon (not stop-at-exit) keeps never-converged trades in the sample, so no survivorship bias. Sample count thins at the right edge (events near the panel end).
Signal Quality: Decay Profile & ML Classifier
Window Comparison — All
| Window | Events | CHEAP | CHEAP Hit% | CHEAP Med Δ | CHEAP Max Δ | RICH | RICH Hit% | RICH Med Δ | RICH Max Δ | Avg |Δ| | Max |Δ| | Conv% | Med Days | P25 | P75 |
|---|
Window Comparison — IG
| Window | Events | CHEAP Hit% | RICH Hit% | Conv% | Med Days | Profit% |
|---|
Window Comparison — HY
| Window | Events | CHEAP Hit% | RICH Hit% | Conv% | Med Days | Profit% |
|---|
Cross-Lens Comparison (6m window)
| Lens | Signal | Events | Profit% | Conv% | Med Days | Avg |Δ| | Med |Δ| | Min |Δ| | Max |Δ| | P5 | P95 | Up Max Δ | Down Max Δ |
|---|
Lens 2: Basis Convergence (Quasi/Corp vs Sovereign)
| Window | Events | WIDE | WIDE Hit% | WIDE Med Δ | WIDE Max Δ | TIGHT | TIGHT Hit% | TIGHT Med Δ | TIGHT Max Δ | Avg |Δ| | Max |Δ| | Conv% | Med Days | Profit% |
|---|
Lens 5: 3s5s Slope
| Window | Events | STEEP Hit% | STEEP Med Δ | STEEP Max Δ | INV Hit% | INV Med Δ | INV Max Δ | Avg |Δ| | Max |Δ| | Conv% | Med Days | Profit% |
|---|
Lens 5: 5s10s Slope
| Window | Events | STEEP Hit% | STEEP Med Δ | STEEP Max Δ | INV Hit% | INV Med Δ | INV Max Δ | Avg |Δ| | Max |Δ| | Conv% | Med Days | Profit% |
|---|
Lens 5: 10s30s Slope
| Window | Events | STEEP Hit% | STEEP Med Δ | STEEP Max Δ | INV Hit% | INV Med Δ | INV Max Δ | Avg |Δ| | Max |Δ| | Conv% | Med Days | Profit% |
|---|
Lens 5: 5Y Butterfly
| Window | Events | STEEP Hit% | STEEP Med Δ | STEEP Max Δ | INV Hit% | INV Med Δ | INV Max Δ | Avg |Δ| | Max |Δ| | Conv% | Med Days | Profit% |
|---|
Regional Profit Rate
Regional Speed (Median Days)
Country Convergence
| Country | Region | Events | Bonds | Conv% | Med Days | Profit% | Med |Δ| |
|---|
Issuer Convergence
| Ticker | Country | Sector | Events | Bonds | Conv% | Med Days | Profit% | Med |Δ| | Score |
|---|
Loading contagion network data...
Rolling Centrality (63d window, top 20)
| Code | Country | Region | Degree | Between | Eigen | Composite | Z-Score |
|---|
STATIC PERIOD COMPARISON (below)
Spread Correlation Network
Centrality Comparison: Pre-Conflict vs Conflict
| Country | Region | ME | Degree (Pre) | Degree (Post) | Deg Chg | Betw (Pre) | Betw (Post) | Betw Chg | Eig (Pre) | Eig (Post) | Eig Chg |
|---|
Fitting Bayesian HMM regime model...
HMM State Probabilities (Last 1Y)
Transition Matrix
20-Day Regime Forecast
HMM vs Z-Score Regime Comparison
Scoring forward jump risk…
Monotonic LightGBM, CV AUC 0.95, P@1% lift 64–82×
Top Jump-Risk Bonds (next 5 trading days)
P(any |Δoas| > 4·σ30 AND > 50bps widening within 5 trading days). Ranked. Use percentile rank — the absolute probabilities saturate at the tails. CV walk-forward AUC 0.95 across 4 folds.
| Ticker | Country | Rating | Sector | OAS | Res z60 | Maha | Ctry z60 | Pair z | GPR z | p_jump | Pct rank |
|---|
Feature importance (top 15)
Loading factor model…
Kalman dynamic factor model — systematic vs idiosyncratic spread decomposition
Conviction Map — Fair Value × Idiosyncratic Z
X = hierarchical fair-value cheapness (structural). Y = idiosyncratic z (cheap after stripping the common EM factor). Top-right = cheap on BOTH = highest conviction. Off-diagonal: fair-value-cheap but idio-rich = you're just long EM beta, no RV edge.
Factor Identity — Rates vs Risk Loading
X = factor-1 loading, Y = factor-2 loading (axes labelled by dominant macro driver). Size = R² (how systematic; small = idiosyncratic/un-hedgeable). Color = fair-value cheapness. Pure rates plays vs pure risk-beta plays separate spatially.
Factor interpretation (macro correlations)
Each latent factor is labelled by its correlation with macro drivers. Common variance explained (avg R²) measures how much of the cross-section is systematic; the rest is name-specific (alpha).
Idiosyncratically Cheap (own story, wide)
Cheap after stripping the common EM move. High positive idiosyncratic z = cheap on its own story.
| Country | Idio z |
|---|
Idiosyncratically Rich (own story, tight)
Rich after stripping the common EM move. Large negative idiosyncratic z = rich on its own story.
| Country | Idio z |
|---|
Loading fair-value model…
Rating / sector / duration-adjusted cheapness with partial pooling across countries
Cheapest Bonds (vs fair)
Cheapness = OAS − fair OAS (positive = cheap/wide). Fair OAS is the model's rating/sector/duration-adjusted expectation. CI-low is the conservative cheapness bound.
| Ticker | Ctry | TTM | OAS | Fair | Cheap bp | CI low |
|---|
Richest Bonds (vs fair)
Negative cheapness = rich/tight vs the model's fair OAS.
| Ticker | Ctry | TTM | OAS | Fair | Cheap bp |
|---|
Cheapest Countries (shrunk)
Country random effect = rating/sector/duration-adjusted country cheapness, with partial-pooling shrinkage already applied (thin buckets pulled toward the population).
| Country | Cheap bp | Bonds |
|---|
Richest Countries (shrunk)
Negative random effect = country trades rich vs its rating/sector/duration peers.
| Country | Cheap bp | Bonds |
|---|
Loading relative-debt peer lens…
Spread z vs debt/GDP z within quality×commodity peers — IMF WP/2026/110 "Peer Pressure"
Spread vs relative debt (within peer group)
X = relative debt z (debt/GDP vs peers); Y = relative spread z (OAS vs peers). The dashed line is the empirically-fitted relationship (slope = within-peer beta). Points ABOVE the line trade wide for their relative debt = CHEAP; BELOW = RICH. Low R² ⇒ treat as a corroborating overlay, not a standalone signal.
Cheap (wide vs relative debt)
| Ctry | Rating | Debt% | OAS | DebtZ | SprdZ | Resid | FV | Peers |
|---|
Rich (tight vs relative debt)
| Ctry | Rating | Debt% | OAS | DebtZ | SprdZ | Resid | FV | Peers |
|---|
Loading oil transmission analysis...
Granger Causality: Oil → Country Spreads
| Country | Lag (days) | F-stat | p-value | Significant |
|---|
Impulse Response: Oil Shock
Impulse Response: VIX Shock
Lead-Lag Cross-Correlations (Oil → Spreads)
Loading scenario stress analysis...
Scenario Impact by Issuer
| Issuer | Country | Region | Oil Grp | Current 5Y | Delta (bps) | 90% Lo | 90% Hi | Projected | % Chg |
|---|
Expected Value Analysis (Probability-Weighted)
Bond Similarity — 3D neighborhood
Composite distance: Mahalanobis (40%) + 250d spread-change correlation (30%) + regime-conditional betas (15%) + static (15%).ISSUER
BOND
TOP-N
Each point = one bond plotted at its component-similarity to the query.
The query bond is the gold star at (1, 1, 1). Top-N matches are large + bright; the rest of the universe is the faint cloud.
Hover for ticker, country, sector, ratings, TTM, OAS, and component breakdown.
Top matches
Swap Reviewer
Input a dealer swap (sell → buy). Verdict scored on residual z + OU z. Alternatives drawn from each leg's issuer universe.SELL LEG (dealer takes)
→
BUY LEG (dealer delivers)
Loading IMF Article IV briefs...
IMF Article IV Country Briefs
| Country | Report Date | Fiscal | Debt | Reserves | Reforms | Spread Dir | Headline |
|---|
Country Detail
New Issue Pricer — thin / no-curve issuers
handles 0/1/2+ existing bonds via peer-anchored extrapolation
Issuer existing curve
| Bond | TTM | OAS | Px | Rtg |
|---|
Top peer comps
| Cty | Bond | TTM | OAS | Rtg |
|---|
Pricing memo
Loading 13F data...
Loading CDS flow…
Loading legal screen…