Sumbra Sumbra live real market data loading… ← Whitepaper

Macro Sumbra Index

0 100

Regime probability

decoded from the cross-asset panel

Liquidity drivers

Macro Sumbra Index — trailing window

shaded bands = decoded regime · solid = index · dashed = mean pairwise correlation

Conditional correlation

trailing-window correlation across the panel
correlation heatmap
strong co-movehedge / negative

Sensitivity to the tide (β)

per-asset beta to Δindex, normalized to SPX = 1
Assetβexposure

Hidden Markov regime model

3 latent states
Risk-On Neutral Risk-Off

Transition matrix A

P(st | st-1) — estimated from the regime sequence

Regime history · 36 weeks

onneutraloff

Regime statistics

measured conditional behaviour of the cross-asset panel
RegimeMean ρ̄Ann. volShare of timeAvg duration

Latest report

model output for the most recent epoch
Index & regime are computed from live market data. The on-chain contract is not yet deployed — operator/finalization fields are testnet placeholders.

Oracle parameters

protocol design · testnet
Settlement chainBase (L2)
Contractnot deployed
Reporting epoch1 day
Challenge window6 h
Operators1 (single signer)
Verificationoptimistic

Recent feed

computed report per epoch (index & regime are real)
EpochRegimeIndexP(on)OperatorStatus

Pipeline

how each report would reach the chain
Off-chain
data pipeline
model engine
operator node
Bridge
signed report
{regime, index, hash, sig}
On-chain · L2
SumbraOracle.sol
store · challenge · finalize

System architecture

how the feed is computed and served
Market data Yahoo Finance Engine model · HMM · GARCH data.json + heatmap.svg Frontend dashboard · paper

Data refresh

how the dashboard stays current — no manual refresh
Scheduler CI 6h · local 2min Recompute engine.pipeline Publish data.json → Pages Browser polls 45s → updates