Markets change.
Evidence matters.
Map structural shifts across global equity indices and digital assets with Bayesian change-point inference, synchronized returns and transparent sensitivity analysis.
Build your market study
Select a research cohort and run the same inference pipeline across observed weekly returns.
Market universe
4 of 14 selectedPick up where you left off.
Save market selections and model settings locally. Presets never include private CSV contents or historical prices.
No saved workspaces yet. Save a configuration to reuse it later on this device.
Source-attributed market series
Agreement in at least 6 of 9 models
Cross-market temporal overlap
Latest observed close date
Structural change, in context.
Weekly log returns and Bayesian posterior reset probabilities. Both series use separate axes.
Your next insight starts with evidence
Choose a market cohort and run the study to plot actual observed returns and model probabilities.
Cross-market relationships
Pairwise Pearson correlations on common observed UTC week ends; minimum 12 overlapping weeks.
Relationships need observations
Run two or more markets with overlapping dated returns to calculate their correlations.
Correlation measures co-movement, not causality or calibrated risk.
Changing risk conditions
Annualized sample standard deviation of the latest 12 weekly log returns.
Volatility, without guesswork
Analysis calculates rolling descriptive volatility only when enough historical observations exist.
Annualization assumes 52 weekly periods; historical dispersion is not a forecast.
When markets shift together.
Clusters of sensitivity-stable peaks within one observed week. These are research candidates, not verified real-world events.
No shared events identified yet
Run the study to compare synchronized candidate structural changes across markets.
The Gaussian mean-shift BOCPD model uses whole-sample variance and a first-observation prior. Results are retrospective; nine-setting sensitivity is parameter agreement, not prediction confidence. Yahoo Finance is an unofficial historical source, and free APIs may fail. Imported files are processed client-side. Read the full methodology.