v1.4.0

Intraday data, simulation and a visible efficient frontier.

• Hourly intraday prices as a second data granularity, with daily realized volatility computed from within-day returns only
• Monte Carlo value path simulation by bootstrap resampling or a fitted normal, with final-value percentiles and the probability of loss
• An Optimization dashboard view with a long-only portfolio cloud and the closed-form minimum variance and maximum Sharpe portfolios marked
• 158 tests, all metric math checked against hand-computed values

v1.3.0

Optimization, model validation and factsheets.

• Closed-form Markowitz optimization: minimum variance and maximum Sharpe portfolios with numpy only, served at POST /portfolio/optimize
• VaR model validation with a rolling no-lookahead backtest and the Kupiec proportion-of-failures test, via API and CLI
• One-page PDF risk factsheets per asset with metrics, price and drawdown charts
• CLI smoke tests covering every subcommand
• 130 tests, all metric math checked against hand-computed values

v1.2.0

Benchmark analytics, backtesting and a model monitor.

• Beta, CAPM alpha, tracking error and information ratio against a configurable benchmark, via API and CLI
• Rebalanced portfolio backtesting, monthly or quarterly against buy-and-hold, with hand-computed path tests
• A Model Monitor dashboard view showing the volatility model comparison and the drift indicators
• 93 tests, all metric math checked against hand-computed values

v1.1.0

Tail risk, a second data source and model monitoring.

• Historical Value at Risk and expected shortfall at 95 percent confidence, in the metrics summary, the portfolio endpoint and the dashboard
• ECB reference exchange rates as a second provider through the official data API, ingested with the new ingest-fx command
• Return distribution drift monitoring with the population stability index and the two-sample KS statistic, served at /assets/{symbol}/drift and the drift CLI command
• 72 tests, all metric math checked against hand-computed values

v1.0.0

First complete version of the platform.

• Daily price ingestion from the public Yahoo Finance chart endpoint, computed on adjusted closes, with an offline demo mode
• Portfolio risk metrics: total return, annualized volatility, Sharpe ratio, max drawdown and correlations, tested against hand-computed values
• One-day-ahead volatility forecasting comparing a rolling baseline, RiskMetrics EWMA and a HAR regression in a walk-forward test
• Streamlit dashboard with single asset and portfolio views
• REST API for assets, prices, metrics and forecasts
• docker compose stack with PostgreSQL, smoke tested in CI
• scheduled workflow that proves the live data pipeline after every US trading day
• CI with Ruff, mypy, a coverage floor, the suite on SQLite and PostgreSQL, CodeQL and Dependabot auto-merge