New Backtest

Configure a portfolio allocation strategy and run it against historical data. Each combination of model, covariance estimator, risk measure, and allocation method produces a distinct portfolio. All 144 HRP combinations are backed by published empirical research.

Choose HRP Model to configure a hierarchical risk parity variant, or Benchmark to run a traditional allocation strategy for comparison.
HRP: recursive bisection (Lopez de Prado 2016). HERC: cluster-aware equal risk (Raffinot 2018). NCO: two-stage optimization (Lopez de Prado 2019). Constrained HRP: with box/group weight limits.
S&P Sectors: 11 sector ETFs. Multi-Asset: equities, bonds, commodities, REITs. Country Indices: global equity ETFs. Factors: momentum, value, quality, etc.
Ward: minimizes within-cluster variance, produces balanced groups (recommended). Complete: uses maximum inter-cluster distance.
Gerber: noise-robust co-movement statistic (Gerber & Markowitz 2021). Ledoit-Wolf: optimal linear shrinkage. RMT Denoised: eigenvalue cleaning via Random Matrix Theory.
Variance: standard portfolio volatility. CVaR: expected loss in the worst 5% of scenarios. CDaR: expected drawdown in the worst 5% — strongest empirical results for HERC (Raffinot 2018).
IVP: inverse variance — higher weight to lower-variance assets. ERC: equal risk contribution — each asset contributes equally to cluster risk.
Beginning of the backtest period. At least 2 years of history recommended for stable covariance estimation.
Monthly is standard. Quarterly reduces turnover.
Set to 0 for frictionless backtests. 10 bps is a reasonable default for liquid ETFs.
Advanced Parameters
252 = 1 year (standard). 126 = 6 months. 504 = 2 years.
63 = 1 quarter minimum history before inclusion.
Portfolio holds cash/equal weight during warm-up.
0.0 = no minimum. 0.05 = at least 5% per asset.
1.0 = no maximum. 0.25 = max 25% per asset.
0.05 = worst 5% of outcomes (standard). 0.01 = worst 1%.

Custom Universe

Create a custom universe by entering ticker symbols, or upload a CSV file with price data.

From Tickers

Enter Yahoo Finance ticker symbols separated by commas.

From CSV

CSV with columns: date, ticker1, ticker2, ... (prices as values)