About Backvest

Backvest backtests real mutual funds and ETFs — not abstract asset classes — and is built around one idea: the most valuable output of a backtest is knowing how much not to trust it.

The honesty principles

Ties are ties

When two portfolios differ by less than the noise in the data (a few hundredths of a percent, or the uncertainty from stand-in history), Backvest shows them at the same rank. Most tools rank them anyway — that ranking is fiction.

Regions, not points

A search that says “exactly 62% stocks” is overfit. Backvest reports ranges — “55–65% large-cap works about equally well” — because answers that survive small changes are the ones that survive the future.

Winners rarely repeat

Every search stress-tests its winners: reshuffle history (bootstrap) and re-pick winners over time (walk-forward). When past champions land mid-pack out of sample — which is typical — Backvest says so on the result.

Every shortcut is disclosed

Funds younger than your test window get history assembled from older share classes or benchmarks. Every splice appears in a methodology table with a confidence grade, and rankings inside that uncertainty render as ties.

Survivorship is unfixable — so it's labeled

Funds that closed or merged aren't in any free dataset, so every screen quietly favors survivors. Backvest can't cure that; it prints the warning wherever it matters instead of pretending.

How the numbers are made

  • Monthly total returns (dividends and distributions reinvested) from fund NAV and ETF price histories; mutual-fund returns are already net of fund expenses.
  • Backtests simulate real rebalancing (monthly to annual, threshold, or none), contributions and withdrawals, and use time-weighted returns for risk metrics.
  • Rolling-window analysis measures every possible start date, so a result that depends on one lucky entry point is visible immediately.
  • Searches test every allocation on a grid (up to millions), then stress-test the leaders. Large searches use a disclosed two-stage scan-then-refine method.
  • Data quality is graded per fund (A–F); gaps and suspicious jumps block analyses loudly instead of being silently patched.

Data sources & limits

Prices and distributions currently come from public market-data feeds; expense ratios are seed values pending verified filings, and are flagged wherever shown. Fund descriptions link to Morningstar and Yahoo Finance for independent research. Institutional-grade exchange data is on the roadmap.

Disclaimer. Backvest is research software, not an investment adviser. Everything shown is historical simulation, not a forecast and not a recommendation to buy or sell any security. Past performance does not guarantee future results. Consider costs, taxes, and your own situation — ideally with a professional — before acting.