Portfolio Backtester (Your Own Price Files)

About the Portfolio Backtester (Your Own Price Files)

See how a mix of assets you choose would have behaved on prices you supply: load one file with a date column and a column of prices for each asset (or one file per asset), set the weights, and the backtester replays the mix with your rebalancing rule — never, every month, quarter or year, or whenever a weight drifts outside a band — your regular contributions or withdrawals and your trading costs.

It reports the final value, the compound annual growth rate, volatility, the maximum drawdown with its dates and recovery, the Sharpe and Sortino ratios, every calendar year, rolling 1-, 3- and 5-year returns, what each asset added, the turnover and the costs, and the same mix against an equal-weight mix and each asset alone. Prices in other currencies are converted with an exchange-rate file or a rate you type.

It works for any market and any currency, from daily or monthly prices. Files are read in your browser and never uploaded, and the tool uses no market data: every figure comes from your own file.

Common uses

• See how often rebalancing a 60/40 mix would have traded, and what it changed, on prices you exported.
• Compare the drawdowns and volatility of two mixes of the same funds before you decide how to split your own money.
• Check how much regular contributions or withdrawals changed the outcome over the period in your file.
• Test a threshold band against calendar rebalancing on the same data and the same costs.

The formulas

• Units are bought at the target weights on the first date; on each rebalancing date the units are reset to the target weights at that date’s prices. Scheduled rebalancing uses the last price of each calendar month, quarter or year in your data; band rebalancing happens on any date a weight is further from its target than the band you set. Neither happens on the last date.
• Contributions are made on the first price of each new month, quarter or year and buy at the target weights; withdrawals sell every holding in proportion.
• Costs are your percentage of the value traded, on every trade, the first purchase included, paid from the portfolio.
• Returns are time-weighted: each period’s return is the value before that date’s contribution over the value after the previous date’s, so money you add is not counted as growth. CAGR = (1 + total return)^(1 ÷ years) − 1.
• Volatility = the sample standard deviation of the period returns × √(periods a year): 12 for monthly prices, 252 for daily prices without weekends, 365 with them.
• Sharpe ratio = mean excess return ÷ its standard deviation × √(periods a year) (Sharpe); Sortino ratio = mean excess return ÷ downside deviation × √(periods a year) (Sortino and van der Meer), both against the risk-free rate you give.
• Maximum drawdown = the deepest fall of the growth of the mix from a running high, with the dates of the high, the low and the recovery.
• Rolling returns are the annualised returns of every 1-, 3- and 5-year window that ends on a date in your data.

Reading the results

• The final value includes everything you added or took out; the total return and CAGR do not count those flows as growth. With regular flows the tool also gives the money-weighted return (XIRR) of your own cash flows.
• A drawdown is a fall from a previous high: −20% means the mix was a fifth below its best value so far. The recovery date is when it got back to that high.
• Rolling returns show how much the result depended on the start date: the spread between the worst and the best window is often wider than the whole-period figure suggests.
• The comparison runs the same dates, flows and costs with equal weights and with each asset alone, so the effect of the mix itself is visible.

What it leaves out

Taxes, dividends and interest that are not in your prices (use total-return prices or adjusted closes to include them), bid-ask spreads, slippage, fund charges not in the prices, minimum trade sizes and whole-unit rounding, and the cost of exchanging currency. A backtest uses prices that are known only afterwards and assumes every trade filled at the closing price; it describes your data, and the rules for simulated past performance are why the tool never projects it forward.

Limitations

• Prices only: dividends count only when your file holds total-return or adjusted prices.
• Every asset needs a price on a date for that date to be used: the backtest runs on the dates all your assets share.
• Long positions only, with weights from 0% to 100%; no leverage and no short sales.
• Up to 20 assets; files up to 50 MB.
• It describes the past of the prices you load and projects nothing forward.

Privacy

Everything happens in your browser. What you enter or open here is not uploaded or stored by MySmartCoPilot.