Portfolio backtest: test an ETF portfolio on real returns
A portfolio backtest replays a mix of funds over past market data to show what it would have returned and how far it would have fallen. This free backtester runs up to ten ETFs with your weights on real monthly total returns, against a fund or a second portfolio, with rebalancing and cash flows. It also shows the risk behind the result: the deepest falls and how long they lasted, past crises, rolling returns and the share of risk of each holding.
1 · Your portfolio
- SPY SPDR S&P 500 ETF Trust · USD%since 2006
- BND Vanguard Total Bond Market ETF · USD%since 2007
● Total 100% · up to 10 holdings
2 · Compare with
3 · Settings
- Annual return (CAGR)
- +7.9%
- $10,000 became
- $44,701
- Worst fall
- −30.3%
- Sharpe ratio
- 0.66
May 2007 to Nov 2026: 235 months that all chosen funds and SPY have in common. Rebalanced every December.
| Your portfolio | SPY | |
|---|---|---|
| Annual return (CAGR) | +7.9% | +10.7% |
| $10,000 became | $44,701 | $72,643 |
| Volatility | 9.8% | 15.4% |
| Sharpe ratio | 0.66 | 0.62 |
| Sortino ratio | 0.99 | 0.93 |
| Worst fall (max drawdown) | −30.3% | −50.8% |
| Longest time below a previous high | 3.0 years | 4.4 years |
| Best calendar year | +22.3% | +32.3% |
| Worst calendar year | −19.3% | −36.8% |
| Best month | +8.1% | +12.7% |
| Worst month | −10.3% | −16.5% |
| Months with a gain | 65% | 66% |
Against SPY: correlation 0.98, beta 0.62. In the months SPY rose the portfolio moved 65% as much on average; in the months it fell, 60% as much.
Sharpe and Sortino use 1–3 month US Treasury bills (BIL) as the risk-free rate, over the 233 months it has data. Return, volatility and the ratios do not depend on the money added or withdrawn.
Growth of $10,000
Falls from the previous high
Distance from the previous high at each month-end.
The five deepest falls of your portfolio
| High | Low | Fall | High to recovery |
|---|---|---|---|
| Oct 2007 | Feb 2009 | −30.3% | 3.0 years |
| Dec 2021 | Sep 2022 | −20.2% | 2.2 years |
| Jan 2020 | Mar 2020 | −11.5% | 6 months |
| Apr 2011 | Sep 2011 | −8.1% | 9 months |
| Sep 2018 | Dec 2018 | −7.9% | 6 months |
How it held up in past crises
Return from the month-end before each episode to the month-end at its low. The dates follow the month-end highs and lows of the S&P 500.
| Episode | Your portfolio | SPY | Difference |
|---|---|---|---|
| Global financial crisisOct 2007 – Feb 2009 | −30.3% | −50.8% | +20.5% |
| 2011 euro debt crisis and US downgradeApr 2011 – Sep 2011 | −8.1% | −16.2% | +8.1% |
| Late-2018 sell-offSep 2018 – Dec 2018 | −7.9% | −13.5% | +5.7% |
| Covid crashDec 2019 – Mar 2020 | −10.8% | −19.4% | +8.7% |
| 2022 inflation and rate risesDec 2021 – Sep 2022 | −20.2% | −23.9% | +3.8% |
Every calendar year
Lighter bars: year not complete in this period.
Table of yearly returnsShowHide
| Year | Portfolio | SPY | Difference |
|---|---|---|---|
| 2026(partial) | +6.6% | +11.5% | −4.9% |
| 2025 | +13.5% | +17.7% | −4.3% |
| 2024 | +15.5% | +24.9% | −9.4% |
| 2023 | +18.0% | +26.2% | −8.2% |
| 2022 | −16.1% | −18.2% | +2.0% |
| 2021 | +16.4% | +28.7% | −12.3% |
| 2020 | +14.0% | +18.3% | −4.3% |
| 2019 | +22.3% | +31.2% | −9.0% |
| 2018 | −2.7% | −4.6% | +1.9% |
| 2017 | +14.5% | +21.7% | −7.3% |
| 2016 | +8.2% | +12.0% | −3.8% |
| 2015 | +1.0% | +1.2% | −0.3% |
| 2014 | +10.4% | +13.5% | −3.1% |
| 2013 | +18.5% | +32.3% | −13.8% |
| 2012 | +11.1% | +16.0% | −4.8% |
| 2011 | +4.3% | +1.9% | +2.4% |
| 2010 | +11.5% | +15.1% | −3.5% |
| 2009 | +17.3% | +26.4% | −9.1% |
| 2008 | −19.3% | −36.8% | +17.5% |
| 2007(partial) | +2.3% | +0.0% | +2.3% |
Every month
| Year | Jan | Feb | Mar | Apr | May | Jun | Jul | Aug | Sep | Oct | Nov | Dec | Year |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 2026 | 1.0 | 0.1 | −3.7 | 6.3 | 3.3 | −0.6 | −0.4 | 1.8 | −1.2 | 0.3 | −0.2 | +6.6%* | |
| 2025 | 1.9 | −0.0 | −3.2 | −0.3 | 3.2 | 3.8 | 1.3 | 1.7 | 2.6 | 1.7 | 0.3 | −0.1 | +13.5% |
| 2024 | 0.9 | 2.6 | 2.3 | −3.4 | 3.8 | 2.5 | 1.6 | 2.0 | 1.8 | −1.5 | 4.2 | −2.2 | +15.5% |
| 2023 | 5.0 | −2.5 | 3.3 | 1.2 | −0.2 | 3.9 | 2.0 | −1.3 | −3.9 | −1.9 | 7.4 | 4.2 | +18.0% |
| 2022 | −4.0 | −2.2 | 1.1 | −6.9 | 0.5 | −5.5 | 6.3 | −3.6 | −7.1 | 4.2 | 4.8 | −3.8 | −16.1% |
| 2021 | −1.0 | 1.0 | 2.3 | 3.6 | 0.5 | 1.8 | 2.0 | 1.8 | −3.4 | 4.5 | −0.5 | 2.9 | +16.4% |
| 2020 | 0.8 | −4.0 | −7.7 | 8.1 | 3.0 | 1.3 | 4.0 | 3.7 | −2.3 | −1.7 | 6.9 | 2.3 | +14.0% |
| 2019 | 5.2 | 2.0 | 1.9 | 2.5 | −3.4 | 4.7 | 1.0 | −0.0 | 1.0 | 1.5 | 2.3 | 1.8 | +22.3% |
| 2018 | 2.9 | −2.6 | −1.4 | −0.0 | 1.8 | 0.3 | 2.2 | 2.2 | 0.2 | −4.7 | 1.4 | −4.7 | −2.7% |
| 2017 | 1.1 | 2.6 | 0.1 | 0.9 | 1.1 | 0.4 | 1.4 | 0.5 | 1.1 | 1.5 | 1.9 | 1.0 | +14.5% |
| 2016 | −2.5 | 0.3 | 4.3 | 0.4 | 1.0 | 1.0 | 2.4 | −0.1 | 0.0 | −1.4 | 1.2 | 1.4 | +8.2% |
| 2015 | −0.8 | 2.8 | −0.7 | 0.5 | 0.6 | −1.7 | 1.7 | −3.8 | −1.2 | 5.0 | 0.1 | −1.1 | +1.0% |
| 2014 | −1.5 | 2.9 | 0.4 | 0.7 | 1.8 | 1.3 | −0.9 | 2.8 | −1.1 | 1.6 | 2.1 | −0.1 | +10.4% |
| 2013 | 2.8 | 1.0 | 2.4 | 1.6 | 0.8 | −1.4 | 3.4 | −2.2 | 2.4 | 3.3 | 1.8 | 1.5 | +18.5% |
| 2012 | 3.0 | 2.7 | 1.8 | −0.0 | −3.4 | 2.5 | 1.2 | 1.6 | 1.6 | −1.2 | 0.5 | 0.5 | +11.1% |
| 2011 | 1.4 | 2.2 | −0.0 | 2.3 | −0.2 | −1.2 | −0.6 | −2.6 | −3.8 | 6.2 | −0.3 | 1.1 | +4.3% |
| 2010 | −1.7 | 2.0 | 3.5 | 1.4 | −4.4 | −2.4 | 4.3 | −2.0 | 5.1 | 2.4 | −0.3 | 3.6 | +11.5% |
| 2009 | −5.8 | −6.5 | 5.1 | 5.9 | 3.8 | 0.2 | 5.0 | 2.7 | 2.6 | −1.1 | 4.4 | 0.7 | +17.3% |
| 2008 | −3.1 | −1.5 | −0.3 | 2.6 | 0.5 | −4.9 | −0.5 | 1.2 | −5.6 | −10.3 | −1.6 | 3.1 | −19.3% |
| 2007 | 1.7 | −1.0 | −1.5 | 1.4 | 2.6 | 1.2 | −1.6 | −0.3 | +2.3%* |
Monthly returns in %. * year not complete in this period.
Rolling returns
Annualised return over every period of that length inside the dates above: what you would have earned starting in any month.
| Holding period | Worst | Median | Best | Ended with a loss | |
|---|---|---|---|---|---|
| 1 year224 periods | Your portfolio | −25.5% | +10.8% | +33.5% | 15% |
| SPY | −43.4% | +15.3% | +56.2% | 16% | |
| 3 years200 periods | Your portfolio | −1.9% | +9.2% | +17.9% | 3% |
| SPY | −9.8% | +13.2% | +26.0% | 5% | |
| 5 years176 periods | Your portfolio | +2.9% | +9.0% | +15.6% | 0% |
| SPY | −1.0% | +13.9% | +22.9% | 1% |
Risk of each holding
Each holding over the same months, and how much of the portfolio’s ups and downs it accounts for at the target weights.
| Holding | Share of the money | Share of the risk | Money vs risk | Return a year | Volatility | Worst fall |
|---|---|---|---|---|---|---|
| SPY | 60% | 92% | +10.7% | 15.4% | −50.8% | |
| BND | 40% | 8% | +2.9% | 4.6% | −17.5% |
SPY has 60% of the money and 92% of the risk. Grey bar: share of the money; coloured bar: share of the portfolio’s variance.
Monthly total returns with dividends reinvested; the funds’ own costs are included, trading costs, taxes and inflation are not. Historical simulation, not a forecast and not investment advice.
Backtest results of six well-known portfolios
The same test on ready-made allocations, from a single stock fund to mixes of four and five funds: total return in US dollars, rebalanced once a year, all over the same period, Nov 2011 to Nov 2026. Each one can be loaded in the calculator above as a comparison.
| Portfolio | Funds | Return a year | Volatility | Sharpe | Worst fall | Back at the high after | Worst year |
|---|---|---|---|---|---|---|---|
| 100% Stocks | 1 | +14.1% | 16.9% | 0.77 | −35.0% | 6 months | −19.5% (2022) |
| 60/40 Portfolio | 2 | +9.3% | 10.3% | 0.77 | −21.8% | 5 months | −16.9% (2022) |
| Three-Fund Portfolio | 3 | +8.1% | 10.0% | 0.67 | −22.3% | 2.3 years | −16.2% (2022) |
| All Weather Portfolio | 5 | +5.7% | 8.0% | 0.54 | −23.8% | 3.8 years | −19.2% (2022) |
| Permanent Portfolio | 4 | +6.0% | 7.1% | 0.63 | −17.5% | 2.4 years | −12.5% (2022) |
| Golden Butterfly Portfolio | 5 | +7.0% | 8.3% | 0.67 | −19.5% | 2.7 years | −13.3% (2022) |
“Worst fall” is the largest peak-to-trough decline on daily values; “Back at the high after” counts from that peak to the day the portfolio returned to it. Before taxes and trading costs. Past performance is not a forecast. Source: MEDGE Capital backtests on end-of-day fund prices. All portfolios and method.
Over these years the all-stock portfolio returned +14.1% a year and fell 35.0% at its worst; the 60/40 portfolio returned +9.3% and fell 21.8%. The highest Sharpe ratio in the table belongs to 100% Stocks, at 0.77. A single period favours whatever did well in it: the comparison says how these mixes behaved, not which one is best.
Five market crises: stocks, bonds, gold and a 60/40 mix
The average return hides the moments that decide whether a portfolio is kept or sold. This table isolates them: the return of each asset from the month-end before the episode to the month-end at its low, for a US stock fund (SPY), a US bond fund (BND), gold (GLD) and a 60/40 mix of the first two rebalanced every year.
| Episode | Month-ends | Stocks (SPY) | Bonds (BND) | Gold (GLD) | 60/40 mix |
|---|---|---|---|---|---|
| Global financial crisis | Oct 2007 – Feb 2009 | −50.8% | +7.0% | +17.8% | −30.3% |
| 2011 euro debt crisis and US downgrade | Apr 2011 – Sep 2011 | −16.2% | +4.8% | +3.7% | −8.1% |
| Late-2018 sell-off | Sep 2018 – Dec 2018 | −13.5% | +1.6% | +7.5% | −7.9% |
| Covid crash | Dec 2019 – Mar 2020 | −19.4% | +2.2% | +3.6% | −10.8% |
| 2022 inflation and rate rises | Dec 2021 – Sep 2022 | −23.9% | −14.5% | −9.5% | −20.2% |
Month-end values with dividends reinvested, so the falls are smaller than the daily peak-to-trough figures usually quoted. The dates follow the month-end highs and lows of the S&P 500. Source: MEDGE Capital calculations on end-of-day prices.
The 60/40 mix did better than stocks in 5 of 5 episodes. Bonds rose in 4 of 5 and gold in 4 of 5. Bonds fell together with stocks in 1 of them (2022 inflation and rate rises: −14.5%), and there the mix had little to cushion it: the protection of bonds depends on the kind of crisis. The calculator above runs the same episodes on your own portfolio.
Does rebalancing change the result?
Rebalancing brings the weights back to their targets by selling part of what rose and buying what fell. Without it the fastest-growing holding takes a larger share every year, and the portfolio you end with is not the one you chose.
Perold and Sharpe compared buying and holding with rebalancing to a constant mix and showed that neither is best in every market: a constant mix does relatively well when markets swing back and forth, and relatively poorly when they move in one direction for a long time.[1] Rebalancing is first of all a way to keep the mix you chose.
A 60/40 mix of SPY and BND under four rules, May 2007 to Oct 2026
| Rebalanced | Return a year | Volatility | Worst fall | $10,000 became |
|---|---|---|---|---|
| Every month | +7.83% | 9.9% | −32.2% | $43,494 |
| Every quarter | +7.93% | 9.9% | −31.6% | $44,260 |
| Every year | +7.99% | 9.8% | −30.3% | $44,792 |
| Never | +8.68% | 10.5% | −28.0% | $50,695 |
Month-end rebalancing: every month, at the end of each quarter, every December, or never. Before taxes and trading costs, which rebalancing more often increases. Source: MEDGE Capital calculations on end-of-day prices.
Across the four rules the yearly return ranges from +7.83% to +8.68% and the volatility from 9.8% to 10.5%. The worst fall was 30.3% with yearly rebalancing and 28.0% without any. Left alone, the 60% in stocks became 86% of the portfolio by Oct 2026.
How to read a portfolio backtest
Annual return (CAGR). The constant yearly rate that turns the starting value into the final one. It says nothing about the path in between.
CAGR = (final value ÷ starting value)^(1 ÷ years) − 1
Example: $10,000 that becomes $20,000 in 10 years has a CAGR of 7.18%, whether it rose steadily or halved along the way.
Volatility. The standard deviation of the monthly returns, scaled to a year by multiplying by √12. It measures how much the returns swing around their average, in both directions.
Maximum drawdown. The deepest fall from a previous high, and the number that most often decides whether an investor holds on. A fall needs a larger gain to be recovered:
| Fall from the high | −10% | −20% | −30% | −40% | −50% |
|---|---|---|---|---|---|
| Gain needed to get back | +11.1% | +25.0% | +42.9% | +66.7% | +100.0% |
gain needed = 1 ÷ (1 − fall) − 1
Time below the previous high. How long the portfolio stayed under a level it had already reached. A fall of 20% that is recovered in six months and one that takes six years are different experiences with the same depth.
Sharpe and Sortino ratios. The Sharpe ratio is the average return in excess of a benchmark such as the risk-free rate, divided by the standard deviation of that excess return.[2] The Sortino ratio replaces total volatility with downside deviation, so that only returns below a minimum acceptable return count as risk.[3]
Sharpe = average (portfolio return − risk-free return) ÷ standard deviation of that difference
Both are estimates with a margin of error. Lo showed that scaling a monthly Sharpe ratio to a year by √12 is exact only when returns are independent from one month to the next, and that with serially correlated returns the annual figure can be overstated.[4] Small differences between two portfolios’ ratios mean little.
Rolling returns. The return over every period of the same length, for example every five years inside the test. They answer the question a single start date cannot: what would have happened starting in a different month.
Share of risk. How much of the portfolio’s ups and downs each holding accounts for. In a 60/40 mix of SPY and BND from May 2007 to Oct 2026, the 60% in stocks accounted for 92% of the risk. The weights you set are not the risks you take.
What a backtest cannot tell you
It is one path of history. The period tested contains one sequence of crises and recoveries, and the next one will differ. The mix that did best in one period is the one that held most of what rose in that period, which says little about the next.
The more variants you try, the better the best one looks. Bailey, Borwein, López de Prado and Zhu showed that high simulated performance is easy to obtain after trying a relatively small number of variants of a strategy, and that a backtest selected this way tends to disappoint out of sample; they call it backtest overfitting.[5] Harvey, Liu and Zhu counted hundreds of factors proposed to explain stock returns and argued that, with so many tests, a new one should clear a much higher bar than the usual one: a t-statistic above 3.0.[6]
Only the survivors are in the data. Brown, Goetzmann, Ibbotson and Ross showed that looking only at funds that survived can make past winners appear to keep winning even when they do not.[7] The catalog here contains funds that exist today; the ones that were closed along the way are not in it.
Costs, taxes and behaviour are outside. The funds’ own costs are in their returns; trading costs, taxes and inflation are not. Nor is the hardest part: holding the portfolio through its worst fall instead of selling at the low.
Five checks before trusting a backtest
- 1. Change the start date. Use the rolling returns and the date range: a result that holds only from one particular month is luck.
- 2. Read the worst fall before the return. Look at its depth and at how long the portfolio stayed below its previous high.
- 3. Compare with something simple. A single broad fund or a 60/40 mix: if the difference is small, the simpler portfolio wins.
- 4. Look at the crises. The episodes show whether the holdings fell together when it mattered.
- 5. Decide the rule first. A mix chosen for a reason and then tested is worth more than the best of twenty mixes tried.
How this backtest works
Data. Monthly total returns of each ETF with dividends reinvested, computed from end-of-day prices, for up to 20 years, and of each ready-made portfolio. Only the months all chosen funds have in common are used, so every fund is measured over the same period.
Rebalancing. At month-end: every month, at the end of each quarter, every December, or never. Between two dates the weights drift with each fund’s returns. A second portfolio follows the same rule.
Cash flows. The amount you add, or withdraw, enters at the start of every month after the first. Annual return, volatility and the ratios are time-weighted and do not depend on it; the money-weighted return of your cash flows is shown separately.
Risk measures. Volatility, Sharpe and Sortino are computed on monthly returns and scaled to a year. The risk-free rate is the monthly return of 1–3 month US Treasury bills (BIL), or of the euro overnight rate (XEON) for portfolios of euro funds only. Drawdowns are measured at month-ends, so a fall that reversed within a month looks smaller than on daily prices.
Share of risk. For each holding, its weight times its covariance with the portfolio, divided by the portfolio’s variance, at the target weights. The shares add up to 100%.
Not included. Trading costs, taxes and inflation: the results are nominal and before tax. Returns are in each fund’s trading currency.
Portfolio backtest: questions
What is a portfolio backtest?
A portfolio backtest replays a mix of funds over past market data: it shows what that mix, with a given rebalancing rule, would have returned and how far it would have fallen. It describes the past, which helps compare choices; it is not a forecast.
How do I backtest a portfolio with this tool?
Add up to ten ETFs from the catalog and set their weights, or start from a ready-made portfolio. Choose what to compare it with: a fund, a ready-made portfolio or a second portfolio of your own. Then set rebalancing, the starting amount, any monthly contribution or withdrawal and the dates. The results update as you type, with no sign-up.
Which data does the backtest use, and how far back does it go?
Monthly total returns with dividends reinvested, computed from end-of-day prices of each ETF, for up to 20 years. The backtest starts in the first month that all chosen funds have in common. For some ETFs the price history in our data starts later than the fund’s launch: the catalog shows the first year available for each fund.
Can I compare two portfolios?
Yes. Choose “A second portfolio” under “Compare with” and build Portfolio B with its own funds and weights. Both portfolios are tested on the same months, with the same rebalancing rule and the same cash flows, and every table shows them side by side.
Does rebalancing change the result?
For a 60/40 mix of SPY and BND from May 2007 to Oct 2026, rebalancing every year returned +8.0% a year with a volatility of 9.8% and a worst fall of −30.3%; never rebalancing returned +8.7% with a volatility of 10.5% and a worst fall of −28.0%, and the stock share drifted from 60% to 86%. Without rebalancing the portfolio you end with is not the one you chose.
Can I include contributions or withdrawals?
Yes. Enter an amount to add each month, or a negative amount to withdraw. The growth chart then shows the value against the money paid in, and the table adds the money-weighted return of your cash flows. Annual return, volatility and the risk ratios stay time-weighted, so they do not depend on the cash flows.
What are CAGR, maximum drawdown and the Sharpe ratio?
CAGR is the constant yearly growth rate that turns the starting value into the final one: $10,000 that becomes $20,000 in 10 years has a CAGR of 7.18%. Maximum drawdown is the deepest fall from a previous high. The Sharpe ratio divides the average return in excess of the risk-free rate by the volatility of that excess return: the higher, the more return per unit of risk.
Can I backtest European ETFs?
Yes: UCITS ETFs such as VWCE, SWDA, CSSPX or EIMI are in the catalog. Returns are in each fund’s trading currency, so mixing US-dollar and euro funds also mixes currency effects; the calculator warns you when that happens.
Is a good backtest a reason to expect the same result in future?
No. A backtest is one path of history, and the more variants you try before choosing, the more likely the best one looks good by chance. Check the result over different start dates with the rolling returns, look at the worst fall and how long it lasted, and compare with a simple benchmark.
What does the full MEDGE backtest add?
Daily data, any listed ticker and many more holdings, transaction costs, per-holding attribution, risk contribution, stress tests on past crises and hypothetical shocks, and Monte Carlo projections on your actual weights.
Sources
- Perold, A. F., & Sharpe, W. F. (1988). Dynamic Strategies for Asset Allocation. Financial Analysts Journal, 44(1), 16–27. doi.org/10.2469/faj.v44.n1.16
- Sharpe, W. F. (1994). The Sharpe Ratio. The Journal of Portfolio Management, 21(1), 49–58. doi.org/10.3905/jpm.1994.409501
- Sortino, F. A., & Price, L. N. (1994). Performance Measurement in a Downside Risk Framework. The Journal of Investing, 3(3), 59–64. doi.org/10.3905/joi.3.3.59
- Lo, A. W. (2002). The Statistics of Sharpe Ratios. Financial Analysts Journal, 58(4), 36–52. doi.org/10.2469/faj.v58.n4.2453
- Bailey, D. H., Borwein, J. M., López de Prado, M., & Zhu, Q. J. (2014). Pseudo-Mathematics and Financial Charlatanism: The Effects of Backtest Overfitting on Out-of-Sample Performance. Notices of the American Mathematical Society, 61(5), 458. doi.org/10.1090/noti1105
- Harvey, C. R., Liu, Y., & Zhu, H. (2016). … and the Cross-Section of Expected Returns. The Review of Financial Studies, 29(1), 5–68. doi.org/10.1093/rfs/hhv059
- Brown, S. J., Goetzmann, W., Ibbotson, R. G., & Ross, S. A. (1992). Survivorship Bias in Performance Studies. The Review of Financial Studies, 5(4), 553–580. doi.org/10.1093/rfs/5.4.553
More free tools
- Rolling returnsBest and worst 1–15 year periods
- Drawdown and recoveryEvery fall and how long it lasted
- Efficient frontierRisk and return of every mix
- ETF correlation matrixHow your funds move together
- DCA calculatorA regular plan on real returns
- Rebalancing calculatorTrades to get back to target weights
- Lazy portfolios comparedBacktests of famous portfolios
- Portfolio diversification guideFormulas, calculators and data
Source: MarketStack end-of-day prices, total return with dividends reinvested; academic sources listed above. Figures are historical and descriptive; they are not forecasts and not investment advice. MEDGE Capital does not recommend buying or selling any instrument. Risk disclaimer.