How to Backtest a Portfolio (Without a Spreadsheet)

If you have ever wondered what would have happened if you had put $10,000 into the S&P 500 ten years ago instead of leaving it in a savings account, you have already had the instinct that drives every serious investor: you want to backtest a portfolio before committing real money to it.

Backtesting is the process of applying a portfolio idea to historical price data and measuring how it would have performed. It is not a crystal ball, but done honestly it tells you three things you cannot easily get any other way: what kind of return the strategy has produced through real market cycles, how painful the worst stretches were, and whether the idea actually beats the boring default of buying a broad index fund. This guide walks through exactly how to do it — what numbers to look at, how to run a backtest in our free portfolio simulator, and the traps that quietly invalidate most of the backtests people share online.

What backtesting actually proves (and what it doesn't)

A good backtest answers one specific question: given the historical price series I am using, with these specific contributions and this specific allocation, what would my portfolio look like today? That is it. It is a measurement, not a forecast.

What backtesting can legitimately do:

• Tell you the realized return of a strategy over a real period that included recessions, bubbles and recoveries.
• Show you the worst drawdown — the largest peak-to-trough drop — you would have had to live through.
• Let you compare two strategies fairly, over the exact same window, with the exact same contributions.

What backtesting cannot do:

• Predict the next ten years. Past returns are evidence, not prophecy.
• Account for emotional decisions you would actually have made. The chart looks calm in hindsight; the experience of watching your portfolio fall 50% over 18 months in 2008 was not calm.
• Replace a clear understanding of what a portfolio is and why diversification matters.

If you remember nothing else from this article, remember that a backtest is a description of a single historical path. Markets only ran the experiment once.

The four numbers that actually matter

When someone shows you a backtest and only quotes the total return, be suspicious. A useful backtest reports at minimum these four numbers:

• Final value. The dollar amount your initial investment plus contributions would be worth today. This is the headline, but on its own it is misleading because it does not say how long it took or how bumpy the ride was.
• CAGR (compound annual growth rate). The smoothed annual return that would have produced the final value. CAGR is the right way to compare strategies that ran for different lengths of time. A 100% total return over 5 years (~14.9% CAGR) is very different from 100% over 15 years (~4.7% CAGR).
• Max drawdown. The deepest peak-to-trough decline along the way, in percent. This is the number that tells you whether you would have actually stuck with the strategy. A 9% CAGR with a 60% drawdown is a much harder portfolio to hold than a 7% CAGR with a 25% drawdown.
• Best and worst calendar year. A quick sanity check on volatility. If the worst year is -45%, you need to be honest about whether you would have added money that January or sold in panic.

Anything else — Sharpe ratio, Sortino, alpha, beta — is useful but secondary. Get these four right first.

How to backtest a portfolio in the simulator

Here is the workflow we use ourselves. It takes about three minutes.

1. Open the What-If Portfolio simulator. Pick "Past" mode. 2. Set the initial investment. A round number like $10,000 makes the final value easy to interpret. 3. Add your assets. Search by ticker (for example VOO, AGG, BTC-USD) and set the percentage weight for each. The weights must sum to 100%. 4. Pick the time range. Start with the longest range your assets support — short windows lie. If one of your assets only has 5 years of data, the whole backtest is capped to 5 years. 5. Optional: add recurring contributions. A monthly DCA dramatically changes the story for volatile assets because it lowers your average entry price. 6. Read the stats bar. You get final value, total return, CAGR, best year, worst year, max drawdown and Sharpe ratio for the exact window you selected.

If you just want to see worked examples without setting anything up, we publish hundreds of pre-computed scenarios. Three good starting points:

• What if I invested in VOO 10 years ago — the boring, hard-to-beat S&P 500 benchmark.
• What if I invested in QQQ 10 years ago — same period, but Nasdaq-100 concentrated.
• What if I invested in AAPL 10 years ago — single-stock concentration, for contrast.

Comparing these three to each other is itself a great first lesson in why diversification exists.

The mistakes that quietly ruin a backtest

Most published backtests are wrong in at least one of these five ways. When you read someone else's, scan for these. When you run your own, avoid them.

1. Cherry-picked start dates. "Strategy X returned 18% per year since 2009" sounds amazing until you notice 2009 was the bottom of the Global Financial Crisis. Start one year earlier and the number collapses. Always check what the same backtest looks like over multiple windows — last 5, 10, 15 and 20 years.

2. Survivorship bias. If you backtest "the top 10 stocks of the S&P 500" you are implicitly using today's top 10, which is exactly the group that won. The losers got delisted and erased. Broad index ETFs largely solve this; hand-picked stock baskets do not.

3. Ignoring fees and taxes. Most backtesters (ours included) use raw price-return data. In real life you pay expense ratios, sometimes transaction fees, and you may owe tax on dividends and rebalancing. For a fair comparison between two ETFs, use total-return data and subtract the expense-ratio difference.

4. Mistaking a single path for the distribution. History happened once. The next 10 years will not be a copy. If you really want to stress-test a portfolio, run a Monte Carlo simulation on top of the historical mean and volatility — but treat both as evidence, not as truth.

5. Backtesting only the good times. If your window does not include at least one real bear market (2000-2002, 2008, 2020, 2022), your max drawdown number is fiction. Always extend the window until you have lived through some pain.

A worked example: 60/40 vs 100% stocks

Let's say you are trying to decide between two allocations:

• Portfolio A: 100% VOO (S&P 500 ETF)
• Portfolio B: 60% VOO / 40% AGG (US aggregate bond ETF)

Run both in the simulator over the longest available window, with the same $10,000 starting amount. You will typically see Portfolio A finish with a meaningfully higher final value and a meaningfully higher CAGR — but also a much deeper max drawdown and a much worse worst year. Which one is "better" depends entirely on whether you would have sold Portfolio A at the bottom of 2008 or 2022. If the honest answer is "probably yes", Portfolio B is the better portfolio for you, even though it underperforms on paper.

This is exactly the kind of question a backtest is designed to answer, and exactly the kind of question that no amount of staring at the news will answer for you.

A short checklist

Before you trust any backtest — yours or someone else's:

• Window includes at least one bear market. ✓
• Same start and end date for every strategy being compared. ✓
• Same contribution schedule for every strategy. ✓
• All four headline numbers reported, not just total return. ✓
• Source data is total return, or expense ratios are accounted for. ✓

If anything on that list is missing, the backtest is decoration, not evidence.

FAQ

Is backtesting the same as investing?

No. Backtesting tests a portfolio idea against history. Investing is what happens when you buy the portfolio. A backtest can rule out bad ideas quickly, but a good backtest does not guarantee a good outcome — it just stacks the odds in your favor.

How many years of data do I need for a meaningful backtest?

At least 10 years, ideally 15-20, and the window must include at least one significant bear market. Shorter windows over a single bull market will make almost any aggressive strategy look brilliant.

Should I backtest with a lump sum or with monthly contributions?

Both, and compare them. A lump sum is the cleanest test of an asset's underlying return. A monthly contribution (dollar-cost averaging) is closer to how most real investors actually buy, and it changes the result significantly for volatile assets.

What does CAGR mean in plain English?

CAGR is the constant annual rate of return that, if compounded every year, would turn your starting value into your ending value. If $10,000 grew to $20,000 over 10 years, the CAGR is roughly 7.2% — meaning the portfolio behaved as if it gained exactly 7.2% every single year, even though the actual yearly returns bounced around.

Why is max drawdown more important than volatility?

Volatility (standard deviation) treats upside moves and downside moves equally. Max drawdown only measures actual losses, which is what investors actually care about and what actually causes people to sell at the wrong time. For practical portfolio decisions, drawdown is the more honest number.

---

Ready to run your own? Open the simulator and try the 60/40 vs 100% stocks experiment from this article — it takes about two minutes and it will probably change how you think about your own allocation. If you want a foundation first, start with what an investment portfolio actually is.