A backtest runs a trading rule over past prices. We tested one simple rule on the Nasdaq-100 index. It holds a fund that tracks the index while the fund's price is above its average of the last 200 trading days. Otherwise, it holds cash. From August 2017 to September 2026, it returned 16.9% a year after costs. Holding the fund returned 20.7%.
On that line alone, the rule lost. The rest of the report tells a fuller story. The rule fell less and swung less. Its return for each unit of risk was about the same as holding. Here is how to read a backtest report line by line, using that one rule as the worked example.
How do you read a backtest report? A backtest report is the scorecard of a trading rule tested on past prices. Read it in six steps: the set-up, the return lines, the risk lines, a fair comparison, the trade lines and the year-by-year view. Judge each one against simply holding the asset.
Return alone misleads. A rule holding QQQ (the Nasdaq-100 exchange-traded fund, or ETF) only while it was above its 200-day average returned 16.9% a year, against 20.7% for holding it. Its worst fall was 22.9%, against 35.1%.
Risk-adjusted return was about the same. The Sharpe ratio is return above cash divided by volatility, which is how much returns swing in a year. It was 0.84 for the rule and 0.81 for holding.
A low win rate can still work. The rule made money on 5 of its 20 completed round trips, each a purchase and the sale after it. But the mean win was 39.6% and the mean loss 2.4%.
Set-up lines: rule, data, period, timing, costs and benchmark
We explain how a backtest works in What Is Backtesting? How a Backtest Can Fool You.
A report should open with its set-up. Without these lines, none of the numbers below can be checked. Ours:
Rule: hold the Invesco QQQ Trust (QQQ), a fund that tracks the Nasdaq-100, when its close is above its 200-day moving average. That average is the mean close of the last 200 trading days. Otherwise, hold 3-month Treasury bills, which are loans to the US government.
Data: YX Insights price data, with dividends counted. Treasury bill rates come from the Federal Reserve Economic Data (FRED) database, series DTB3.
Period: 11 August 2017, the first day with a full 200-day average, to 24 September 2026.
Timing: decide at the close, then trade at the next day's close.
Costs: 0.1% of the amount traded on every switch in or out.
Benchmark: holding QQQ throughout, with no costs after the first purchase.
The full backtest report for the QQQ 200-day rule
Line | 200-day rule, after 0.1% costs | Holding QQQ |
|---|---|---|
Total return | +315.8% | +454.5% |
Annual return | 16.9% | 20.7% |
Annual return before costs | 17.4% | 20.7% |
Volatility (yearly) | 17.3% | 23.4% |
Sharpe ratio | 0.84 | 0.81 |
Maximum drawdown | −22.9% | −35.1% |
Longest time below a previous high | 17.5 months | 23.5 months |
Time in the market | 82.3% | 100% |
Switches in or out | 41 | 1 (the first purchase) |
Completed round trips won | 5 of 20 (25%) | n/a |
Mean win / mean loss per round trip | +39.6% / −2.4% | n/a |
Best / worst calendar year, 2018 to 2025 | +37.3% (2023) / −16.0% (2022) | +54.9% (2023) / −32.6% (2022) |
Source: YX Insights price data; FRED (DTB3); YX Insights
Each line is read against the holding column. The terms are explained below, in order. A round trip is one purchase and the sale that follows it. Drawdown is the fall from a previous high.
Return lines: total and annual return
Source: YX Insights price data; FRED (DTB3); YX Insights
The chart shows $100 growing in each. Total return is the gain over the whole period. The rule turned $100 into $416, a gain of 315.8%. Holding QQQ turned $100 into $555, a gain of 454.5%.
Annual return is the steady yearly rate that gives the same end value. It was 16.9% for the rule and 20.7% for holding, a gap of 3.7 points. The rule held Treasury bills on the other 17.7% of days, when they paid a mean of 2.61% a year.
Check the cost line too. Before costs, the rule returned 17.4% a year, so costs took 0.5 points. At 0.25% a trade, the return falls to 16.1%. At 0.5%, it falls to 14.8%.
Risk lines: drawdown, volatility and Sharpe ratio
Source: YX Insights price data; FRED (DTB3); YX Insights
The chart shows each line's fall from its previous high, its drawdown.
Maximum drawdown is the deepest fall. The rule's was 22.9%, from August 2018 to June 2019. Holding QQQ fell 35.1%, from December 2021 to November 2022. Between August 2018 and June 2019, the rule switched 15 times. All seven round trips that began in that time lost money.
Time below a previous high shows how long the pain lasted. The rule's longest was 17.5 months. For holding, it was 23.5 months.
Volatility is how much the daily returns swing, scaled to a year. It was 17.3% for the rule and 23.4% for holding.
Sharpe ratio is the return above cash divided by volatility. The rule scored 0.84. Holding scored 0.81. The two are close.
Fair comparison: a mix with the same exposure
A fair test asks how much of the result comes from simply owning less of the asset. Take a mix that held 82.3% in QQQ every day, the rule's average, with the rest in Treasury bills. That mix returned 17.7% a year. Holding returned 20.7%, while the rule returned 16.9%. So owning less QQQ explains 2.9 of the 3.7-point gap.
The mix's worst fall was 29.3%. The rule's was 22.9%. So the timing cut the worst fall by a further 6.4 points, at a cost of 0.8 points a year in return.
Trade lines: time in the market, switches and win rate
The rule held QQQ on 82.3% of days. It switched 41 times, about 4.5 times a year. That gives 20 completed round trips. A 21st trade, bought on 8 April 2026, was still running at the end.
Only 5 of the 20 round trips made money, a win rate of 25%. The mean win was 39.6%, while the mean loss was 2.4%. The biggest single win was 81.8%. Of the 15 losses, 12 lasted 10 trading days or fewer. These are whipsaws, where the price crosses the average and quickly crosses back.
So read the win rate with the size of wins and losses. This rule made its money in a few long trends. It paid for them with many small losses.
Consistency lines: is the result reliable year by year?
Source: YX Insights price data; FRED (DTB3); YX Insights
Of the eight full calendar years, 2018 to 2025, the rule beat holding in only one. That was 2022, when it lost 16.0% against 32.6% for QQQ. It matched holding in 2021 and 2024, when it never left the market. It lagged most in 2019, at 18.6% against 39.0%. In 2023, it made 37.3% against 54.9%. Both times, it bought back in late. QQQ had risen 19.3% from its December 2018 low when the rule bought on 14 February 2019. It had risen 13.4% from its November 2022 low when the rule bought on 31 January 2023.
A split test asks whether the result holds in both halves of the period. From August 2017 to December 2021, the rule's Sharpe ratio was 0.97. Holding scored 1.10. From January 2022 to September 2026, the rule scored 0.71. Holding scored 0.54. So the rule did better only in the second half. Year by year, that comes down to 2022.
The start date matters as well. Over March 2011 to August 2026, our 200-day moving average essay found the same rule on QQQ had a lower Sharpe ratio than holding, even before costs.
What this backtest report says about the rule
Read line by line, the report says the rule lost on return. It fell less and swung less. Its Sharpe ratio was about the same as holding. Most of its return gap came from holding less QQQ. It did better only in 2022, so one year carries the result.
How to read a backtest report: a checklist
Set-up lines: rule, data, period, timing, costs and benchmark.
Return lines: annual return against holding, before and after costs.
Risk lines: maximum drawdown, time below the high, volatility and the Sharpe ratio.
Fair comparison: a mix with the same average exposure.
Trade lines: number of trades, win rate, mean win and mean loss.
Consistency lines: year-by-year returns and a split of the period.
A live record is the result of trades made after the rule was fixed, so it cannot be re-run with hindsight. In a 2016 study, R. David McLean and Jeffrey Pontiff found the returns of 97 published patterns were 58% smaller after publication. When a live record exists, compare it with the backtest over the same months.
Read the set-up first. Then judge return, risk and trades against holding the asset. Compare the rule with a mix at the same exposure. Check it year by year as well as in total.
Learn more with YX Insights
This explainer is part of the YX Insights Academy. Each one takes a single idea and checks it against real data.
The same approach runs through everything else we publish:
Systematic Portfolio: ready-made portfolios for Macro & Megacaps and for Commodities. We publish the holdings and every change.
Multi-model Signals: a daily long-or-flat call on every name we cover. Each call shows how strongly our four models agree.
Research: company deep dives, macro commentary and essays on how we test.
Good places to start on the website:
Academy: more explainers like this one.
Start here: get our free weekly letter by email.
Macro & Megacaps Systematic Portfolio and Commodities: the two portfolios in detail.
Track record: live results since 3 August 2026, plus the ten-year backtest.
Upgrade: get the Systematic Portfolio, Multi-model Signals and research in full.
DISCLAIMER: This article is strictly educational. Any information or analysis in this note is not an offer to sell or the solicitation of an offer to buy any securities. Nothing in this note is intended to be investment advice and nor should it be relied upon to make investment decisions. Any opinions, analyses, or probabilities expressed in this note are those of the author as of the note's date of publication and are subject to change without notice.