ONE TICKER, ONE RULE, 20+ YEARS
Stock backtesting software to backtest a stock strategy on 20+ years
Name one ticker, say the rule out loud ("buy AAPL when the 50-day crosses above the 200-day, sell when it crosses back"), and the bench tests it on 20+ years of split- and dividend-adjusted daily history against simply holding the same share.
- 1 THESIS
- 2 EVIDENCE
- 3 BACKTEST
- 4 RISK
- 5 VERDICT
02 EVIDENCE · FUNDAMENTALS
04 RISK · IN PLAIN ENGLISH
Past performance does not guarantee future results. Educational analysis only, not financial advice.
In short
Stock backtesting means applying a trading rule to one stock's historical prices to see how the rule would have behaved on that specific name: when it got you in, when it got you out, how much of the move it captured, how deep the drawdown went, and whether it beat simply holding the share. In Agenttrading you type the rule as a sentence, such as "buy AAPL when the 50-day moving average crosses above the 200-day and sell on the reverse cross" or "hold MSFT while it trades above its 200-day average, otherwise cash", confirm the restated rule, and the test runs on 20+ years of split- and dividend-adjusted daily history with a 0.1% cost per trade charged by default. Single-stock backtests carry three risks a fund-level test does not. Sample size: a crossover rule on one ticker may fire only 15 to 30 times in twenty years, which is a thin sample no matter how smooth the equity curve looks. Company-specific events: one product cycle, one lawsuit, or one CEO can dominate the entire result, so the test measures that company's history more than it measures your rule. And selection bias: choosing the ticker after you already know it did well builds the conclusion into the data before the test begins. The academic record on overfitting is worth reading before you trust any single-name result. Bailey, Borwein, Lopez de Prado and Zhu showed in the Notices of the American Mathematical Society (May 2014) that with only five years of daily data, trying more than 45 independent strategy configurations is close to guaranteed to produce an in-sample Sharpe ratio of 1.0 whose expected out-of-sample Sharpe is zero. McLean and Pontiff (Journal of Finance, 2016) tracked 97 published stock-return predictors and found returns were 26% lower out-of-sample and 58% lower after publication. That is why this bench prints every assumption, flags thin samples, and benchmarks against buy-and-hold on the identical share rather than against a broad index that makes any single winner look ordinary. Stock backtesting is included from $19 per month. No trades are executed, nothing is recommended, and the output is educational analysis only: past performance does not guarantee future results.
Past performance does not guarantee future results. For educational and informational purposes only. Not financial advice. Consult a licensed advisor.
WHAT YOU GET - STOCK BACKTESTING
Stock backtesting, run on the bench
Benchmarked against holding the same share
The only fair comparison for a single-stock rule is buying that stock and doing nothing. Beating the S&P while losing to the share you actually traded is the most common way a single-name backtest flatters itself, so the buy-and-hold line for the identical ticker is always drawn.
Thin samples get flagged, not hidden
A crossover rule on one ticker often fires 15 to 30 times across twenty years. When the signal count is too low to support a conclusion, the result says so instead of presenting nine trades as evidence.
Splits and dividends already handled
Raw closing prices invent fake signals at every split and quietly understate every dividend the stock paid. Every test runs on split- and dividend-adjusted daily history, so the rule and the benchmark are measured on the same basis.
The company behind the ticker, not just the chart
Alongside the price rule you get a plain-English fundamentals summary for the name: revenue trend, margins, balance sheet, valuation. A rule that only ever worked because of one product cycle is much easier to spot with both views on screen.
HOW IT WORKS - 4 STEPS
From a sentence to a stamped verdict
Name the stock and state the rule
One sentence: the ticker, the entry condition, the exit condition, and where the money sits when you are out. "Buy AAPL above the 200-day and hold cash below it" is enough detail to test.
Confirm the restated rule
The bench shows the explicit rule it extracted, entry and exit, before running anything. Rephrase until the card matches what you actually meant.
Run it across 20+ years
The rule tests on split- and dividend-adjusted daily history for that ticker, with 0.1% charged per trade, plotted against buying and holding the same share with the worst drawdown window shaded.
Read the sample size, the risks, and the verdict
How many trades fired, how long the worst drawdown lasted, what the fundamentals say about the company, and whether the rule HELD UP, was MIXED, or UNDERPERFORMED. Historical, never a recommendation.
Past performance does not guarantee future results. For educational and informational purposes only. Not financial advice. Consult a licensed advisor.
On the same bench
A single-stock test is one setting of the wider engine described on backtesting software, and the verdicts get stamped by the trading strategy tester. The adjusted daily record behind each run is documented on historical stock data, and the concentration risk a one-ticker rule leaves you holding is explained on investment risk analysis. Because a rule says nothing about whether the company is any good, most people pair it with the fundamental analysis tool. If you have not settled on a platform yet, the best backtesting software roundup weighs ten of them, including the free options your broker may already give you. Testing the same idea across a basket instead of one name belongs on portfolio backtesting or ETF backtesting, and the two strategy families people test most on single stocks are covered on momentum backtesting and dividend backtesting. For the method, read how to backtest a trading strategy and is backtesting accurate; plans start at $19 per month.
QUESTIONS - ASKED AND ANSWERED
Stock backtesting: the common questions
How do you backtest a stock?
Write the rule as one sentence naming the ticker, the entry condition, the exit condition, and what you hold when you are out, then apply it to that stock's split- and dividend-adjusted daily history over 20 years or more, charge a realistic cost per trade, and compare the result against simply buying and holding the same share. In Agenttrading you type that sentence, confirm the restated rule the bench extracted, and read the result with the date range, cost assumption, and trade count printed on it.
What is backtesting in stocks?
Backtesting in stocks is running a buy and sell rule against a stock's past prices to see how the rule would have behaved, so an idea is judged on the historical record rather than on memory or intuition. A backtest worth acting on states the rule before it runs, uses adjusted prices, charges trading costs, covers at least one full market cycle, and reports the result against holding the same stock over the identical window.
What is the best stock backtesting software?
It depends on what you are willing to learn. Code-based tools such as AmiBroker (one-time $299 to $499, Windows only, data feed not included) and QuantConnect give you the most control and expect real programming. Chart-first tools such as TrendSpider (about $82 to $321 per month) and TradingView (free to $239.95 per month) suit traders fluent in indicators. Fidelity offers a genuinely free web Strategy Testing tool for technical equity templates if you hold an account there. Agenttrading ($19 to $129 per month) is the plain-English option: you describe the rule in a sentence rather than building it, and the trade-off is that it tests stated rules rather than optimizing across thousands of parameter sets.
Can you backtest a single stock, or do you need a portfolio?
You can test a single stock, and for most rule ideas that is the right place to start, because it isolates the rule from allocation decisions. What you have to accept is a smaller sample and a heavier dose of company-specific luck. One acquisition, one earnings collapse, or one decade-long product run can carry the entire result. The usual discipline is to test the same rule on several unrelated tickers: if it only works on one, you found that company's history rather than a strategy.
How much historical data do you need to backtest a stock?
At least 20 years where the stock has traded that long. A shorter window can easily miss every bear market, and a rule that has never been tested through the dot-com unwind, 2008, 2020, and 2022 has not really been tested. Length also protects against overfitting: Bailey, Borwein, Lopez de Prado and Zhu showed that with only five years of daily data, trying more than 45 strategy variations is nearly certain to produce an impressive in-sample Sharpe ratio with no out-of-sample edge at all.
How many trades should you backtest?
More than most single-stock backtests actually produce. A rule with fewer than about 30 trades tells you very little, because a couple of lucky entries can dominate the average, and a smooth-looking equity curve built on nine round trips is a sample of nine. If your rule fires rarely, either extend the window, test the same rule across several tickers to pool the evidence, or accept that the result is a hint rather than a finding.
How do you backtest stocks without coding?
Describe the rule in plain English instead of writing it. Traditional stock backtesting expected Python, EasyLanguage, or Pine Script, which is why most investors never ran one. Agenttrading takes the sentence, restates it as an explicit testable rule for you to confirm, then runs it, so the skill required is being clear about what you mean rather than knowing a syntax. Fidelity's free web Strategy Testing tool is another no-code route if you already hold an account there, though it is limited to technical equity templates.
Is stock backtesting accurate?
It accurately describes what a stated rule would have done in history under stated assumptions. It does not predict. The gap between backtest and reality is well measured: McLean and Pontiff tracked 97 published stock-return predictors and found returns 26% lower out-of-sample and 58% lower after publication. Costs, slippage, survivorship, and quiet parameter tuning all push results in the flattering direction, which is why this bench charges costs by default, prints every assumption, and stamps UNDERPERFORMED as prominently as HELD UP. Past performance does not guarantee future results.
Past performance does not guarantee future results. For educational and informational purposes only. Not financial advice. Consult a licensed advisor.
Your next idea deserves a verdict, not a hunch.
Bring a thesis or a ticker. Agenttrading restates the rule, shows the evidence, runs 20+ years of history, and stamps an honest verdict. You decide.
Past performance does not guarantee future results. For educational and informational purposes only. Not financial advice. Consult a licensed advisor.