How to Run a Manual Backtest: Trade-by-Trade Testing on Historical Charts

A manual backtest is the purest form of strategy testing: you load a price chart, look at historical bars one by one, and record what your strategy would have done. No code. No formulas. Just you, a chart, a notebook, and your rules. Manual backtesting is slower than automated testing, but it teaches you pattern recognition, forces deep engagement with your strategy, and reveals nuances that automated systems sometimes miss. This walkthrough takes you through a real example so you can run your own manual backtest.

The process is simple in theory: start at a historical date, step forward bar by bar, identify entry signals, record entry price and date, hold the position until your exit rule triggers, record exit price and profit or loss, then repeat for the entire backtest period. In practice, discipline matters. You must resist the temptation to use future information to justify entries and exits, record every trade even the ones you wish you hadn't taken, and track your equity curve to measure drawdown.

Quick definition: A manual backtest is a hands-on process of stepping through historical price data bar by bar, recording entries and exits according to your rules, and calculating profit, loss, and risk metrics without automated software.

Key takeaways

• Manual backtesting requires a chart, your written rules, a notebook, and discipline to record every trade mechanically without hindsight bias.
• You step through bars chronologically, looking backward only to identify support, resistance, moving averages, or other technical levels—never forward.
• For each entry, record the date, price, signal type, and position size; for each exit, record date, price, reason, and resulting profit or loss.
• Track cumulative equity (starting capital plus all profits and losses) to measure drawdown and visualize how your strategy performed over time.
• A manual backtest on 2–3 years of data typically yields 50–200 trades, enough to identify edge but small enough to complete in days or weeks.

What you need to run a manual backtest

A charting platform with historical data. Free platforms like TradingView, Finviz, or MetaTrader provide daily charts back 10+ years. Paid services like Bloomberg Terminal or FactSet offer intraday data and more granular historical depth.

Your rules written down in clear, mechanical language. Not "buy strength," but "buy when the close breaks above the 50-day moving average AND the RSI <70." Every rule must be testable without opinion. If your rule says "buy when sentiment is bullish," you can't backtest it. If it says "buy when the MACD histogram turns positive and the volume is above the 20-day average," you can.

A record-keeping system to track every trade. A notebook, a Google Sheet, or a simple spreadsheet works. You'll record: entry date, entry price, position size, reason for entry, exit date, exit price, profit or loss, and notes. Over a 3-year backtest, you might have 100 trades, so this is manageable by hand but time-consuming.

Discipline to follow your rules without cheating. When you're looking at a chart, it's tempting to use future information: "I can see the price will drop tomorrow, so I'll move my stop-loss up today." That's cheating. A valid manual backtest can only use information available at the time of the trade. Once you decide to exit, don't look ahead to see if you should have held longer.

Step-by-step walkthrough: A simple moving-average breakout strategy

Let's backtest a concrete example. Your strategy: "Buy when the close breaks above the 50-day moving average on above-average volume. Hold until the close drops below the 50-day moving average or a 10% loss, whichever comes first."

Starting capital: $10,000. Position size: $1,000 per trade (10% of capital). Data: Apple stock (AAPL) daily bars from January 1, 2022 to December 31, 2022.

Week 1: January–February 2022 (Setup and first trade)

You load the AAPL daily chart for January 2022. Price is around $180. You calculate the 50-day moving average using the past 50 closes (which means your first data point is November 2021). The 50-day MA is at $176.

January 14, 2022: The close is $180.75, above the 50-day MA ($176). Volume is 40 million shares, above the 20-day average of 38 million. Entry signal triggered.

You record:

• Entry date: January 14, 2022
• Entry price: $180.75
• Position size: $1,000 ÷ $180.75 = 5.53 shares (you'd buy 5 or 6 shares in real trading)
• Reason: Close above 50-day MA on above-average volume
• Stop-loss: 10% below entry = $162.68

You hold. Each day you check: does the close drop below the 50-day MA? Has price fallen >10%?

January 19, 2022: Close is $177.80. The 50-day MA has moved to $177.20 (new closes entered, old closes dropped off). Price is still above the MA.

January 28, 2022: Price has dropped to $165.00. That's >10% below entry ($180.75 × 0.9 = $162.68). Exit signal triggered (stop-loss).

You record:

• Exit date: January 28, 2022
• Exit price: $165.00
• Reason: Stop-loss hit (down 10%)
• Profit/loss: $1,000 entry ÷ $180.75 × $165.00 = $911.35, a loss of $88.65 (or about -8.9% after friction)

Cumulative equity: $10,000 - $88.65 = $9,911.35.

You continue scanning for the next entry signal.

Week 2: March–April 2022 (Second trade, a winner)

February 1, 2022: Close is $181.50, above the 50-day MA ($178.80), volume 35 million (above average). Entry signal.

Entry date: February 1, 2022 Entry price: $181.50 Stop-loss: $163.35 (10% below entry)

You hold.

February 10, 2022: Price has risen to $186. The 50-day MA is $180.20. Still above, holding.

March 1, 2022: Price has risen to $195. The 50-day MA is $183.50. Still above, holding.

March 15, 2022: Close is $192, but the 50-day MA has moved to $191.50. Close dropped just below, so technically the MA close below the moving average rule is triggered. Actually, let me recalculate: 50-day MA at $191.50, close at $192, price is still above. Continue.

Wait—I'm being sloppy. Let me re-read the rule: "hold until the close drops below the 50-day MA or a 10% loss, whichever comes first."

March 20, 2022: Close is $188.80. The 50-day MA is $188.00. Close is still slightly above, hold.

March 22, 2022: Close is $187.50. The 50-day MA is $187.95. Close is now below the 50-day MA. Exit signal triggered.

You record:

• Exit date: March 22, 2022
• Exit price: $187.50
• Reason: Close drops below 50-day MA
• Profit/loss: $1,000 entry ÷ $181.50 × $187.50 = $1,032.88, a profit of $32.88 (or about +3.3%)

Cumulative equity: $9,911.35 + $32.88 = $9,944.23.

Continuing the backtest through 2022

You repeat this process for the entire year. Some trades hit your stop-loss. Some hit the moving-average exit. Some sit for two weeks, others for two months. By December 31, 2022, you've recorded perhaps 12 to 15 trades on AAPL.

Let's say the full-year backtest yields:

• Total trades: 14
• Winning trades: 8
• Losing trades: 6
• Win rate: 8 ÷ 14 = 57%
• Total profit: $1,245
• Total loss: $890
• Net profit: $355 (or 3.5% return on $10,000 starting capital)
• Maximum drawdown: -$1,200 on June 15 (from a peak of $10,900)
• Profit factor: $1,245 ÷ $890 = 1.4

Tracking equity and drawdown

A critical step is plotting your equity curve—a line chart showing your account balance after each trade. Start at $10,000, add or subtract each trade's profit/loss, and chart the running total.

If you've recorded every trade correctly, your equity curve tells you:

• When was the strategy in drawdown (equity dropping)?
• How long did it take to recover?
• What was the worst-case balance (maximum drawdown)?
• Did the strategy make or lose money overall?

In our example above, a peak of $10,900 followed by losses down to $9,700 is a drawdown of $1,200, or about 11%. If that $1,200 drop occurred over three months and recovery took five more months, you know: "This strategy can lose 11% of my capital for an entire season. Can I tolerate that?"

If you're paper-trading (hypothetically) with a $100,000 account, an 11% drawdown is $11,000. You need to know you can handle that without panic-exiting the strategy.

Avoiding common mistakes in manual backtesting

Hindsight bias: You know, on March 22, that the close will be below the 50-day MA, so you exit. But on March 19, you didn't know that. A valid backtest uses only information available at the time. Step through the chart chronologically and resist peeking ahead.

Survivor bias: You might unconsciously skip trades that "obviously" would have lost and focus on the winners. Discipline requires recording every single signal, whether the trade feels good or not.

Unrealistic exits: If your rule says "exit on a stop-loss," don't skip the stop because "the price recovered the next day." The exit signal triggered, so you exited. That's what the backtest is: a record of following your rules.

Sloppy position sizing: If your first trade uses $1,000 and your account grows to $10,900, does your position size now use 10% of $10,900? You must decide upfront how position sizing evolves. The simplest approach is fixed position size (always $1,000) or fixed percentage of account (always 10%). Write it down before you start.

Incomplete record-keeping: Some traders record the profit but not the date, the entry price but not the reason, the exit but not the reasoning. Six months later, you can't remember why the trade happened. Complete records are essential so you can analyze why certain types of trades work or fail.

Why manual backtesting is valuable despite being slow

Automated backtests are faster. But a manual backtest teaches you things code doesn't. You see price action. You notice that your strategy works well in trending months and fails in choppy consolidations. You feel the weight of a three-trade losing streak. You understand, viscerally, that your edge is real but requires patience.

A manual backtest is also a form of paper trading. You're training your brain to follow rules, to resist emotion, to think in probabilities. By the time you paper-trade or go live, you've already internalized your strategy. You know it works. You've seen it work. Your backtest is not just data; it's an education.

Decision tree

Real-world example: Gold futures, 2022

A trader backtested a simple rule on gold futures: "Buy when the 10-day moving average crosses above the 20-day moving average. Sell when the 10-day crosses below the 20-day. Risk $500 per trade."

Testing January–December 2022 on daily gold charts:

• 18 trades total
• 10 winners, 8 losers
• Win rate: 56%
• Total profit on winners: $3,240
• Total loss on losers: $2,100
• Net profit: $1,140
• Drawdown: -$1,050 in March (when the strategy got whipsawed by low volatility)
• Profit factor: 1.54

The trader noted: "The strategy works but struggles in choppy, low-volatility consolidation months. In trending months (January, February, August, November), it's profitable. In sideways months (April, July), it breaks even or loses. I should consider adding a volatility filter or a trend filter to avoid choppy periods."

This insight—"the strategy works in trends but fails in chop"—came from manually stepping through the bars and noticing the pattern. An automated backtest might report the same numbers, but the trader wouldn't understand why.

Summary

Manual backtesting is a hands-on process of stepping through historical price data, recording entries and exits according to your rules, and calculating profit, loss, and drawdown. You need a chart, clear rules, and disciplined record-keeping. Start with 2–3 years of data to generate 50–200 trades. Track cumulative equity to visualize drawdown and recovery. Avoid hindsight bias by moving chronologically and using only information available at the time of each trade. Avoid survivor bias by recording every signal, not just the winners. Manual backtests teach you your strategy deeply, reveal when it works and fails, and prepare you for paper trading and live trading. The process is slow but invaluable.

Next

Ready to speed up your backtest with charting software? Learn Charting Software Backtesting and discover how built-in backtest features streamline testing on visual platforms.