Backtesting Your Strategy: From Idea to Profitable System

90% of traders skip backtesting and wonder why they lose money. Test your strategy on historical data before risking real capital. Here's how to do it properly.

What is Backtesting?

Backtesting is applying your trading rules to historical price data to see if the strategy would have been profitable. It answers: "Does my edge actually exist?"

• Manual Backtesting: Scroll through charts, mark trades, record results (best for beginners)
• Automated Backtesting: Code your strategy, run it on data (requires programming)
• Forward Testing: Test on recent data after backtesting (paper trading)

Why Most Traders Skip This

The Manual Backtesting Process

You don't need coding skills. Here's the step-by-step manual process:

Step 1: Define Your Strategy Rules

Write down EXACT entry and exit rules. No discretion allowed.

Step 2: Set Up Your Spreadsheet

Create columns to track every trade:

Trade #	Date	Ticker	Entry	Exit	% Gain/Loss	R Multiple
1	2025-01-15	AAPL	$180	$207	+15%	+2.14R
2	2025-01-22	TSLA	$245	$228	-7%	-1R

Step 3: Scroll Through Historical Charts

Go back 2-3 years. Mark every trade that meets your rules. Be honest - no cherry-picking!

Step 4: Record Every Trade

Log all trades in your spreadsheet. Winners AND losers. Track:

• Entry price and date
• Exit price and date (stop or target)
• % gain or loss
• R multiple (how many times your risk did you make/lose)
• Hold time in days

Key Metrics to Calculate

After 100+ trades, calculate these metrics to evaluate your strategy:

Metric	Formula	Good Target
Win Rate	Wins / Total Trades	> 50%
Average Win	Sum of wins / # of wins	> 8%
Average Loss	Sum of losses / # of losses	< 5%
Expectancy	(Win% × Avg Win) - (Loss% × Avg Loss)	> 0.5%
Profit Factor	Gross Profit / Gross Loss	> 1.5
Max Drawdown	Largest peak-to-trough decline	< 20%
Average Hold Time	Sum of hold days / # trades	Depends on style

Real Backtest Example: Breakout Strategy

Here's a real 100-trade backtest of the breakout strategy defined above:

Interpreting the Results

Verdict: Profitable strategy with positive expectancy.

• 58% win rate is solid (above 50%)
• Average win (12.3%) is 2x average loss (6.2%) = good R:R
• Expectancy of +4.5% per trade means long-term profitability
• Profit factor of 2.73 means you make $2.73 for every $1 lost
• Max drawdown of 18.5% is manageable with proper position sizing

Common Backtesting Mistakes

Mistake	Why It's Wrong	Fix
Looking ahead	You see the future, invalidates test	Use bar replay or cover chart
Cherry-picking trades	Only counting winners inflates results	Record EVERY signal, no exceptions
Too few trades	20 trades proves nothing statistically	Minimum 100 trades for significance
Ignoring slippage	Real trading has costs	Subtract 0.1-0.3% per trade
Curve fitting	Optimizing for past = fails in future	Keep rules simple, test out-of-sample
Survivorship bias	Only testing stocks that survived	Include delisted/bankrupt stocks

Expectancy: The Most Important Metric

Expectancy tells you how much you make per trade on average. It's the ONLY metric that matters for long-term profitability.

What Different Expectancies Mean

• Negative expectancy: You lose money long-term. Don't trade this strategy.
• 0-1% expectancy: Barely profitable. Commissions might wipe you out.
• 1-3% expectancy: Decent edge. Can be profitable with discipline.
• 3-5% expectancy: Strong edge. Very profitable if executed properly.
• 5%+ expectancy: Exceptional edge. Rare. Verify it's not curve-fitted.

Forward Testing

After backtesting, forward test on recent data (last 3-6 months) that you didn't backtest on:

1. Paper trade: Apply your rules in real-time with fake money
2. Track results: Same spreadsheet, same metrics
3. Compare to backtest: Results should be similar (within 20%)
4. If forward test fails: Your backtest was curve-fitted. Start over.

When to Start Real Money Trading

Sample Backtest Spreadsheet Template

Here's what your spreadsheet should look like:

Advanced: Monte Carlo Simulation

After backtesting, run Monte Carlo simulation to understand worst-case scenarios:

1. Take your 100 trade results
2. Randomly shuffle them 1000 times
3. Calculate max drawdown for each shuffle
4. The 95th percentile is your "worst realistic case"

This shows you: "If I'm unlucky, how bad could it get?" Essential for position sizing.

Key Takeaways

• Backtest minimum 100 trades before risking real money
• Write exact rules - no discretion allowed during backtest
• Don't look ahead or cherry-pick trades - invalidates results
• Expectancy is the only metric that matters for profitability
• Forward test on recent data to verify backtest wasn't curve-fitted
• Start real money only after successful backtest + forward test
• Track same metrics in live trading - should match backtest within 20%