Building a Simple Backtesting Framework in Python

Why Backtest?

Without historical testing, a trading strategy is just a hypothesis. Backtesting lets you stress-test an idea against years of market data before risking real capital.

import pandas as pd

def backtest(df, signal_col, stop_pct=0.01, target_pct=0.02):
    trades = []
    for i, row in df.iterrows():
        if row[signal_col] == 1:
            entry = row['close']
            stop  = entry * (1 - stop_pct)
            tp    = entry * (1 + target_pct)
            trades.append({'entry': entry, 'stop': stop, 'tp': tp})
    return pd.DataFrame(trades)

Avoiding Curve-Fitting

The biggest trap in backtesting is optimising parameters until historical results look perfect. This creates a strategy that works on the past but fails on the future.

• Use out-of-sample data — test on data the strategy never saw during development
• Minimise the number of parameters
• Accept that live performance will always be below backtest performance

If your backtest looks too good, it probably is.