Backtesting Strategies in Pine Script

Backtesting is essential for validating trading strategies. This guide covers everything from basic backtesting to advanced optimization.

---

Why Backtesting Matters

Metric	Importance
Win Rate	Know your success %
Profit Factor	Risk vs reward
Drawdown	Maximum risk
Sharpe Ratio	Risk-adjusted returns
expectancy	Average R per trade

---

Basic Backtest Setup

Simple Strategy Backtest

//@version=6
strategy("Basic Backtest", 
     overlay=true, 
     default_qty_type=strategy.percent_of_equity,
     default_qty_value=10)
 
// Entry
if ta.crossover(ta.sma(close, 20), ta.sma(close, 50))
    strategy.entry("Long", strategy.long)
 
// Exit
if ta.crossunder(ta.sma(close, 20), ta.sma(close, 50))
    strategy.close("Long")
 
plot(ta.sma(close, 20))
plot(ta.sma(close, 50))

Commission Settings

//@version=6
strategy("With Commission",
     commission_type=strategy.commission.percent,
     commission_value=0.1,  // 0.1% per trade
     slippage=3)  // 3 tick slippage
 
if ta.crossover(ta.sma(close, 20), ta.sma(close, 50))
    strategy.entry("Long", strategy.long)
 
if ta.crossunder(ta.sma(close, 20), ta.sma(close, 50))
    strategy.close("Long")

---

Backtest Parameters

Date Range

//@version=6
strategy("Date Range Backtest",
     overlay=true,
     process_orders_on_close=true)
 
// Set date range
startDate = timestamp(2020, 1, 1, 0, 0)
endDate = timestamp(2023, 12, 31, 0, 0)
 
inDateRange = time >= startDate and time <= endDate
 
if inDateRange and ta.crossover(ta.sma(close, 20), ta.sma(close, 50))
    strategy.entry("Long", strategy.long)
 
if inDateRange and ta.crossunder(ta.sma(close, 20), ta.sma(close, 50))
    strategy.close("Long")
 
plot(ta.sma(close, 20))

Instrument Settings

//@version=6
strategy("Instrument Settings",
     overlay=true,
     currency="USD",
     initial_capital=10000,
     default_qty_type=strategy.percent_of_equity,
     default_qty_value=10)
 
if ta.crossover(ta.sma(close, 20), ta.sma(close, 50))
    strategy.entry("Long", strategy.long)

---

Performance Metrics

Custom Metrics

//@version=6
strategy("Metrics", overlay=true)
 
var int wins = 0
var int losses = 0
var float totalProfit = 0
 
// Entry
if ta.crossover(ta.sma(close, 20), ta.sma(close, 50))
    strategy.entry("Long", strategy.long)
 
// Track trades
if strategy.closedtrades > strategy.closedtrades[1]
    if strategy.closedtrades.profit > 0
        wins := wins + 1
    else
        losses := losses + 1
    
    totalProfit := totalProfit + strategy.closedtrades.profit
 
plot(totalProfit)

Win Rate Calculation

//@version=6
strategy("Win Rate", overlay=true)
 
var int totalTrades = 0
var int winningTrades = 0
 
// On trade close
if strategy.closedtrades > strategy.closedtrades[1]
    totalTrades := totalTrades + 1
    
    if strategy.closedtrades.profit > 0
        winningTrades := winningTrades + 1
 
winRate = totalTrades > 0 ? (winningTrades / totalTrades) * 100 : 0

---

Walk Forward Testing

Rolling Window

//@version=6
strategy("Walk Forward", 
     overlay=true,
     process_orders_on_close=true)
 
// Use recent data for parameters
lookback = 200
trainLength = 150
testLength = 50
 
// Simplified: use fixed for demo
maFast = input.int(12, "Fast")
maSlow = input.int(26, "Slow")
 
if ta.crossover(ta.ema(close, maFast), ta.ema(close, maSlow))
    strategy.entry("Long", strategy.long)
 
if ta.crossunder(ta.ema(close, maFast), ta.ema(close, maSlow))
    strategy.close("Long")
 
plot(ta.ema(close, maFast))
plot(ta.ema(close, maSlow))

In-Sample/Out-Sample

//@version=6
strategy("In/Out Sample",
     overlay=true,
     process_orders_on_close=true)
 
// In-sample: 2020-2022
// Out-sample: 2023
inSampleEnd = timestamp(2022, 12, 31)
isInSample = time <= inSampleEnd
 
maFast = input.int(12, "Fast")
maSlow = input.int(26, "Slow")
 
if isInSample and ta.crossover(ta.ema(close, maFast), ta.ema(close, maSlow))
    strategy.entry("Long", strategy.long)
 
if ta.crossunder(ta.ema(close, maFast), ta.ema(close, maSlow))
    strategy.close("Long")

---

Optimization

Single Parameter Optimization

//@version=6
strategy("Optimize MA", 
     overlay=true, 
     optimize=true)
 
fastMA = input.int(12, "Fast MA", 8, 20, 2)
slowMA = input.int(26, "Slow MA", 20, 50, 5)
 
if ta.crossover(ta.ema(close, fastMA), ta.ema(close, slowMA))
    strategy.entry("Long", strategy.long)
 
if ta.crossunder(ta.ema(close, fastMA), ta.ema(close, slowMA))
    strategy.close("Long")
 
plot(ta.ema(close, fastMA))
plot(ta.ema(close, slowMA))

Multi-Parameter Optimization

//@version=6
strategy("Optimize Multi",
     overlay=true,
     optimize=true)
 
fastMA = input.int(12, "Fast", 5, 20, 1)
slowMA = input.int(26, "Slow", 20, 50, 5)
rsiLength = input.int(14, "RSI", 7, 21, 2)
rsiLevel = input.int(30, "RSI Level", 20, 40, 5)
 
rsi = ta.rsi(close, rsiLength)
 
maCross = ta.crossover(ta.ema(close, fastMA), ta.ema(close, slowMA))
 
if maCross and rsi < rsiLevel
    strategy.entry("Long", strategy.long)
 
if ta.crossunder(ta.ema(close, fastMA), ta.ema(close, slowMA))
    strategy.close("Long")

---

Common Backtesting Pitfalls

Look-Ahead Bias

//@version=6
// PROBLEM: Using future data
// BAD:
if close > close[1] and close[2] > close[3]
 
// GOOD: Only use past data
if close > close[1]

Over-Optimization

// PROBLEM: Too many parameters
// BAD:
param1 = input(1)
param2 = input(2)
param3 = input(3)
param4 = input(4)
param5 = input(5)
 
// GOOD: Few parameters
maFast = input.int(12, "Fast")
maSlow = input.int(26, "Slow")

Survivorship Bias

// PROBLEM: Only testing active symbols
// GOOD: Test across multiple symbols
// Note: Pine Script tests one symbol at a time

---

Monte Carlo Simulation

Basic Monte Carlo

//@version=6
strategy("Monte Carlo", overlay=true)
 
// Run multiple times with different seeds
// Note: Pine doesn't support true Monte Carlo
// Use strategy tester for this
 
if ta.crossover(ta.sma(close, 20), ta.sma(close, 50))
    strategy.entry("Long", strategy.long)
 
if ta.crossunder(ta.sma(close, 20), ta.sma(close, 50))
    strategy.close("Long")

---

Drawdown Analysis

Track Drawdown

//@version=6
strategy("Drawdown Tracker", overlay=true)
 
var float maxEquity = strategy.equity
var float drawdown = 0
 
// Update max equity
if strategy.equity > maxEquity
    maxEquity := strategy.equity
 
// Calculate drawdown
drawdown := (maxEquity - strategy.equity) / maxEquity * 100

Max Drawdown

//@version=6
strategy("Max DD", overlay=true)
 
var float maxDrawdown = 0
 
if strategy.equity > 0
    equity = strategy.equity
    peak = math.max(peak, equity)
    dd = (peak - equity) / peak * 100
    maxDrawdown := math.max(maxDrawdown, dd)
 
plot(maxDrawdown, title="Max DD")

---

Complete Backtest Template

//@version=6
strategy("Complete Backtest", 
     overlay=true,
     default_qty_type=strategy.percent_of_equity,
     default_qty_value=10,
     commission_type=strategy.commission.percent,
     commission_value=0.1,
     slippage=3,
     process_orders_on_close=true)
 
// === PARAMETERS ===
maFast = input.int(12, "Fast MA", 8, 20, 2)
maSlow = input.int(26, "Slow MA", 20, 60, 5)
atrLength = input.int(14, "ATR")
atrSL = input.float(2.0, "SL ATR")
atrTP = input.float(3.0, "TP ATR")
 
// === CALCULATIONS ===
fastMA = ta.ema(close, maFast)
slowMA = ta.ema(close, maSlow)
atr = ta.atr(atrLength)
 
// === SIGNALS ===
longSignal = ta.crossover(fastMA, slowMA)
shortSignal = ta.crossunder(fastMA, slowMA)
 
// === ENTRIES ===
if longSignal
    strategy.entry("Long", strategy.long)
 
if shortSignal
    strategy.entry("Short", strategy.short)
 
// === EXITS ===
if strategy.position_size > 0
    strategy.exit("Long Exit", "Long", 
         stop=strategy.position_avg_price - atr * atrSL,
         limit=strategy.position_avg_price + atr * atrTP)
 
if strategy.position_size < 0
    strategy.exit("Short Exit", "Short",
         stop=strategy.position_avg_price + atr * atrSL,
         limit=strategy.position_avg_price - atr * atrTP)
 
// === PLOT ===
plot(fastMA, color=color.green)
plot(slowMA, color=color.red)

---

Backtest Checklist

• Clear entry/exit rules
• Commission included
• Slippage considered
• Realistic position sizing
• Date range specified
• Multiple market conditions tested
• Not over-optimized
• Out-of-sample validation

---

Metrics to Review

Metric	Good	Bad
Profit Factor	> 1.5	< 1.0
Win Rate	> 40%	< 35%
Sharpe Ratio	> 1.0	< 0.5
Max Drawdown	< 20%	> 50%
Expectancy	> 0.5R	< 0

---

Next Steps

• timeframes - Multi-timeframe analysis
• custom-functions - Advanced coding
• arrays - Arrays and data structures

---

Backtest thoroughly, trade cautiously.