Backtesting

Use strategy() mode to simulate a trading strategy over historical data. The server executes entries and exits, then returns a performance report.

How It Works

• Declare your script with strategy() instead of indicator()
• Use strategy.entry() and strategy.close() (or strategy.exit()) to simulate trades
• Use --format json to get the full performance report
• Parse the report with jq: .data.report_json | fromjson

Strategy Settings

Common parameters for the strategy() declaration and their defaults:

Parameter	Default	Description
initial_capital	1000000	Starting capital
commission_type	strategy.commission.percent	Commission calculation method
commission_value	0	Commission rate / amount (0 = no commission)
slippage	0	Slippage in ticks per fill
default_qty_type	strategy.fixed	How position size is specified: strategy.fixed (contracts), strategy.percent_of_equity, strategy.cash
default_qty_value	1	Default position size
pyramiding	0	Max simultaneous entries in the same direction (0 = one at a time)
risk_free_rate	2	Annual risk-free rate (%) for Sharpe / Sortino

Example with custom settings:

strategy("My Strategy",
    initial_capital    = 50000,
    commission_type    = strategy.commission.percent,
    commission_value   = 0.1,
    default_qty_type   = strategy.percent_of_equity,
    default_qty_value  = 10)

EMA Crossover Strategy

Buy when EMA8 crosses above EMA21; sell when it crosses below.

longbridge quant run NVDA.US \
  --start 2025-01-01 --end 2026-04-28 \
  --format json \
  --script '
strategy("EMA Cross", overlay=true)
fast = ta.ema(close, 8)
slow = ta.ema(close, 21)
if ta.crossover(fast, slow)
    strategy.entry("Long", strategy.long)
if ta.crossunder(fast, slow)
    strategy.close("Long")
' | jq '.data.report_json | fromjson | .performanceAll'

{
  "netProfit": 4231.00,
  "netProfitPercent": 42.31,
  "grossProfit": 7850.00,
  "grossLoss": 3619.00,
  "profitFactor": 2.17,
  "buyHoldReturnPercent": 31.20,
  "maxDrawdownPercent": -28.15,
  "sharpeRatio": 0.87,
  "sortinoRatio": 1.24,
  "totalClosedTrades": 18,
  "percentProfitable": 50.0,
  "avgWinningTradePercent": 8.72,
  "avgLosingTradePercent": -4.02,
  "commissionPaid": 180.00
}

RSI Mean-Reversion Strategy

Buy when RSI drops below 30 (oversold); exit when RSI recovers above 55.

longbridge quant run AAPL.US \
  --start 2025-01-01 --end 2026-04-28 \
  --format json \
  --script '
strategy("RSI Reversion", overlay=false)
r = ta.rsi(close, 14)
if ta.crossunder(r, 30)
    strategy.entry("Long", strategy.long)
if ta.crossover(r, 55)
    strategy.close("Long")
' | jq '.data.report_json | fromjson | .performanceAll'

Report Reference

Parse the full report object:

longbridge quant run NVDA.US ... --format json --script '...' \
  | jq '.data.report_json | fromjson'

Top-Level Structure

Field	Description
performanceAll	Performance metrics across all trades
performanceLong	Performance metrics for long trades only
performanceShort	Performance metrics for short trades only
closedTrades	Array of completed trade records
openTrades	Array of unrealized positions at backtest end
equityCurve	Per-bar account equity
drawdownCurve	Per-bar drawdown from equity peak
buyHoldCurve	Per-bar buy-and-hold benchmark equity
config	Strategy configuration snapshot

Performance Metrics

performanceAll, performanceLong, and performanceShort share the same shape. Fields marked all only are always 0 / null in performanceLong and performanceShort — they reflect the combined equity curve and cannot be split by direction.

Profit & Loss

Field	Description
netProfit	Net profit in account currency
netProfitPercent	Net profit % of initial capital
grossProfit	Total profit from winning trades
grossProfitPercent	Gross profit %
grossLoss	Total loss from losing trades (positive number)
grossLossPercent	Gross loss %
profitFactor	Gross profit ÷ gross loss
buyHoldReturn	Buy-and-hold return in account currency (all only)
buyHoldReturnPercent	Buy-and-hold return % (all only)

Drawdown & Runup (all only)

Field	Description
maxDrawdown	Largest equity drawdown in account currency
maxDrawdownPercent	Max drawdown %
maxRunup	Largest equity runup in account currency
maxRunupPercent	Max runup %

Risk-Adjusted Returns (all only)

Field	Description
sharpeRatio	Annualized Sharpe ratio
sortinoRatio	Annualized Sortino ratio

Trade Statistics

Field	Description
totalClosedTrades	Completed trades
totalOpenTrades	Unrealized positions at end of backtest
numWinningTrades	Trades with profit > 0
numLosingTrades	Trades with profit < 0
numEvenTrades	Break-even trades (all only)
percentProfitable	Win rate (0–100)

Average Trade

Field	Description
avgTrade	Average P&L per trade
avgTradePercent	Average P&L % per trade
avgWinningTrade	Average profit of winning trades
avgWinningTradePercent	Average winning trade profit %
avgLosingTrade	Average loss of losing trades
avgLosingTradePercent	Average losing trade loss %
ratioAvgWinLoss	Avg winning trade ÷ avg losing trade
largestWinningTrade	Single largest profit
largestWinningTradePercent	Single largest profit %
largestLosingTrade	Single largest loss
largestLosingTradePercent	Single largest loss %

Holding Period

Field	Description
avgBarsInTrades	Average bars held per trade
avgBarsInWinningTrades	Average bars held for winning trades
avgBarsInLosingTrades	Average bars held for losing trades

Other

Field	Description
commissionPaid	Total commissions paid
maxContractsHeld	Peak simultaneous contracts held
marginCalls	Number of margin calls triggered

Trade History

Each entry in closedTrades is a completed round-trip:

Field	Description
tradeNum	Trade number (0-based)
entrySide	"Long" or "Short"
entryId	Entry order ID
entryPrice	Entry fill price
entryTime	Entry timestamp (Unix ms)
exitId	Exit order ID
exitPrice	Exit fill price
exitTime	Exit timestamp (Unix ms)
quantity	Contracts / shares traded
profit	Realized P&L after commission
profitPercent	Realized P&L % relative to entry value
cumulativeProfit	Running total P&L including this trade
cumulativeProfitPercent	Running total P&L % vs. initial capital
maxRunup / maxRunupPercent	Best unrealized gain during the trade
maxDrawdown / maxDrawdownPercent	Worst unrealized loss during the trade
commission	Total commission for this trade (entry + exit)

# Print a trade-by-trade summary
longbridge quant run NVDA.US --start 2025-01-01 --end 2026-04-28 \
  --format json --script '...' \
  | jq -r '.data.report_json | fromjson | .closedTrades[]
    | "#\(.tradeNum) \(.entrySide)  entry=\(.entryPrice)  exit=\(.exitPrice)  P&L=\(.profitPercent)%"'

openTrades uses the same shape minus exit fields, and profit reflects current unrealized P&L.

Equity Curves

Three parallel arrays, one value per bar (index 0 = first bar):

Field	Description
equityCurve	Account equity at bar close
drawdownCurve	Drawdown from equity peak (always ≥ 0)
buyHoldCurve	Hypothetical buy-and-hold equity (benchmark)

# Final equity
jq '.data.report_json | fromjson | .equityCurve[-1]'

# Worst drawdown value
jq '.data.report_json | fromjson | .drawdownCurve | max'

Strategy Config

config captures the declared strategy() settings:

Field	Description
initialCapital	Starting account equity
commissionType	PerContract, PerTrade, or PercentOfValue
commissionValue	Commission amount
slippage	Order slippage in ticks
pyramiding	Max simultaneous entries in the same direction
riskFreeRate	Annual risk-free rate used for Sharpe/Sortino (%)

Table Output (Quick Review)

Without --format json, the table shows each plotted series — useful for visually checking signal timing before running a full backtest:

longbridge quant run NVDA.US \
  --start 2025-01-01 --end 2026-04-28 \
  --script '
strategy("EMA Cross", overlay=true)
fast = ta.ema(close, 8)
slow = ta.ema(close, 21)
plot(fast, "EMA8")
plot(slow, "EMA21")
if ta.crossover(fast, slow)
    strategy.entry("Long", strategy.long)
if ta.crossunder(fast, slow)
    strategy.close("Long")
'