Pine Script中变量计算止盈止损时strategy.exit失效问题

问题现象 #

• 实现市价入场+限价止盈止损的括号订单，变量Long1_Take_Profit和Long1SL的计算结果在图表标签、webhook输出均正确，通过IB_Insync机器人能在TWS正常提交订单
• 但将这两个变量传入strategy.exit时，止盈止损完全不触发
• 仅当手动输入固定价格、直接写计算式（如close*1.05）或手动赋值变量（如TP=17585+7）时，strategy.exit才能正常工作
• 策略逻辑：止损设为前3根K线最低价ta.lowest(3)，止盈为1:1风险比，该逻辑在V3/V4中可实现，V5下失效

核心原因 #

问题出在变量赋值时机与strategy.exit执行时机不匹配：

1. 触发long1信号执行strategy.entry时，当前K线的strategy.position_size还未更新为持仓状态，导致Long1SL和Long1_Take_Profit此时仍为初始值0.0
2. strategy.opentrades.entry_price(strategy.opentrades - 1)在开仓瞬间无法获取正确入场价格，因为订单还未完成成交确认

解决方案 #

1. 提前计算止损价格：在开仓信号触发时直接使用当前可用的ta.lowest(3)，无需等待持仓状态更新
2. 同步计算止盈止损：将止损止盈的计算逻辑与strategy.exit放在同一代码块，确保开仓时能传递正确的价格参数
3. 使用当前K线价格作为基准：开仓时用close（或根据订单类型调整为open）作为入场基准计算止盈，保证参数有效性

修改后的代码 #

//@version=5
strategy("PPO Strat Entry Test",  overlay=true, default_qty_value=1, calc_on_order_fills=false, calc_on_every_tick=true, pyramiding=0)

//==============================================DATE FILTER===========================================================================================================================================

useDateFilter = input.bool(true, title='Filter Date Range of Backtest', group='Backtest Time Period')
backtestStartDate = input.time(title='Start Date', defval=timestamp('1 Apr 2024'), group='Backtest Time Period', tooltip='This start date is in the time zone of the exchange where the chart\'s instrument trades. It doesn\'t use the time zone of the chart or of your computer.')
backtestEndDate = input.time(title='End Date', defval=timestamp('30 Apr 2024'), group='Backtest Time Period', tooltip='This end date is in the time zone of the exchange where the chart\'s instrument trades. It doesn\'t use the time zone of the chart or of your computer.')

inDateRange = not useDateFilter or time >= backtestStartDate and time < backtestEndDate

//________________________________________________________________________________________________________________________________________________________________________
//              EXPONENTIAL MOVING AVERAGE

EMAlength = input.int(20, minval=1)
EMAexp = input(true, 'exponential')

esma(realClose, EMAlength)=>
    s = ta.sma(realClose, EMAlength)
    e = ta.ema(realClose, EMAlength)
    EMAexp ? e : s

//________________________________________________________________________________________________________________________________________________________________________
//              PPO

//PPO Chart Time
fast_length1 = input(title='Fast Length', defval=7, group = 'PPO')
slow_length1 = input(title='Slow Length', defval=21, group = 'PPO')

signal_length = input.int(title='Signal Smoothing', minval=1, maxval=50, defval=5, group = 'PPO')
sma_source = input(title='Simple MA(Oscillator)', defval=true, group = 'PPO')
sma_signal = input(title='Simple MA(Signal Line)', defval=true, group = 'PPO')
PPOsrc = close

// Calculating
fast_ma = sma_source ? ta.sma(PPOsrc, fast_length1) : ta.ema(PPOsrc, fast_length1)
slow_ma = sma_source ? ta.sma(PPOsrc, slow_length1) : ta.ema(PPOsrc, slow_length1)
PPO = ( ((fast_ma - slow_ma) / slow_ma) * 100 )
PPOsig = (sma_signal ? ta.sma(PPO, signal_length) : ta.ema(PPO, signal_length))

//Price Oscillator Chart Time
OSC_shortlen=input.int(7, "Short Length", minval=1, group = 'Price Osc')
OSC_longlen=input.int(21, "Long Length", minval=1, group = 'Price Osc')
OSC_exp = input(true, "exponential", group = 'Price Osc')

OSCshort = esma(PPOsrc, OSC_shortlen)
OSClong = esma(PPOsrc, OSC_longlen)
PPOosc = ( ((OSCshort - OSClong)/OSClong*100) )

long1 = ta.crossunder(PPOsig, PPO)

//ORDER EXECUTION
if inDateRange and long1 and barstate.isconfirmed and strategy.position_size == 0
    // 开仓时直接计算止损价格（前3根K线最低价）
    current_sl = math.round_to_mintick(ta.lowest(3))
    // 用当前bar收盘价作为入场基准，计算1:1风险比的止盈价格
    entry_base = close
    current_tp = math.round_to_mintick(entry_base + (entry_base - current_sl))
    
    // 提交入场订单
    strategy.entry('Long', strategy.long, qty = 1, comment = "Long1\nBrkt Entry", oca_name = "Long1")
    // 提交止损止盈订单，直接传入刚计算的价格
    strategy.exit('Exit Long', from_entry = 'Long', limit = current_tp, stop = current_sl, 
                  oca_name = "Long1_Exit", comment_profit = "Profit Bkt\nClose Long1", comment_loss = "SL Bkt\nClose Long1")
    
    // 绘制测试标签验证价格
    label.new(bar_index, high, str.tostring(current_tp), yloc=yloc.abovebar, textcolor=color.white, size=size.large, style=label.style_label_left)
    label.new(bar_index, low, str.tostring(current_sl), yloc=yloc.belowbar, textcolor=color.white, size=size.large, style=label.style_label_left)

额外说明 #

• 如果需要更精准的入场价格，可以开启calc_on_order_fills=true，通过订单成交事件更新止损止盈价格，但会增加代码复杂度
• 确保math.round_to_mintick正确处理价格，符合品种最小报价单位要求
• 回测时保留barstate.isconfirmed，避免tick级信号导致的过度交易

内容的提问来源于stack exchange，提问作者Trent