构建K线图趋势约束模型（第九部分）：多策略EA（第一部分）（基础篇）
「多策略EA的框架起点」
这一节是系列第九部分的开篇，定位在把七种已知策略塞进同一个EA骨架里跑，而不是新写一套信号逻辑。原文列了总览、魔法数字引入、趋势跟踪落地、回测与结论五个块，说明工程重点是‘多策略共存’而非‘单策略优化’。 魔法数字（magic number）在这里是关键隔离手段：不同子策略的下单用不同 magic 值标记，便于MT5后台按策略统计胜率与回撤，也避免平仓时误杀其他策略的持仓。对外汇与贵金属来说，多策略同账户运行会放大隔夜利息与滑点风险，实盘前必须用策略测试器分 magic 跑一遍。 作者给出的基础数据点：该文发布于2025年6月16日，原文交互量820次、评论0条，说明这套多策略框架当时还未经过社区充分质疑，直接照搬需自行验证。
日线定方向，小周期找切入点
做趋势约束型 EA，第一性原则是把仓位方向锁在 D1 日 K 线的主流情绪上。日线收阳（close > open）视为多头日，收阴（close < open）视为空头日，这是上层约束；真正下单信号则下沉到 M1 这类小周期，等 RSI 撞到极端值再触发，初版逻辑刻意不做复杂堆叠，方便手动跟代码。 回头看已发布的约束片段，判断函数非常直白：用 iOpen 和 iClose 抓当前日线的开收价，比大小即得 bool 标志。下面这段就是日线多空判定的最小可用核，建议直接开 MT5 新建脚本粘进去跑一遍验证。
class="type">void OnTick() { class=class="str">"cmt">// Determine current daily trend(bullish ) class="type">class="kw">double daily_open = iOpen(_Symbol, PERIOD_D1, class="num">0); class="type">class="kw">double daily_close = iClose(_Symbol, PERIOD_D1, class="num">0); class="type">bool is_bullish = daily_close > daily_open; } class="type">void OnTick() { class=class="str">"cmt">// Determine current daily trend(bearish) class="type">class="kw">double daily_open = iOpen(_Symbol, PERIOD_D1, class="num">0); class="type">class="kw">double daily_close = iClose(_Symbol, PERIOD_D1, class="num">0); class="type">bool is_bearish = daily_close < daily_open; } if (is_bullish) { class=class="str">"cmt">// Logic for bullish trend Print("The daily trend is bullish."); class=class="str">"cmt">// You can add your trading logic here, for example: class=class="str">"cmt">// if (OrderSelect(...)) { class=class="str">"cmt">/* Place buy order */ } } if (is_bearish) { class=class="str">"cmt">// Logic for bearish trend Print("The daily trend is bearish."); class=class="str">"cmt">// You can add your trading logic here, for example: class=class="str">"cmt">// if (OrderSelect(...)) { class=class="str">"cmt">/* Place sell order */ } } class="type">long position_type = PositionGetInteger(POSITION_TYPE);
class="type">void OnTick() { class=class="str">"cmt">// Determine current daily trend(bullish ) class="type">class="kw">double daily_open = iOpen(_Symbol, PERIOD_D1, class="num">0); class="type">class="kw">double daily_close = iClose(_Symbol, PERIOD_D1, class="num">0); class="type">bool is_bullish = daily_close > daily_open; } class="type">void OnTick() { class=class="str">"cmt">// Determine current daily trend(bearish) class="type">class="kw">double daily_open = iOpen(_Symbol, PERIOD_D1, class="num">0); class="type">class="kw">double daily_close = iClose(_Symbol, PERIOD_D1, class="num">0); class="type">bool is_bearish = daily_close < daily_open; } if (is_bullish) { class=class="str">"cmt">// Logic for bullish trend Print("The daily trend is bullish."); class=class="str">"cmt">// You can add your trading logic here, for example: class=class="str">"cmt">// if (OrderSelect(...)) { class=class="str">"cmt">/* Place buy order */ } } if (is_bearish) { class=class="str">"cmt">// Logic for bearish trend Print("The daily trend is bearish."); class=class="str">"cmt">// You can add your trading logic here, for example: class=class="str">"cmt">// if (OrderSelect(...)) { class=class="str">"cmt">/* Place sell order */ } } class="type">long position_type = PositionGetInteger(POSITION_TYPE);
◍ 把参数和句柄先钉死
EA 跑起来之前，先把外部可调参数摊开：超买 70、超卖 30 是 RSI 的经典边界，手数 0.1 属于轻仓试错，止损 100 点、止盈 200 点构成 1:2 的风险回报比，追踪止损 50 点用于在趋势延续时锁利。外汇与贵金属波动剧烈，这类固定点数止损在跳空时可能直接穿损，实盘前须在 MT5 策略测试器按品种点值重算。 初始化阶段只做一件事：用 iRSI 给当前品种、当前周期、收盘价算出的 RSI 建句柄。句柄若返回 INVALID_HANDLE，说明指标创建失败，OnInit 直接返回 INIT_FAILED 阻止 EA 挂载；正常则返回 INIT_SUCCEEDED。 退出时别漏了 IndicatorRelease 释放句柄，否则多次加载脚本会堆积内存占用。OnTick 里先抓日线开盘价与收盘价判断当日方向——收高于开视为偏多，收低于开视为偏空，再 CopyBuffer 取最新一根 RSI 值，失败就打印并 return，不让脏数据触发下单。
class="kw">input class="type">class="kw">double RSI_Overbought = class="num">70.0; class=class="str">"cmt">// RSI overbought level class="kw">input class="type">class="kw">double RSI_Oversold = class="num">30.0; class=class="str">"cmt">// RSI oversold level class="kw">input class="type">class="kw">double Lots = class="num">0.1; class=class="str">"cmt">// Lot size class="kw">input class="type">class="kw">double StopLoss = class="num">100; class=class="str">"cmt">// Stop Loss in points class="kw">input class="type">class="kw">double TakeProfit = class="num">200; class=class="str">"cmt">// Take Profit in points class="kw">input class="type">class="kw">double TrailingStop = class="num">50; class=class="str">"cmt">// Trailing Stop in points class=class="str">"cmt">// Global variables class="type">class="kw">double rsi_value; class="type">int rsi_handle; CTrade trade; class=class="str">"cmt">// Declare an instance of the CTrade class class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| Expert initialization function | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">int OnInit() { class=class="str">"cmt">// Create an RSI indicator handle rsi_handle = iRSI(_Symbol, PERIOD_CURRENT, RSI_Period, PRICE_CLOSE); if (rsi_handle == INVALID_HANDLE) { Print("Failed to create RSI indicator handle"); class="kw">return(INIT_FAILED); } class="kw">return(INIT_SUCCEEDED); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| Expert deinitialization function | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void OnDeinit(class="kw">const class="type">int reason) { class=class="str">"cmt">// Release the RSI indicator handle IndicatorRelease(rsi_handle); } class=class="str">"cmt">//+------------------------------------------------------------------+ class=class="str">"cmt">//| Expert tick function | class=class="str">"cmt">//+------------------------------------------------------------------+ class="type">void OnTick() { class=class="str">"cmt">// Determine current daily trend(bullish or bearish) class="type">class="kw">double daily_open = iOpen(_Symbol, PERIOD_D1, class="num">0); class="type">class="kw">double daily_close = iClose(_Symbol, PERIOD_D1, class="num">0); class="type">bool is_bullish = daily_close > daily_open; class="type">bool is_bearish = daily_close < daily_open; class=class="str">"cmt">// Get the RSI value for the current bar class="type">class="kw">double rsi_values[]; if (CopyBuffer(rsi_handle, class="num">0, class="num">0, class="num">1, rsi_values) <= class="num">0) { Print("Failed to get RSI value"); class="kw">return; } rsi_value = rsi_values[class="num">0]; class=class="str">"cmt">// Close open positions if the trend changes
「逆势持仓清场与信号开仓的循环逻辑」
这段逻辑解决一个实盘常见坑：当趋势判定翻转（如 bullish 转 bearish）而账户里还躺着反向单，必须先按 ticket 精准平仓，而不是靠 PositionClose(_Symbol) 偷懒——后者在同源多单时会漏单。 循环从 PositionsTotal()-1 递减到 0，用 PositionSelect(PositionGetSymbol(i)) 定位第 i 个持仓，再读 POSITION_TYPE 与 POSITION_TICKET。若持仓是多单且当前 is_bearish，或空单且 is_bullish，直接 trade.PositionClose(ticket) 平掉；外汇与贵金属杠杆高，逆势单过夜可能扩大浮亏，这种清理要快。 开仓侧只在 PositionsTotal()==0 时动手：bullish 且 rsi_value < RSI_Oversold 读 SYMBOL_ASK 开多，sl=price-StopLoss*_Point、tp=price+TakeProfit*_Point；bearish 且 rsi_value > RSI_Overbought 读 SYMBOL_BID 开空，方向相反。把 StopLoss / TakeProfit 调成 500 / 1000 这类点数，回测里能直观看到盈亏比变化。 尾部还有第二个同构循环做 trailing stop：重新遍历持仓，取 POSITION_PRICE_OPEN 与 POSITION_SL，按 POSITION_TYPE 分支处理。复制下面代码到 MT5 策略测试器，把 RSI 周期设 14、超买超卖 70/30，跑 EURUSD 小时图大概率能看到逆势单被扫掉后再开新单的节奏。
for (class="type">int i = PositionsTotal() - class="num">1; i >= class="num">0; i--) { if (PositionSelect(PositionGetSymbol(i))) class=class="str">"cmt">// Corrected usage { class="type">long position_type = PositionGetInteger(POSITION_TYPE); class="type">class="kw">ulong ticket = PositionGetInteger(POSITION_TICKET); class=class="str">"cmt">// Get the position ticket if ((position_type == POSITION_TYPE_BUY && is_bearish) || (position_type == POSITION_TYPE_SELL && is_bullish)) { trade.PositionClose(ticket); class=class="str">"cmt">// Use the class="type">class="kw">ulong variable directly } } } class=class="str">"cmt">// Check for buy condition(bullish trend + RSI oversold) if (is_bullish && rsi_value < RSI_Oversold) { class=class="str">"cmt">// No open positions? Place a buy order if (PositionsTotal() == class="num">0) { class="type">class="kw">double price = SymbolInfoDouble(_Symbol, SYMBOL_ASK); class="type">class="kw">double sl = price - StopLoss * _Point; class="type">class="kw">double tp = price + TakeProfit * _Point; class=class="str">"cmt">// Open a buy order trade.Buy(Lots, _Symbol, price, sl, tp, "TrendConstraintExpert Buy"); } } class=class="str">"cmt">// Check for sell condition(bearish trend + RSI overbought) if (is_bearish && rsi_value > RSI_Overbought) { class=class="str">"cmt">// No open positions? Place a sell order if (PositionsTotal() == class="num">0) { class="type">class="kw">double price = SymbolInfoDouble(_Symbol, SYMBOL_BID); class="type">class="kw">double sl = price + StopLoss * _Point; class="type">class="kw">double tp = price - TakeProfit * _Point; class=class="str">"cmt">// Open a sell order trade.Sell(Lots, _Symbol, price, sl, tp, "TrendConstraintExpert Sell"); } } class=class="str">"cmt">// Apply trailing stop for (class="type">int i = PositionsTotal() - class="num">1; i >= class="num">0; i--) { if (PositionSelect(PositionGetSymbol(i))) class=class="str">"cmt">// Corrected usage { class="type">class="kw">double price = PositionGetDouble(POSITION_PRICE_OPEN); class="type">class="kw">double stopLoss = PositionGetDouble(POSITION_SL); class="type">class="kw">double current_price; if (PositionGetInteger(POSITION_TYPE) == POSITION_TYPE_BUY) {
多单与空单的跟进止损分支
上面这段逻辑把双向持仓的 trailing stop 拆成了两个独立判断。多头用 BID 价比对开仓价，空头用 ASK 价，方向不同取价方式不能反，否则回测里会出现止损永远不触发或者瞬间被扫的怪象。 以多头为例：当 current_price 与记录价 price 的差额超过 TrailingStop 个 _Point 时，才进入修改判断。这里 price 通常是持仓期间某次记录的最低/最高价，TrailingStop 是交易者自己设的步长，比如 30 点。 修改条件要求新止损位比原 stopLoss 更优（多头时原止损更小），才调用 trade.PositionModify 把止损挪到 current_price - TrailingStop*_Point，止盈保持不变。空头分支镜像处理，用 current_price + TrailingStop*_Point，且允许 stopLoss==0 时直接初始化跟进。 外汇与贵金属杠杆高，这类自动改止损在滑点行情中可能成交在预期外价位，建议先在 MT5 策略测试器用 2023 年 XAUUSD 的 M1 数据跑一遍，观察 TrailingStop 从 15 调到 50 点时的回撤变化。
current_price = SymbolInfoDouble(_Symbol, SYMBOL_BID); if (current_price - price > TrailingStop * _Point) { if (stopLoss < current_price - TrailingStop * _Point) { trade.PositionModify(PositionGetInteger(POSITION_TICKET), current_price - TrailingStop * _Point, PositionGetDouble(POSITION_TP)); } } } else if (PositionGetInteger(POSITION_TYPE) == POSITION_TYPE_SELL) { current_price = SymbolInfoDouble(_Symbol, SYMBOL_ASK); if (price - current_price > TrailingStop * _Point) { if (stopLoss > current_price + TrailingStop * _Point || stopLoss == class="num">0) { trade.PositionModify(PositionGetInteger(POSITION_TICKET), current_price + TrailingStop * _Point, PositionGetDouble(POSITION_TP)); } } } }
◍ 把数学原理落成可跑的算法
外汇与贵金属交易天然带高杠杆与双向波动风险，任何策略都只是提高概率，而非消除亏损可能。 文中列出的七类策略并非凭空而来，它们各自锚定了某位经济学家或实战派交易者提出的数学框架——比如均值回归对应正态分布假设，突破类则依赖波动率聚类现象。先把原理吃透，后续写代码才不会只抄个形态外壳。 真正分水岭在第二步：当你能口述清楚某策略的进出场条件，再用 MQL5 把它写成 EA，验证成本就只剩电费。很多人口中的「策略懂了」其实停在直觉层，一碰到滑点和点差重写就露怯。 所以这一节的价值不在罗列名字，而是提醒你——原理记熟后，编码实现才是把主观经验变成可回测对象的唯一路径，打开 MT5 新建一个空 EA 试着落一条规则，比反复读图文有用。