在 GitHub 上查看
Exp AFIRMA 策略
概述
Exp AFIRMA 策略在 StockSharp 平台上重现了 MetaTrader 顾问 Exp_AFIRMA.mq5。核心为 AFIRMA 指标(Adaptive Finite
Impulse Response Moving Average),该指标先用带窗函数的 FIR 滤波器平滑价格,再利用 ARMA 预测未来方向。本移植版完全
保留原始逻辑:当 ARMA 预测线由下转上时开多,由上转下时开空。所有计算基于选定时间框架的已完成 K 线(默认 4 小时)。
策略在每根蜡烛收盘后记录最新的 ARMA 值,并检查三个连续收盘的斜率。当出现反向信号时,先撤销当前方向的持仓,再建立
新的反向仓位。可选的止损/止盈通过 StockSharp 的 StartProtection 机制实现。
交易逻辑
- 指标计算
- 自定义的
AfirmaIndicator 复刻了 AFIRMA 算法。FIR 阶段使用 Taps 个系数,按所选窗口函数(矩形、两种 Hanning、
Blackman、Blackman-Harris)加权,参数 Periods 控制频带宽度(与原 MQL 输入一致)。
- 预测阶段采用与原始代码相同的最小二乘公式(系数
sx2…sx6),输出 FIR 与 ARMA 两条曲线供图形显示。
- 信号判定
- 每根 K 线收盘时保存新的 ARMA 值。
SignalBar 指定跳过的收盘数,默认值 1 表示分析 t-1、t-2、t-3 三根已收盘 K
线并在第 t 根开盘执行。
- 多头条件:
ARMA[t-2] < ARMA[t-3] 且 ARMA[t-1] > ARMA[t-2]。此时允许平空并根据 TradeVolume 建立/加仓多头。
- 空头条件:
ARMA[t-2] > ARMA[t-3] 且 ARMA[t-1] < ARMA[t-2]。此时允许平多并建立/加仓空头。
- 仓位管理
- 策略仅维持一个净头寸。触发信号后,仓位目标调整到
±TradeVolume。
- 若启用
StopLossPoints 或 TakeProfitPoints,在 OnStarted 中调用 StartProtection,以价格单位设置保护单。
- 开仓前会取消所有挂单,避免旧订单与新信号冲突。
参数
| 参数 |
说明 |
TradeVolume |
基础交易量,决定多空目标仓位。 |
CandleType |
指标计算所用的蜡烛类型/时间框架。 |
Periods |
FIR 滤波器的带宽参数(等同于原顾问的 1/(2*Periods) 设置)。 |
Taps |
FIR 系数数量,内部会自动调整为奇数。 |
Window |
FIR 的窗函数:Rectangular、Hanning1、Hanning2、Blackman、BlackmanHarris。 |
SignalBar |
用于确认信号的历史收盘数量;1 代表最近已收盘的一根 K 线。 |
EnableBuyEntries / EnableSellEntries |
是否允许开多/开空。 |
EnableBuyExits / EnableSellExits |
是否允许自动平多/平空。 |
StopLossPoints |
止损距离,使用价格单位。 |
TakeProfitPoints |
止盈距离,使用价格单位。 |
转换说明
- 原策略中的资金管理选项(
MM、MMMode、Deviation_)未迁移,改为单一的 TradeVolume。可通过账户或外部模块实现更
灵活的仓位控制与滑点处理。
- MQL 版本以“点”为单位发送止损/止盈,这里直接使用价格差值,方便在不同品种之间复用。如需与点数匹配,请自行乘以价格步长。
- 当
SignalBar = 1 时,策略读取最近三根已完成蜡烛的 ARMA 值,并在下一根蜡烛开盘下单,与原 CopyBuffer 调用保持一致。
将 SignalBar 设为 0 也可以使用,但由于计算在收盘后进行,因此仍然基于最新收盘价。
AfirmaIndicator 完整实现了 FIR+ARMA 的数学模型,可与 DrawIndicator 结合,在图表上同时显示 FIR 与 ARMA 曲线。
使用建议
- 绑定所需的证券与投资组合,设置
TradeVolume 并选择合适的 CandleType。
- 根据策略方向需求调整多空开仓和平仓开关。
- 若需要固定的止损/止盈,设置
StopLossPoints 与 TakeProfitPoints;留为 0 时策略仅依赖反向信号退出。
- 在调整
Periods、Taps、SignalBar 等参数时配合图表观察 FIR/ARMA 线的走势以及成交记录,确保行为与预期一致。
using System;
using System.Collections.Generic;
using Ecng.Common;
using StockSharp.Algo.Indicators;
using StockSharp.Algo.Strategies;
using StockSharp.BusinessEntities;
using StockSharp.Messages;
namespace StockSharp.Samples.Strategies;
/// <summary>
/// Exp AFIRMA strategy using EMA crossover as adaptive filter approximation.
/// Buys when fast EMA crosses above slow EMA, sells on reverse.
/// </summary>
public class ExpAfirmaStrategy : Strategy
{
private readonly StrategyParam<int> _fastPeriod;
private readonly StrategyParam<int> _slowPeriod;
private readonly StrategyParam<int> _stopLossPoints;
private readonly StrategyParam<int> _takeProfitPoints;
private ExponentialMovingAverage _fast;
private ExponentialMovingAverage _slow;
private decimal _prevFast;
private decimal _prevSlow;
private decimal _entryPrice;
private int _cooldown;
/// <summary>
/// Fast EMA period.
/// </summary>
public int FastPeriod
{
get => _fastPeriod.Value;
set => _fastPeriod.Value = value;
}
/// <summary>
/// Slow EMA period.
/// </summary>
public int SlowPeriod
{
get => _slowPeriod.Value;
set => _slowPeriod.Value = value;
}
/// <summary>
/// Stop-loss distance in price steps.
/// </summary>
public int StopLossPoints
{
get => _stopLossPoints.Value;
set => _stopLossPoints.Value = value;
}
/// <summary>
/// Take-profit distance in price steps.
/// </summary>
public int TakeProfitPoints
{
get => _takeProfitPoints.Value;
set => _takeProfitPoints.Value = value;
}
/// <summary>
/// Initializes a new instance of the <see cref="ExpAfirmaStrategy"/> class.
/// </summary>
public ExpAfirmaStrategy()
{
_fastPeriod = Param(nameof(FastPeriod), 21)
.SetGreaterThanZero()
.SetDisplay("Fast Period", "Fast EMA period", "Indicator");
_slowPeriod = Param(nameof(SlowPeriod), 100)
.SetGreaterThanZero()
.SetDisplay("Slow Period", "Slow EMA period", "Indicator");
_stopLossPoints = Param(nameof(StopLossPoints), 200)
.SetNotNegative()
.SetDisplay("Stop Loss", "Stop-loss in price steps", "Risk");
_takeProfitPoints = Param(nameof(TakeProfitPoints), 400)
.SetNotNegative()
.SetDisplay("Take Profit", "Take-profit in price steps", "Risk");
}
/// <inheritdoc />
public override IEnumerable<(Security sec, DataType dt)> GetWorkingSecurities()
{
yield return (Security, TimeSpan.FromMinutes(5).TimeFrame());
}
/// <inheritdoc />
protected override void OnReseted()
{
base.OnReseted();
_fast = null;
_slow = null;
_prevFast = 0;
_prevSlow = 0;
_entryPrice = 0;
_cooldown = 0;
}
/// <inheritdoc />
protected override void OnStarted2(DateTime time)
{
base.OnStarted2(time);
_fast = new ExponentialMovingAverage { Length = FastPeriod };
_slow = new ExponentialMovingAverage { Length = SlowPeriod };
var subscription = SubscribeCandles(TimeSpan.FromMinutes(5).TimeFrame());
subscription.Bind(_fast, _slow, ProcessCandle);
subscription.Start();
}
private void ProcessCandle(ICandleMessage candle, decimal fastValue, decimal slowValue)
{
if (candle.State != CandleStates.Finished)
return;
if (!_fast.IsFormed || !_slow.IsFormed)
{
_prevFast = fastValue;
_prevSlow = slowValue;
return;
}
if (_cooldown > 0)
{
_cooldown--;
_prevFast = fastValue;
_prevSlow = slowValue;
return;
}
var close = candle.ClosePrice;
var step = Security?.PriceStep ?? 1m;
// Check SL/TP
if (Position > 0 && _entryPrice > 0)
{
if (StopLossPoints > 0 && close <= _entryPrice - StopLossPoints * step)
{
SellMarket();
_entryPrice = 0;
_cooldown = 80;
_prevFast = fastValue;
_prevSlow = slowValue;
return;
}
if (TakeProfitPoints > 0 && close >= _entryPrice + TakeProfitPoints * step)
{
SellMarket();
_entryPrice = 0;
_cooldown = 80;
_prevFast = fastValue;
_prevSlow = slowValue;
return;
}
}
else if (Position < 0 && _entryPrice > 0)
{
if (StopLossPoints > 0 && close >= _entryPrice + StopLossPoints * step)
{
BuyMarket();
_entryPrice = 0;
_cooldown = 80;
_prevFast = fastValue;
_prevSlow = slowValue;
return;
}
if (TakeProfitPoints > 0 && close <= _entryPrice - TakeProfitPoints * step)
{
BuyMarket();
_entryPrice = 0;
_cooldown = 80;
_prevFast = fastValue;
_prevSlow = slowValue;
return;
}
}
// EMA crossover
if (_prevFast <= _prevSlow && fastValue > slowValue && Position <= 0)
{
if (Position < 0)
BuyMarket();
BuyMarket();
_entryPrice = close;
_cooldown = 80;
}
else if (_prevFast >= _prevSlow && fastValue < slowValue && Position >= 0)
{
if (Position > 0)
SellMarket();
SellMarket();
_entryPrice = close;
_cooldown = 80;
}
_prevFast = fastValue;
_prevSlow = slowValue;
}
}
import clr
clr.AddReference("StockSharp.Messages")
clr.AddReference("StockSharp.Algo")
clr.AddReference("StockSharp.Algo.Indicators")
clr.AddReference("StockSharp.Algo.Strategies")
from System import TimeSpan
from StockSharp.Messages import DataType, CandleStates
from StockSharp.Algo.Indicators import ExponentialMovingAverage
from StockSharp.Algo.Strategies import Strategy
class exp_afirma_strategy(Strategy):
def __init__(self):
super(exp_afirma_strategy, self).__init__()
self._fast_period = self.Param("FastPeriod", 21) \
.SetDisplay("Fast Period", "Fast EMA period", "Indicator")
self._slow_period = self.Param("SlowPeriod", 100) \
.SetDisplay("Slow Period", "Slow EMA period", "Indicator")
self._stop_loss_points = self.Param("StopLossPoints", 200) \
.SetDisplay("Stop Loss", "Stop-loss in price steps", "Risk")
self._take_profit_points = self.Param("TakeProfitPoints", 400) \
.SetDisplay("Take Profit", "Take-profit in price steps", "Risk")
self._fast = None
self._slow = None
self._prev_fast = 0.0
self._prev_slow = 0.0
self._entry_price = 0.0
self._cooldown = 0
@property
def fast_period(self):
return self._fast_period.Value
@property
def slow_period(self):
return self._slow_period.Value
@property
def stop_loss_points(self):
return self._stop_loss_points.Value
@property
def take_profit_points(self):
return self._take_profit_points.Value
def OnReseted(self):
super(exp_afirma_strategy, self).OnReseted()
self._fast = None
self._slow = None
self._prev_fast = 0.0
self._prev_slow = 0.0
self._entry_price = 0.0
self._cooldown = 0
def OnStarted2(self, time):
super(exp_afirma_strategy, self).OnStarted2(time)
self._fast = ExponentialMovingAverage()
self._fast.Length = self.fast_period
self._slow = ExponentialMovingAverage()
self._slow.Length = self.slow_period
subscription = self.SubscribeCandles(DataType.TimeFrame(TimeSpan.FromMinutes(5)))
subscription.Bind(self._fast, self._slow, self._process_candle)
subscription.Start()
def _process_candle(self, candle, fast_value, slow_value):
if candle.State != CandleStates.Finished:
return
fast_val = float(fast_value)
slow_val = float(slow_value)
if not self._fast.IsFormed or not self._slow.IsFormed:
self._prev_fast = fast_val
self._prev_slow = slow_val
return
if self._cooldown > 0:
self._cooldown -= 1
self._prev_fast = fast_val
self._prev_slow = slow_val
return
close = float(candle.ClosePrice)
step = float(self.Security.PriceStep) if self.Security is not None and self.Security.PriceStep is not None else 1.0
if self.Position > 0 and self._entry_price > 0:
if self.stop_loss_points > 0 and close <= self._entry_price - self.stop_loss_points * step:
self.SellMarket()
self._entry_price = 0.0
self._cooldown = 80
self._prev_fast = fast_val
self._prev_slow = slow_val
return
if self.take_profit_points > 0 and close >= self._entry_price + self.take_profit_points * step:
self.SellMarket()
self._entry_price = 0.0
self._cooldown = 80
self._prev_fast = fast_val
self._prev_slow = slow_val
return
elif self.Position < 0 and self._entry_price > 0:
if self.stop_loss_points > 0 and close >= self._entry_price + self.stop_loss_points * step:
self.BuyMarket()
self._entry_price = 0.0
self._cooldown = 80
self._prev_fast = fast_val
self._prev_slow = slow_val
return
if self.take_profit_points > 0 and close <= self._entry_price - self.take_profit_points * step:
self.BuyMarket()
self._entry_price = 0.0
self._cooldown = 80
self._prev_fast = fast_val
self._prev_slow = slow_val
return
if self._prev_fast <= self._prev_slow and fast_val > slow_val and self.Position <= 0:
if self.Position < 0:
self.BuyMarket()
self.BuyMarket()
self._entry_price = close
self._cooldown = 80
elif self._prev_fast >= self._prev_slow and fast_val < slow_val and self.Position >= 0:
if self.Position > 0:
self.SellMarket()
self.SellMarket()
self._entry_price = close
self._cooldown = 80
self._prev_fast = fast_val
self._prev_slow = slow_val
def CreateClone(self):
return exp_afirma_strategy()