在 GitHub 上查看

Market Predictor 策略

概述

Market Predictor 策略是对 MetaTrader 平台 MarketPredictor 智能交易顾问的 StockSharp 高级 API 版本。策略通过蒙特卡洛预测与基于最新蜡烛的自适应统计参数相结合,持续评估下一步价格走势,只处理已经收盘的蜡烛。

核心思想

  • 自适应均值 (mu): 使用简单移动平均线动态更新期望值,对应原始 EA 中的参数优化步骤。
  • 波动性驱动的幅度: ATR 控制幅度系数 (alpha),让策略能够捕捉波动率的增加。
  • 蒙特卡洛预测: 每根蜡烛结束时运行可配置数量的随机模拟,计算未来价格 (P_t1) 的平均值。
  • 方向判定: 当预测价格与最新收盘价的差异超过 sigma 阈值时,以市价单开仓;只有在没有持仓时才允许入场。

交易规则

  1. 等待蜡烛收盘并确认所有指标形成。
  2. 通过 SMA 更新 mu,通过 ATR 更新 alpha
  3. 围绕最新收盘价执行蒙特卡洛模拟。
  4. 如果平均模拟价格高于 Close + sigma,在空仓时买入开多。
  5. 如果平均模拟价格低于 Close - sigma,在空仓时卖出开空。
  6. 持仓直至出现反向信号。

参数

  • InitialAlpha – ATR 尚未形成时使用的默认幅度。
  • InitialBeta – 保留的系数,与原始 EA 保持一致(当前未直接参与计算)。
  • InitialGamma – 保留的阻尼系数(当前未直接使用)。
  • Kappa – Sigmoid 概念的敏感度参数,作为未来扩展的占位符。
  • InitialMu – SMA 尚未形成时的默认平均值。
  • Sigma – 触发信号所需的预测价格偏离阈值。
  • MonteCarloSimulations – 预测下一价格所使用的模拟次数。
  • CandleType – 使用的蜡烛类型(时间框架)。

备注

  • 策略依赖 StockSharp 的高级 API 实现蜡烛订阅、指标绑定及市价委托。
  • 源码中的注释全部为英文,便于国际化协同。
  • 根据需求,未创建 Python 版本及其目录。
using System;
using System.Linq;
using System.Collections.Generic;

using Ecng.Common;
using Ecng.Collections;
using Ecng.Serialization;

using StockSharp.Algo.Indicators;
using StockSharp.Algo.Strategies;
using StockSharp.BusinessEntities;
using StockSharp.Messages;

namespace StockSharp.Samples.Strategies;

/// <summary>
/// Strategy that adapts the MarketPredictor expert advisor logic to StockSharp.
/// It recalculates statistical parameters from finished candles and
/// runs a Monte Carlo forecast to determine directional bias.
/// </summary>
public class MarketPredictorStrategy : Strategy
{
	private readonly StrategyParam<decimal> _initialAlpha;
	private readonly StrategyParam<decimal> _initialBeta;
	private readonly StrategyParam<decimal> _initialGamma;
	private readonly StrategyParam<decimal> _kappa;
	private readonly StrategyParam<decimal> _initialMu;
	private readonly StrategyParam<decimal> _sigma;
	private readonly StrategyParam<int> _monteCarloSimulations;
	private readonly StrategyParam<DataType> _candleType;

	private decimal _alpha;
	private decimal _mu;

	/// <summary>
	/// Default amplitude used before ATR values become available.
	/// </summary>
	public decimal InitialAlpha
	{
		get => _initialAlpha.Value;
		set => _initialAlpha.Value = value;
	}

	/// <summary>
	/// Placeholder coefficient retained from the original expert advisor.
	/// </summary>
	public decimal InitialBeta
	{
		get => _initialBeta.Value;
		set => _initialBeta.Value = value;
	}

	/// <summary>
	/// Placeholder damping constant retained for documentation purposes.
	/// </summary>
	public decimal InitialGamma
	{
		get => _initialGamma.Value;
		set => _initialGamma.Value = value;
	}

	/// <summary>
	/// Sensitivity parameter associated with the sigmoid concept.
	/// </summary>
	public decimal Kappa
	{
		get => _kappa.Value;
		set => _kappa.Value = value;
	}

	/// <summary>
	/// Default mean price used until the moving average is formed.
	/// </summary>
	public decimal InitialMu
	{
		get => _initialMu.Value;
		set => _initialMu.Value = value;
	}

	/// <summary>
	/// Required deviation between forecast and latest close to trigger entries.
	/// </summary>
	public decimal Sigma
	{
		get => _sigma.Value;
		set => _sigma.Value = value;
	}

	/// <summary>
	/// Number of Monte Carlo simulations used to forecast the next price.
	/// </summary>
	public int MonteCarloSimulations
	{
		get => _monteCarloSimulations.Value;
		set => _monteCarloSimulations.Value = value;
	}

	/// <summary>
	/// Candle type used for the strategy calculations.
	/// </summary>
	public DataType CandleType
	{
		get => _candleType.Value;
		set => _candleType.Value = value;
	}

	/// <summary>
	/// Initializes parameters with defaults matching the original EA configuration.
	/// </summary>
	public MarketPredictorStrategy()
	{
		_initialAlpha = Param(nameof(InitialAlpha), 0.1m)
			.SetDisplay("Initial Alpha", "Default amplitude before ATR is formed", "Prediction")
			
			.SetOptimize(0.05m, 0.5m, 0.05m);

		_initialBeta = Param(nameof(InitialBeta), 0.1m)
			.SetDisplay("Initial Beta", "Fractal weight placeholder", "Prediction")
			;

		_initialGamma = Param(nameof(InitialGamma), 0.1m)
			.SetDisplay("Initial Gamma", "Fractal damping placeholder", "Prediction")
			;

		_kappa = Param(nameof(Kappa), 1.0m)
			.SetDisplay("Kappa", "Sigmoid sensitivity placeholder", "Prediction")
			;

		_initialMu = Param(nameof(InitialMu), 1.0m)
			.SetDisplay("Initial Mu", "Fallback mean price", "Prediction")
			
			.SetOptimize(0.5m, 2.0m, 0.25m);

		_sigma = Param(nameof(Sigma), 10.0m)
			.SetGreaterThanZero()
			.SetDisplay("Sigma", "Deviation threshold for trades", "Trading")
			
			.SetOptimize(1.0m, 30.0m, 1.0m);

		_monteCarloSimulations = Param(nameof(MonteCarloSimulations), 1000)
			.SetGreaterThanZero()
			.SetDisplay("Monte Carlo Simulations", "Number of simulations per candle", "Prediction")
			
			.SetOptimize(100, 2000, 100);

		_candleType = Param(nameof(CandleType), TimeSpan.FromHours(4).TimeFrame())
			.SetDisplay("Candle Type", "Timeframe for candle subscription", "General");
	}

	/// <inheritdoc />
	protected override void OnReseted()
	{
		base.OnReseted();

		_alpha = InitialAlpha;
		_mu = InitialMu;
	}

	/// <inheritdoc />
	protected override void OnStarted2(DateTime time)
	{
		base.OnStarted2(time);

		_alpha = InitialAlpha;
		_mu = InitialMu;

		var sma = new SimpleMovingAverage { Length = 14 };
		var atr = new AverageTrueRange { Length = 14 };

		var subscription = SubscribeCandles(CandleType);

		subscription
			.Bind(sma, atr, (candle, smaValue, atrValue) => ProcessCandle(candle, sma, atr, smaValue, atrValue))
			.Start();
	}

	private void ProcessCandle(ICandleMessage candle, SimpleMovingAverage sma, AverageTrueRange atr, decimal smaValue, decimal atrValue)
	{
		// Process only finished candles to avoid premature trading decisions.
		if (candle.State != CandleStates.Finished)
			return;

		// Confirm that the strategy is allowed to trade and all prerequisites are met.
		// Update adaptive mean when the moving average is formed.
		if (sma.IsFormed)
		{
			_mu = smaValue;
		}
		else
		{
			_mu = InitialMu;
		}

		// Adjust amplitude based on volatility when ATR values are reliable.
		if (atr.IsFormed && atrValue > 0m)
		{
			_alpha = atrValue * 0.1m;
		}
		else
		{
			_alpha = InitialAlpha;
		}

		var currentPrice = candle.ClosePrice;
		ExecuteTrade(currentPrice);
	}

	private void ExecuteTrade(decimal currentPrice)
	{
		var deviation = _alpha > 0 ? Sigma * _alpha : Sigma;

		// Mean-reversion: buy when significantly below mean, sell when significantly above
		if (currentPrice < _mu - deviation && Position <= 0)
		{
			if (Position < 0)
				BuyMarket();
			BuyMarket();
		}
		else if (currentPrice > _mu + deviation && Position >= 0)
		{
			if (Position > 0)
				SellMarket();
			SellMarket();
		}
		// Exit when price returns to mean
		else if (Position > 0 && currentPrice >= _mu)
		{
			SellMarket();
		}
		else if (Position < 0 && currentPrice <= _mu)
		{
			BuyMarket();
		}
	}
}