Deutsch DE
Auf GitHub ansehen

MTrendLine Strategy

Overview

The MTrendLine Strategy ports the MetaTrader script MTrendLine.mq4 to StockSharp's high-level strategy API. The original expert advisor repeatedly adjusts the price of existing pending orders so they remain aligned with a trend line drawn on the chart. The StockSharp version automates the same behaviour by rebuilding the moving trend line with a configurable LinearRegression indicator. Up to three independent pending order slots can follow the calculated regression line with their own order type, distance, and volume. Every time a new candle closes the strategy recomputes the line value, evaluates the required offsets, and refreshes the pending orders accordingly.

The port adds modern risk and usability improvements such as structured parameters, automatic conversion from MetaTrader points into real price steps, and optional stop-loss / take-profit distances that move together with the pending orders. Bid/ask updates are monitored via SubscribeLevel1() so the strategy honours the minimal distance that brokers demand between the current market price and resting orders.

Trading logic

  1. Subscribe to the configured candle series through SubscribeCandles() and feed a LinearRegression indicator with each finished bar. The indicator represents the manual trend line from the MetaTrader version.
  2. Maintain Level1 subscriptions to cache the latest best bid and best ask values. They are used to enforce the minimum distance parameter before relocating a pending order.
  3. For every enabled slot compute the desired price as regression value + distance × point size. The point size defaults to the security price step but can be overridden to match MetaTrader's Point constant.
  4. Convert the slot configuration into StockSharp order helpers (BuyLimit, SellLimit, BuyStop, SellStop). Optional stop-loss and take-profit prices are derived from the requested distance in points so they track the order after each move.
  5. If an active pending order already exists for the slot and the new target price differs, cancel the current order first and wait for the next candle to place the updated one. This mirrors the behaviour of OrderModify from the MQL code without risking duplicate requests.
  6. When a slot is disabled or the computed price becomes invalid (for example, negative), cancel the associated pending order and clear its cached state.

Pending order slots

Each slot emulates one call to modify() in the original EA. Slots can be configured independently:

  • Type — choose between Buy Limit, Buy Stop, Sell Limit, or Sell Stop.
  • Distance — distance in MetaTrader points added to the regression value to obtain the new price. Use negative values to position orders below the regression line.
  • Volume — size of the pending order. If set to zero or negative, the strategy falls back to the global TradeVolume.
  • Enable flag — allows disabling a slot without removing its configuration. Disabled slots automatically cancel any active orders that belong to them.

Parameters

Name Type Default Description
CandleType DataType 1-hour candles Primary timeframe used to build the regression trend line.
RegressionLength int 24 Number of completed candles fed into the LinearRegression indicator.
PointValue decimal 0 Monetary value of one MetaTrader point. When zero the strategy uses the security price step.
TradeVolume decimal 1 Default volume used by all slots when their own volume is zero.
StopLossPoints decimal 0 Stop-loss distance in points. Set to zero to disable automatic stop-loss placement.
TakeProfitPoints decimal 0 Take-profit distance in points. Set to zero to disable automatic take-profit placement.
MinDistancePoints decimal 0 Minimum gap (in points) that must exist between the best bid/ask and the pending order.
PendingOrder{1,2,3}Enabled bool Slot specific Enables or disables the given slot.
PendingOrder{1,2,3}Mode enum Slot specific Pending order type: BuyLimit, BuyStop, SellLimit, or SellStop.
PendingOrder{1,2,3}DistancePoints decimal Slot specific Distance (in points) added to the regression value to compute the order price.
PendingOrder{1,2,3}Volume decimal Slot specific Volume for the slot. Zero reuses TradeVolume.

Differences from the original MetaTrader script

  • MetaTrader modifies existing orders in place. StockSharp uses cancel-and-replace semantics while waiting for confirmation before registering the replacement order on the next candle.
  • The original code reads the value of a manually drawn trend line. The port replaces this with an automatic LinearRegression indicator so the behaviour is deterministic and can run unattended.
  • MODE_STOPLEVEL is not available on StockSharp. Instead, the strategy provides the configurable MinDistancePoints parameter and enforces it using real-time bid/ask updates.
  • Stop-loss and take-profit distances are optional parameters instead of reading existing order settings. This keeps the values consistent across order re-registrations.

Usage tips

  • Set PointValue to match the broker's point definition if it differs from the security PriceStep; this guarantees the distance parameters mirror their MetaTrader counterparts.
  • Enable only the slots you need. Each slot maintains its own pending order and comment ("MTrendLine slot N") so identifying them in reports or the Order Log is straightforward.
  • Consider combining the strategy with StockSharp's built-in risk protection helpers if you require trailing stops or account level controls. The implementation focuses on mirroring the original order-modification logic.

Indicators

  • LinearRegression applied to finished candles.
using System;

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

namespace StockSharp.Samples.Strategies;

/// <summary>
/// Trend line strategy using linear regression slope for direction.
/// Enters long when slope is positive and price is above regression, short otherwise.
/// </summary>
public class MTrendLineStrategy : Strategy
{
	private readonly StrategyParam<DataType> _candleType;
	private readonly StrategyParam<int> _regressionLength;

	private decimal _prevSlope;
	private bool _hasPrev;

	public MTrendLineStrategy()
	{
		_candleType = Param(nameof(CandleType), TimeSpan.FromMinutes(5).TimeFrame())
			.SetDisplay("Candle Type", "Timeframe for analysis.", "General");

		_regressionLength = Param(nameof(RegressionLength), 20)
			.SetDisplay("Regression Length", "Length of the linear regression.", "Indicators");
	}

	public DataType CandleType
	{
		get => _candleType.Value;
		set => _candleType.Value = value;
	}

	public int RegressionLength
	{
		get => _regressionLength.Value;
		set => _regressionLength.Value = value;
	}

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

		_prevSlope = 0;
		_hasPrev = false;
	}

		protected override void OnStarted2(DateTime time)
	{
		base.OnStarted2(time);

		_prevSlope = 0;
		_hasPrev = false;

		var lr = new LinearReg { Length = RegressionLength };
		var ema = new ExponentialMovingAverage { Length = RegressionLength };

		var subscription = SubscribeCandles(CandleType);
		subscription
			.Bind(lr, ema, ProcessCandle)
			.Start();

		var area = CreateChartArea();
		if (area != null)
		{
			DrawCandles(area, subscription);
			DrawIndicator(area, lr);
			DrawIndicator(area, ema);
			DrawOwnTrades(area);
		}
	}

	private void ProcessCandle(ICandleMessage candle, decimal lrValue, decimal emaValue)
	{
		if (candle.State != CandleStates.Finished)
			return;

		// Compute slope as difference between LR and EMA
		var slope = lrValue - emaValue;

		if (!_hasPrev)
		{
			_prevSlope = slope;
			_hasPrev = true;
			return;
		}

		var close = candle.ClosePrice;

		// Exit conditions
		if (Position > 0 && slope < 0 && _prevSlope >= 0)
		{
			SellMarket();
		}
		else if (Position < 0 && slope > 0 && _prevSlope <= 0)
		{
			BuyMarket();
		}

		// Entry conditions
		if (Position == 0)
		{
			if (slope > 0 && _prevSlope <= 0 && close > lrValue)
			{
				BuyMarket();
			}
			else if (slope < 0 && _prevSlope >= 0 && close < lrValue)
			{
				SellMarket();
			}
		}

		_prevSlope = slope;
	}
}