New trend indicator: MSR by David Varadi

Author: Eugene

Creation Date: 10/29/2010 11:45 AM

Wealth Lab » Technical Indicators » New trend indicator: MSR by David Varadi

Eugene

10/29/2010 11:45 AM

A New Trend Indicator: MSR

MSR= (10-day median of (H, L, C) – 20-day MAX (H, L, C))/(20-day MAX (H, L, C))
then take the 252-day percentrank of MSR or percentile ranking

QUOTE:
long trades initiated>.5
short trades are initiated<.5

I understood the "(20-day MAX (H, L, C))" as Highest High of 20 days (give it a thought).

Requires Community Indicators library installed.

CODE:
Please log in to see this code.

Size:

Color:

[LINK=http://cssanalytics.wordpress.com/2010/10/27/a-new-trend-indicator-msr/]A New Trend Indicator: MSR[/LINK]

[b]MSR[/b]= (10-day median of (H, L, C) – 20-day  MAX (H, L, C))/(20-day  MAX (H, L, C))
then take the 252-day percentrank of MSR or percentile ranking

[QUOTE]long trades initiated>.5
short trades are initiated<.5[/QUOTE]

I understood the "(20-day MAX (H, L, C))" as Highest High of 20 days (give it a thought).

Requires [LINK=/Extensions/Details/CommunityIndicators-11][b]Community Indicators[/b] library[/LINK] installed.

[CODE]using System;
using System.Collections.Generic;
using System.Text;
using System.Drawing;
using WealthLab;
using WealthLab.Indicators;
using Community.Indicators;

namespace WealthLab.Strategies
{
	public class MSRStrategy : WealthScript
	{
		static double GetMedian(double[] pNumbers)
		{
			int size = pNumbers.Length;
			int mid = size / 2;
			double median = (size % 2 != 0) ? (double)pNumbers[mid] :
				((double)pNumbers[mid] + (double)pNumbers[mid-1]) / 2;
			return median;
		}
		
		static double GetMedian(List<double> pNumbers)
		{
			return GetMedian(pNumbers.ToArray());
		}
		
		private StrategyParameter paramPer1;
		private StrategyParameter paramPer2;
		
		public MSRStrategy()
		{
			paramPer1 = CreateParameter("Period 1", 10, 2, 300, 1);
			paramPer2 = CreateParameter("Period 2", 20, 3, 300, 1);
		}
		
		protected override void Execute()
		{
			//MSR= (10-day median of (H, L, C) – 20-day  MAX (H, L, C))/(20-day  MAX (H, L, C))
			//then take the 252-day percentrank of MSR or percentile ranking
			
			int per1 = paramPer1.ValueInt;
			int per2 = paramPer2.ValueInt;
			
			DataSeries MSR = new DataSeries( Bars, "MSR" );
			
			for(int bar = per2; bar < Bars.Count; bar++)
			{
				double max20 = Highest.Series( High,per2 )[bar];
				List<double> prices = new List<double>(per1 * 3);
				for( int i = bar; i > bar - per1; i-- )
				{
					prices.Add( High[i] ); prices.Add( Low[i] ); prices.Add( Close[i] );
				}
				prices.Sort();
				MSR[bar] = ( GetMedian( prices ) - max20 ) / max20;
			}
			
			ChartPane msr = CreatePane( 30,true,true ); HideVolume();
			PlotSeries( msr, MSR, Color.Blue, LineStyle.Solid, 2 );
			
			PercentRank pr = PercentRank.Series(MSR,252);
			PrcRank pc = PrcRank.Series(MSR,252);
			
			ChartPane p = CreatePane( 30,true,true );
			PlotSeries( p, pr, Color.Red, LineStyle.Solid, 2 );
			PlotSeries( p, pc, Color.Blue, LineStyle.Solid, 2 );
			DrawHorzLine( p, 0.5, Color.Red, LineStyle.Dashed, 1 );

for(int bar = 252; bar < Bars.Count; bar++)
			{
				if (IsLastPositionActive)
				{
					if( CrossUnder( bar, pr, 0.5d ) )
						SellAtMarket( bar+1, LastPosition );
				}
				else
				{
					if( CrossOver( bar, pr, 0.5d ) )
						BuyAtMarket( bar+1 );
				}
			}
		}
	}
}[/CODE]

thodder

11/19/2010 9:40 AM

Hum. Does GetMedian give you the result you want? The code for GetMedian looks correct; however, it seems like you would need to sort the prices before calling GetMedian if you wanted to get the median value.

Eugene

11/19/2010 9:57 AM

Never hurts having someone reviewing your code as you said earlier ;)

My bad, I forgot to sort the List<double> prices before calling GetMedian. Thank you for spotting the bug.

thodder

11/19/2010 1:39 PM

You're welcome.

That's why I prefer to put the sort within methods like this. It is easy to forget to sort the list before you call the method. You might lose a little performance calling sort each time if it is already sorted, but most algorithms are quick if the data is already in the right order.