Gap Close Data - Wealth Lab Forum

Gap Close Data

Author: Cone

Creation Date: 5/10/2010 8:31 AM

Wealth Lab » Trading Strategies » Gap Close Data

Cone

5/10/2010 8:31 AM

Some interesting data on big gap openings -

CODE:
Please log in to see this code.

Here's the script -

CODE:
Please log in to see this code.

Size:

Color:

Some interesting data on big gap openings -

[CODE]^DJI
Following an up gap of 4% (or more)
Gap Date	Closed Gap	Days	Gap (%)	Max Excursion (%)
10/7/1929	10/21/1929	10	5.0%		10.3%
6/22/1931	9/8/1931	55	7.5%		20.3%
2/15/1932	3/17/1932	22	8.6%		14.3%
6/6/1932	6/7/1932	1	4.6%		4.6%
8/8/1932	8/12/1932	4	6.3%		14.4%
6/19/1933	7/21/1933	23	4.1%		23.9%
11/29/1937	12/29/1937	22	4.3%		10.9%
4/11/1938	5/2/1938	14	5.2%		10.9%
9/5/1939	5/14/1940	172	5.3%		16.7%
10/21/1987	10/22/1987	1	6.0%		13.0%
9/8/1998	9/10/1998	2	4.2%		6.1%
[/CODE]

Here's the script - 
[CODE]using System;
using System.Collections.Generic;
using System.Text;
using System.Drawing;
using WealthLab;
using WealthLab.Indicators;

namespace WealthLab.Strategies
{
	public class ClosingGaps : WealthScript
	{
		StrategyParameter gapPercent;

public ClosingGaps()
		{
			gapPercent = CreateParameter("Percent Gap", 3, 1, 10, 0.5);
		}

internal class Gapper
		{
			private Bars _bars;
			public int gapBar = 0;
			public int closedBar = 0;
			
			public Gapper(Bars bars, int bar)
			{
				gapBar= bar;
				_bars = bars;
			}
			
			public double PctGap()
			{
				return (_bars.Open[gapBar]/_bars.Close[gapBar - 1] - 1);
			}
			public double MaxExcursion()
			{
				int hiBar = (int)HighestBar.Value(closedBar, _bars.High, (closedBar - gapBar + 1));
				return (_bars.High[hiBar] / _bars.Close[gapBar - 1]  - 1);
			}			
		}
		
		protected override void Execute()
		{			
			ClearDebug();
			double gapUp = 1 + gapPercent.Value/100d;
			List<Gapper> list = new List<Gapper>();
			
			PrintDebug(Bars.Symbol);
			PrintDebug("Following an up gap of " + gapPercent.Value + "% (or more)");
			PrintDebug("Gap Date\tClosed Gap\tDays\tGap (%)\tMax Excursion (%)");
			
			for(int bar = 1; bar < Bars.Count; bar++)
			{
				if (Open[bar] > Close[bar - 1] * gapUp)
				{					
					Gapper g = new Gapper(Bars, bar);
					list.Add(g);
				}					
				else
				{
					for (int n = list.Count - 1; n >= 0; n--)
					{
						Gapper g = list[n];
						if (Low[bar] <= Close[g.gapBar - 1])
						{
							g.closedBar = bar;						
							PrintDebug(Date[g.gapBar].ToShortDateString() 
								+ "\t" + Date[g.closedBar].ToShortDateString() 
								+ "\t" + (g.closedBar - g.gapBar)
								+ "\t" + g.PctGap().ToString("0.0%")
								+ "\t\t" + g.MaxExcursion().ToString("0.0%"));	
							list.Remove(g);
						}
					}
				}
			}			
		}
	}
}[/CODE]

dan_rozenberg

5/10/2010 9:50 AM

That's some cool info, Cone.

However, I think its important to add the gap% of each case to the table you built. This is because, say, if on 10/4/1929 the gap was 10.3% then it the market only fell from that gap, and entering at open of gap day would result in losses!

Its important to consider the (Max Excursion% - Opening Gap%) to see if buying at open would be a nice deal.

:)

Cone

5/10/2010 1:31 PM

I updated the version above so that it doesn't miss any gaps and also prints the value of the gap %.

In response to your comment, the take-away I was trying to transmit is that you can almost be assured that the market will close this gap. In the two "modern day" cases, this took only 1 or 2 days. In all but two of the other cases, it took 4 weeks or less of trading (usually less).

By the way, if you add this in the logic block where the gap is detected, you'll see the market closed lower than the opening gap level on the day following the gap in every case.

CODE:
Please log in to see this code.

Maybe this time it will be different?

Cone

5/20/2010 8:02 AM

Just a follow up ... while the Dow data doesn't seem to be capturing the gaps as it used to, that 4% gap up was officially closed yesterday, and today it looks like we'll be opening well below it.

Take note that almost most of those big bull gaps occur during bear markets. End of the cyclical bull and return to the larger trend?