Thread: *ANSWER* Question 9 View Single Post
#3
04-08-2013, 08:15 AM
 Michael Reach Senior Member Join Date: Apr 2013 Location: Baltimore, Maryland, USA Posts: 71
Re: *ANSWER* Question 9

My mean was considerably smaller, but there is a very long tail. If you graph the lines at various stages, you can actually see the h hypothesis line jiggling back and forth as it gets closer to the "g" line. The size of the jiggles damps down as you go along. It needs to get close enough so that none of your random chosen set of points is on the wrong side.

If you picture shifting the g line around as far as you can till it bumps into one of the points, you get the range of potential solution "h" lines. You'll see that it depends on the initial points how much room you have for that; often you'll have lots of room. But one could imagine three points that are almost exactly colinear and the g line goes just between them - and there's almost no room to jiggle! In a case like that, it can take an awfully long time till the jiggling of the the h line settles down far enough so that it doesn't keep jumping across that very narrow range.