<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"
	xmlns:content="http://purl.org/rss/1.0/modules/content/"
	xmlns:dc="http://purl.org/dc/elements/1.1/"
	xmlns:atom="http://www.w3.org/2005/Atom"
	>
<channel>
	<title>Comments on: bracketology&#8230;</title>
	<atom:link href="http://sockrebel.wordpress.com/2008/03/17/bracketology/feed/" rel="self" type="application/rss+xml" />
	<link>http://sockrebel.wordpress.com/2008/03/17/bracketology/</link>
	<description>...and the chaos that is my life...</description>
	<pubDate>Fri, 25 Jul 2008 01:42:52 +0000</pubDate>
	<generator>http://wordpress.org/?v=MU</generator>
		<item>
		<title>By: sockrebel</title>
		<link>http://sockrebel.wordpress.com/2008/03/17/bracketology/#comment-21598</link>
		<dc:creator>sockrebel</dc:creator>
		<pubDate>Mon, 24 Mar 2008 16:58:38 +0000</pubDate>
		<guid isPermaLink="false">http://sockrebel.wordpress.com/?p=342#comment-21598</guid>
		<description>Just saw this post on http://answers.yahoo.com/question/index?qid=20070308154659AA6j0mI, which I like better:
I have analyzed this over and over again from a mathematical / statistical point of view for quite a few years now. For the past few years I have used per-round scoring of 2-3-5-8-13-21 by round, plus a seed score of seed# x round#. It works out really well for my pool and the point balances seem to be very fair. Let me explain why I chose this system...

First, the standard round scoring of 2-3-5-8-13-21. For those of you who are mathematically inclined (or are big DaVinci Code fans), you will recognize these scores as Fibonacci numbers. This type of progression just happens to fit well with a bracket elimination type of structure, and the weighting of games in progressive rounds seems to make sense. With the more traditional 1-2-4-8-16-32 system, the championship game is worth 32X as much as any first round game, which in effect makes the first round games almost useless. With "Fibonacci" scoring, the championship game is worth 10.5X as much as each first round game, which is much more reasonable. It's not much different than your 1-2-4-6-8-12 system in that regard.

Now, for the seed# x round# part, I will admit that I was initially opposed to any kind of weighted scoring based on seeds. However, my pool participants repeatedly asked for some kind of reward for picking underdogs, so I eventually caved in. After a bunch of number crunching, I decided that seed# x round# fit in very nicely with the Fibonacci number scoring. Let's look at a couple of scenarios. Consider the 2006 tourney where George Mason went to the Final Four as an 11 seed, and Florida, a #1 seed, won the national championship. Anyone who picked Florida to go all the way would have been rewarded with 73 total points. Naturally, lots of people picked them as a champion since they were a top seed. Now, for George Mason....if you had picked them to go to the Final Four, you would have been rewarded with 128 points which is 1.75X the points you would have received for picking Florida as the champ. Considering that few 11 seeds make it to the Final Four, George Mason was a risky pick, and if you had made that pick it would have paid off accordingly.

Now let's say someone picked all four 14 seeds just to try to rack up points early, and let's say that 2 of them win. That player would get 32 points - 16 for each team (2 standard + 14 seed). However, that person would lose 10 of those points right off the bat from the two 3 seeds that won, so his net gain is only 22 points. If either of those 3 seeds wins a second round game, he's out another 14 points for each. So, it clearly does not pay to go heavy on the low seeds to get the early points.

Bottom line is you still have to pick winners, regardless of the scoring system that is in place.</description>
		<content:encoded><![CDATA[<p>Just saw this post on <a href="http://answers.yahoo.com/question/index?qid=20070308154659AA6j0mI" rel="nofollow">http://answers.yahoo.com/question/index?qid=20070308154659AA6j0mI</a>, which I like better:<br />
I have analyzed this over and over again from a mathematical / statistical point of view for quite a few years now. For the past few years I have used per-round scoring of 2-3-5-8-13-21 by round, plus a seed score of seed# x round#. It works out really well for my pool and the point balances seem to be very fair. Let me explain why I chose this system&#8230;</p>
<p>First, the standard round scoring of 2-3-5-8-13-21. For those of you who are mathematically inclined (or are big DaVinci Code fans), you will recognize these scores as Fibonacci numbers. This type of progression just happens to fit well with a bracket elimination type of structure, and the weighting of games in progressive rounds seems to make sense. With the more traditional 1-2-4-8-16-32 system, the championship game is worth 32X as much as any first round game, which in effect makes the first round games almost useless. With &#8220;Fibonacci&#8221; scoring, the championship game is worth 10.5X as much as each first round game, which is much more reasonable. It&#8217;s not much different than your 1-2-4-6-8-12 system in that regard.</p>
<p>Now, for the seed# x round# part, I will admit that I was initially opposed to any kind of weighted scoring based on seeds. However, my pool participants repeatedly asked for some kind of reward for picking underdogs, so I eventually caved in. After a bunch of number crunching, I decided that seed# x round# fit in very nicely with the Fibonacci number scoring. Let&#8217;s look at a couple of scenarios. Consider the 2006 tourney where George Mason went to the Final Four as an 11 seed, and Florida, a #1 seed, won the national championship. Anyone who picked Florida to go all the way would have been rewarded with 73 total points. Naturally, lots of people picked them as a champion since they were a top seed. Now, for George Mason&#8230;.if you had picked them to go to the Final Four, you would have been rewarded with 128 points which is 1.75X the points you would have received for picking Florida as the champ. Considering that few 11 seeds make it to the Final Four, George Mason was a risky pick, and if you had made that pick it would have paid off accordingly.</p>
<p>Now let&#8217;s say someone picked all four 14 seeds just to try to rack up points early, and let&#8217;s say that 2 of them win. That player would get 32 points - 16 for each team (2 standard + 14 seed). However, that person would lose 10 of those points right off the bat from the two 3 seeds that won, so his net gain is only 22 points. If either of those 3 seeds wins a second round game, he&#8217;s out another 14 points for each. So, it clearly does not pay to go heavy on the low seeds to get the early points.</p>
<p>Bottom line is you still have to pick winners, regardless of the scoring system that is in place.</p>
]]></content:encoded>
	</item>
</channel>
</rss>
