NCAA Tournament Bubble Watch 2011: The Importance of RPI Top 50 Wins

The overanalysis is back.

The sports media's obsession with the discussion of who is in, who is out and which teams will make the tournament is constant. The fact that most bubble teams do little or no damage in the tournament is meaningless.

This year is the first year I have noticed the ascension of the belief that a team should be judged in or out based on top-50 RPI wins. I know RPI Top 50 wins have been discussed before, but never as the critical factor as to why a team should get in. The reason behind this new method of prediction is simple.

Sports writers are lazy.

A few weeks back, I wrote about Harvard being the stealth at-large team, and although a few analysts may give them a passing look, most experts don't even mention Harvard as a team that should be in the discussion in what is universally thought to be a very soft bubble. The reason given is that Harvard doesn't have enough top-50 RPI wins.

The real reason is that nobody writing these predictions has seen Harvard play this year. I watch tons of basketball, and I haven't even seen an Ivy League games this year.

The laziness leads to people only looking at a select consensus group of teams when they make predictions and look at resumes. This isn't Careerbuilder. It's a basketball tournament, the committee looks at records and schedules. Making the RPI Top 50 wins a key point eliminates the mid-major conference teams. They get far fewer chances to get those wins.

The analysis almost always mentions top-50 wins. For example, Marquette has four wins against teams with an RPI in the top 50. Rarely is it mentioned they also have 11 losses against the same group. 4-11 looks way worse than four wins.

St. Mary's isn't getting 15 chances to win four games, seven of them at home. San Diego State and Cincinnati both have five top-50 wins, and San Diego State had five less games to do it in. If this is really to be a comparative, it should be noted in terms of wins and losses.

The argument for the major conference schools would be that they play in tougher conferences, which is true for the Big East, Big 12, Big 10, and SEC East. The ACC and Pac-10 are not as talented as in years past, and the SEC West has not had a good year. I'm trying to be as nice as possible.

The major conference schools get a lot more opportunities to get that key win. A debate that Alabama should get an at-large bid is that they are 12-4 in the SEC and they won at Tennessee and at home against Kentucky. Oakland needs to win their conference tournament to get in despite being 17-1 in their conference and winning at Tennessee.

The big difference in resumes is the conference name and that Alabama beat Kentucky at home, something Oakland didn't have the chance to do. Oakland's at-large candidacy could still be debated because of losses to Michigan State and Michigan, but they aren't even discussed because they have only one win against the RPI Top 50. Oakland should be ashamed.

How important are RPI Top 50 wins?

I got in the time machine and went all the way back to 2010, and came up with some RPI Top 50 win totals from contests in the NCAA Tournament:

Team Matchups in terms of the number of RPI Top 50 wins in the 2010 NCAA tournament:

12 (wins vs. RPI Top 50) vs. 4

9 vs. 1

8 vs. 3

8 vs. 3

7 vs. 0

7 vs. 3

7 vs. 3

6 vs. 0

5 vs. 1

4 vs. 0

4 vs. 3

4 vs. 3

If RPI Top 50 wins are so important, in 10 of those games the advantage should have been huge. In the other two, the team with four RPI Top 50 wins would have a slight edge. Those numbers are better put in context when I add the names of the teams into the chart.

Team matchups in terms of the number RPI Top 50 Wins in the 2010 NCAA tournament:

12 vs. 4        Kansas vs Northern Iowa

9 vs. 1          Georgetown vs Ohio

8 vs. 3          Syracuse vs Butler

8 vs. 3          Kansas St vs Butler

7 vs. 0          Temple vs Cornell

7 vs. 3          Villanova vs St Mary's

7 vs. 3          Pittsburgh vs Xavier

6 vs. 0          Wisconsin vs Cornell

5 vs. 1          Notre Dame vs Old Dominion

4 vs. 0          Vanderbilt vs Murray St

4 vs. 3          Marquette vs Washington

4 vs. 3          Florida St vs Gonzaga

That's 12 games from the tournament last year. I didn't analyze all of the games. The teams with the two highest totals, Georgetown and Kansas, won exactly one game. Cornell and Murray State would automatically be eliminated from consideration if top 50 wins was the main criteria, yet they won three games in the tournament last year.

An earlier point I made is that most bubble teams do little in the tournament. From 2005 to 2009, eight double-digit seeds have made the Sweet 16. Looking at their records against the RPI Ttop 50 will help in showing that it's a mediocre indicator of success at best.

The eight double-digit seeds to make the Sweet 16 from 2005-2009 were (RPI Top 50 records):

2009 Arizona (6-10)

2008 Villanova (3-7)

2008 Western Kentucky (0-4)

2008 Davidson (0-3)

2006 Bradley (7-6)

2006 George Mason (2-4)

2005 UW Milwaukee (0-2)

2005 NC State (4-8)

It's really a mixed bag. Luckily, teams like Western Kentucky and Davidson won the automatic bid, since their lack of top 50 wins might have left them out, even if they were clearly one of the best 34 at-large teams based on performance and not resumes. Arizona and NC State also show that teams with tough schedules should be included in the discussion.

If I use the history from above, and try to match up the RPI Top 50 wins of those teams with teams currently looking at bubble seeds, these are the teams that best match:

Colorado, Penn State, Cleveland State, Belmont, USC, Richmond, Missouri St and Michigan

This analysis should help show that RPI Top 50 wins as a main factor of picking teams for the tournament is too exclusive and inaccurate. The fact that so many analysts on ESPN, CNNSI and other media outlets view that measure as the Holy Grail ignores the fact that the committee needs to pick the best teams, and that schedule is just one factor.

Luckily most of the time road wins, last 10 games and non-conference SOS also play a factor. That would take too much effort to analyze. The 45 minutes of research I put into this would be mind boggling to bracketologists.