NCAA Rankings - Does this seem correct to you?

[Rocketech1]

The latest NCAA rankings are out and the RPI rankings and ESPN rankings are very different where the top 3 in RPI are SEC teams and the top 3 in ESPN/Coaches polls are PAC 12 teams. I guess the WCWS will be the final decision maker.

RPI

1 Florida Southeastern
2 Alabama Southeastern
3 Georgia Southeastern
4 UCLA Pac-12
5 Oregon Pac-12
6 La.-Lafayette Sun Belt
7 Florida St. Atlantic Coast
8 Oklahoma Big 12
9 Missouri Southeastern
10 Arizona St.

ESPN

1 Oregon
2 UCLA
3 Arizona State
4 Florida State
5 Alabama
6 Florida
7 Tennessee
8 Washington
9 Arizona
10 Oklahoma

 

[sluggers]

Since this is math based, I have studied the RPI algorithms.

SO:

An RPI is based on "who beat who". The logic is something like, "Sally beat Suzie. Suzie beat Cindy. Therefore, Sally is better than Cindy." It gets more complicated, but that is the approach the RPI takes.

The RPI was developed to do chess rankings. For individual competition (like golf, tennis, chess, etc.), it works very, very well. In individual sports, the same people compete against each other, so comparing players works out great.

In team sports, there is a problem because players change. So, if a team's big dog pitcher has the flu one weekend, and the team loses to the Sisters of the Poor, U, then that team's RPI drops quite a bit.

The other problem with the RPI is there is the assumption that the teams play each other randomly throughout the season. In softball, the teams play other teams outside of their conference for two months, and then play *only* teams within their conference for the final two months of the season. E.g., Alabama plays many different teams for the first two months. Then, in April and May, it only plays SEC teams.

Non-conference games played at the beginning of the season where teams are trying to "work out the bugs" to be just as important as the conference games at the end of the season.

So, the RPI for team sports is more of a curiosity than anything else.

Of course, the polls aren't better.

 

[SoCal_Dad]

RPI is the primary factor in awarding at-large berths and seeding teams so it has a bigger influence on teams getting to the WCWS. NCAA makes some adjustments, but usually it is within 1-2 of the RPI ranking.

No system is perfect. The SEC teams fare well in RPI due to it's heavy weighting of opponents' winning percentages. The top PAC 12 teams fare better in the polls, but haven't figured out yet how to enhance their RPI via their non-conference scheduling.

 

[SoCal_Dad]

An RPI is based on "who beat who". The logic is something like, "Sally beat Suzie. Suzie beat Cindy. Therefore, Sally is better than Cindy." It gets more complicated, but that is the approach the RPI takes.

NCAA's RPI isn't that specific. The base formula is 25% team's Winning Pct + 50% Opponents' WP + 25% Opponents' Opponents' WP. Each sport has different ways of adjusting it for home/away games and bonus/penalty points. From what I've read, softball doesn't adjust for home/away, but they do give bonus points for wins against non-conference opponents in the top 25/50/75 RPI and penalize for losses against teams in the bottom 25/50/75 RPI.

A conference can help its members by having them select non-conference opponents with a good winning pct. They don't have to be particularly strong teams, they just need to win games against weaker teams. The high OWP not only helps that member's RPI, it also helps the rest of the members too because it counts in their OOWP.

Most of the non-conference schedule is played early in the season, but many/most D1 teams also schedule midweek non-conference games between their conference games.

 

[Lenski]

RPI ranks strength of schedule and how a team does against that strength of schedule. ESPN ranks teams from their perceived "best to worst". So in ESPN's case, they believe the top PAC-12 teams are better and the RPI formula suggests the top SEC teams play a tougher schedule. Comparing RPI to the other rankings is not a true apples to apples comparison.

 

[CoogansBluff]

RPI is not the best math model because it doesn't consider margin of victory. Many fear that using MOV will encourage or reward teams that run up the score, but in reality, it allows a more thorough evaluation of the teams. If you used a good model that accounts for MOV, it would be as good or probably slightly better than a coaches poll.

I will add, however, that softball and baseball provide a little bit tougher challenge for rankings because of how the strength of a team changes depending on the starting pitcher. You might lose mid-week to UAB w/ your #2 pitcher and then beat Alabama with your #1. That makes Alabama look worse than they really are. Those kinds of things can happen in any sport (starting QB was hurt one week), but it's more of a built-in issue w/ softball and baseball than anything else. The hope is that it evens out.

 

[29dad]

Michigan drops one meaningless series at the end of the regular season to Wisconsin and they drop all the way down to 18. Ridiculous.

 

[CoogansBluff]

Michigan drops one meaningless series at the end of the regular season to Wisconsin and they drop all the way down to 18. Ridiculous.

Why was it meaningless? All games count toward being seeded in the NCAA tournament. Winning the conference is meaningless in the national picture.

 

[29dad]

Why was it meaningless? All games count toward being seeded in the NCAA tournament. Winning the conference is meaningless in the national picture.

Fair enough. I used the word "meaningless" because Michigan had already clinched the Big 10 title and it was the last series of the regular season. Wisconsin took 2 of 3 and it was Michigan's only Big 10 series loss. But go ahead and take the word "meaningless" out of my statement. How do you explain dropping them all the way down to 18, even behind Minnesota? Michigan won the Big 10. They took 2 of 3 from my Gophers. How do they drop below Minnesota? I don't get it.

Rankings | NCAA.com

 

[CoogansBluff]

Fair enough. I used the word "meaningless" because Michigan had already clinched the Big 10 title and it was the last series of the regular season. Wisconsin took 2 of 3 and it was Michigan's only Big 10 series loss. But go ahead and take the word "meaningless" out of my statement. How do you explain dropping them all the way down to 18, even behind Minnesota? Michigan won the Big 10. They took 2 of 3 from my Gophers. How do they drop below Minnesota? I don't get it.

Rankings | NCAA.com

I like these rankings better. Gophers are #12. Michigan #14. They're more sophisticated than RPI. Maybe Minnesota has a better body of work. I haven't studied it, though.

Massey Ratings