The RPI story: Of 349 teams, Flashes are 234; their schedule is 306

As we and a lot of other people have been saying for weeks, this is clearly the best Kent State women’s basketball team in six years.

The Flashes are two games above .500 at 6-4. The last time they were in that range was 2010-11, when they won their first six games and went on to a 20-10 record, 11-5 and second in the MAC East. The next year they were 6-21 in Bob Lindsay’s last season, then 21-98 over four years under Danny O’Banion. Best year in her time was 7-23 in 2013-14.

How do we evaluate the Flashes?

Both Kent State’s RPI and strength of schedule right now are below average, I’d argue that schedule strength is somewhat misleading.

KSU’s RPI is 234 of 349 Division I teams. Its strength of schedule is 308. (For comparison, The Flashes’ RPI at the end of last season was 318. Its strength of schedule was 192.)

RPI is Rating Percentage Index, which is a way of ranking teams based on their record and strength of schedule.

Here’s how it’s figured: 25 percent of the score comes from a team’s won-loss percentage. 50 percent comes from its opponents’ won-loss record, and the last 25 from opponents’ opponents’ record. Road wins are weighted 1.4, home wins. 0.6, road losses .6 and home losses 1.4. Neutral site games (of which KSU had three this season) are weighted 1.0.

The RPI is used to some extent in seeding in the NCAA tournament, though people who watch such things say that’s not as true as it once was.

Best way to boost your RPI is to beat a good team on the road, and that’s what happened when KSU won at Wright State last week. Its RPI jumped 60 spots.

So why is KSU’s schedule strength so low so far, especially since coach Todd Starkey has talked a number of times about how much tougher this year’s non-conference schedule is?

First answer is early season RPIs can be misleading. For example, the No. 3 RPI team in the country right now is the Horizon League’s Green Bay. As the season goes on, mid-majors like Green Bay almost always fall. Major conference schools move up because they play each other.

Second answer is that the non-conference schedule isn’t over. Next week the Flashes travel to Iowa on Tuesday and Minnesota on Thursday. Iowa’s current RPI is 44; Minnesota’s is 19. So Kent State’s strength of schedule rating will jump significantly before Christmas. (Of course, the Flashes will have a hard time winning those games, too.)

Third answer is that the records of the mid-major teams Kent State has played haven’t been as good as they teams were predicted to be.

Eastern Kentucky, for example, was picked third in the Ohio Valley Conference. So far its record is 2-6 against Division I teams (RPI uses only Division I opponents) and its RPI is 311.

Similarly Youngstown State, the team KSU beat Saturday, was picked fourth in the Horizon League. The Penguins are 2-6 and have an RPI of 310. Part of their record is due to injuries to three players expected to be starters this season. Youngstown’s RPI last season was 130.

Here’s a rundown of KSU’s 10 opponents, their records against Division I teams, strength of schedule and preseason prospectus:

Bradley (1-6). RPI 263, schedule strength 95. Bradley was picked seven of 10 teams in the Missouri Valley Conference. KSU won at home, 77-52.

Eastern Kentucky (2-6). RPI 311, schedule 269. Eastern was picked third in the 12-team Ohio Valley. KSU beat Eastern in Kent, 80-67.

Robert Morris (3-6). RPI 198, schedule 186. Picked second in the 10-team Northeastern Conference. KSU lost at home, 68-65 in overtime.

Detroit (4-6). RPI 176, schedule 168. Picked second in the nine-team Horizon League. KSU lost in Detroit, 73-52.

Baylor (10-1). RPI 21, schedule 70. Picked first in the 10-team Big 12. KSU lost 84-42 in Gulf Coast Showcase tournament.

Western Kentucky (5-4). RPI 119, schedule 113. Picked first in 14-team Conference USA. KSU lost 79-73 in overtime at Gulf Coast tournament.

Florida Gulf Coast (4-5). RPI 145, schedule 91. Picks first in nine-team Atlantic Sun Conference. KSU won 77-64 at the tournament.

Fort Wayne (1-7). RPI 335, schedule 329. Picked ninth (last) in Summit League. KSU won at home, 66-55.

Wright State (3-4). RPI 225, schedule 272. Picked fifth in nine-team Horizon League. KSU won on road, 79-69.

Youngstown State (2-6) (RPI 310), schedule 278. Picked fourth in nine-team Horizon League. KSU won at home, 75-60.

What do we conclude from all that?

Pretty much what KSU’s record says: KSU is a slightly above average mid-major team.

The Power Conferences — Pac 12, SEC, ACC, Big 12, American Athletic, Big East, Big Ten in that order so far this season — dominate the RPIs. Eleven of the 12 Pac 12 schools, for example, have RPIs below 75.

So KSU’s 234 overall isn’t too bad. KSU’s schedule isn’t good, but that will change in the next 10 days.

In the MAC, KSU is decidedly below average. Its RPI is eighth of 12 teams; its schedule is 10th.

We’ll take a detailed look at the MAC later this week. Right now, best in the MAC are Buffalo (18 RPI at 8-0), Ohio (74 at 7-1) and Toledo (109 at 7-1).

 

Toughest schedule belongs to Buffalo at 102. Second is Miami (2-6) at 184.

(There are various places online you can find RPI Ratings. I use a service call RealTimeRPI. The NCAA’s own list is here. They use basically the same formula. I like RealTimeRPI because it includes schedule strength and breaks it down by conference, among other things. Outside of the main list, though, it costs $27 a year.)

Advertisements

5 comments

  1. goldenflash101

    Pretty much right on. The odd thing about RPI is that once conference season starts your opponents records tend to come together. All teams play teams that average out at 0.500(+/- your own record) so all teams come together on the 75% factors. Home/away wins/losses and unbalanced conference schedules mean that is not 100% true but it is pretty close to that. So only two items really move your RPI significantly after January 1. One is your record. A 6-6 Kent team would gain about 0.041 if it would go 11-7 in conference. It actually would be a bit more because of the bonuses/penalties for road wins and home losses. The second is if your non-conference foes all do well in their conferences. That will help your opponents record, 50% factor at the expense of your opponents/opponents record, only a 25% factor. A good trade off.

    The best schedule for a mid major is play one team from about 3 power conferences win or lose to help the SOS and then be good enough to win games against some average teams and in your conference to make sure your record is good enough. At least Buffalo and Ohio have a chance to do that this year. The only problem with that is the committee doesn’t fall for that as easily as it used too. Kent had a couple of seasons in the high 30’s to mid 40’s and still could not get an at large. Now you have to win one of those power conference games to make it work. The committee won’t reward you for using mathematical tricks to get your RPI up.

    Like

  2. goldenflash101

    By the way Real time predicts Buffalo will end the season at 24-4 and Kent will end the season at 8-22. They use RPI to make the predictions so let’s hope they are wrong.

    Like

  3. Carl Schierhorn

    My impression is that RPI doesn’t mean much any more in the NCAA selections.

    The other thing about RPI after conference season starts is that the Power Conferences teams all tend to go up because they’re playing each other, and most of those teams have good non-conference records. And in very competitive mid-major league — like the MAC men in some years, teams tend to beat each other up and drive down their records and, therefore, their RPI.

    Best case for the MAC getting two bids — which hasn’t happened in 20 years — is two really good teams and nobody else with much of a conference record. That’s how volleyball got two bids in fall. I’ll have more on the MAC RPI later in the week. (League is 11th of 32 right now.)

    I think RealTimeRPI makes its game predictions by picking the team with higher RPI, adjusted for home games. Since KSU’s RPI is only eighth — and not that far off from Miami, BG and Eastern — it’s predicting we’ll win only home games against Miami and BG. They’ve been predicting KSU winning just two league games from the start, and that hasn’t changed even though we’ve won more non-league games than they predicted. I think we’ll do better than that.

    Like

  4. goldenflash101

    Your right about their predictions but I’ve seen over the years that while they don’t do well on specific games their overall prediction for conference record is usually off by no more than three or so games. I’m looking for better than 5-13. Eight to ten would be more to my current expectations.

    Like

  5. Carl Schierhorn

    That win total would be my current guess, too, but I don’t know if I’m being a homer or have some judgment. MAC looks awfully good this season.

    Like

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s