# RPI and NCAA Seeds

Recently I have noticed a tendency to predict Stateâ€™s seed in the NCAA tournament based on projections for a second-place finish in the ACC. While finishing second in the conference is certainly within reach, most of the projected seeds seemed too high to me. Given the decline in the ACC this year and the impact that Stateâ€™s cupcake schedule is having on its RPIâ€¦it just seemed to me that a high seed was far from guaranteed.

So, I looked at the teams in the RPI Top-35 for the last four years and matched the teams up with their seeds in the NCAA tourney. Here is what I found in tabular and graphical format:

RPI Ranking vs NCAAT Seed

0in 5.4pt 0in 5.4pt'>
 NCAAT RPI Average Seed Range RPI 1 1-7 3 2 4-32 9 3 1-20 11 4 9-25 16 5 13-28 19 6 12-37 22 7 17-47 30

For the next several points, I recommend that you go back and right-click on the link to the graph and open it in a new window so that you can look at the graph and hopefully, the following comments will make some sense.

1) The bounding lines have no mathematical basis. They were simply drawn to encompass the majority of the data.

2) The shorter line in the middle is the linear regression through the average RPI for seeds 1-6. I was not really surprised with the spread in RPI rankings for a given seed, but I was surprised by the quality of the regression of average RPI vs seed. (The linear regression breaks down for the 7th and higher seeds.)

3) In 2004, State received a 3 seed with an RPI of 17. That means that State tied for the second worse RPI for a 3 seed over the last four years. That seeding probably makes sense when you consider that State finished second in the highest rated conference. However, the ACC is not nearly so strong this year.

4) Using the left most bounding line and the averages shown above, we can make some thumb-rules to cover what RPI ranking is generally needed for a given seed:

0in 5.4pt 0in 5.4pt'>
 NCAAT Min RPI Ave RPI Seed Required Required 1 Top-10 3 2 Top-15 9 3 Top-20 11 4 Top-25 16 5 Top-30 19 6 Top-35 22

I have always been surprised by the ability of some State fans to ignore data when they donâ€™t like the conclusion that it brings. So as a final attempt to inject some realism into bracket discussions, here is a short lesson from recent ACC history.

In 2003, the ACC finished the year as the third-rated conference (Sound familiar?). Here is a summary of the top three ACC teams from that year:

0in 5.4pt 0in 5.4pt'>
 Conference Final NCAAT Team Record RPI Seed WF 13-3 7 2 Duke 11-5 12 3 w:st="on">Maryland 11-5 37 6

Note: Duke won the ACCT in 2003.

Those fans expecting to match Duke’s seeding in 2003 with worse credentials this year will probably not be convinced by anything that they read here…but they have been warned.

Be Sociable, Share!

### 18 Responses to “RPI and NCAA Seeds”

01/31/2006 at 8:31 AM #

You are amazing!

Fantastic work!

2. choppack
01/31/2006 at 9:15 AM #

A couple things to point out…First off, no one knows what our final record will be, so it’s quite a stretch to say that some will be angry at our seeding. The analysis is good, but then you make several assumptions.

IMHO, a couple of items that need to be pointed out. In 2003, we didn’t have the marque team that we do this year. If the tourney was held tomorrow, Duke would easily be a No. 1 seed. If NC State, BC, UMd or UNC-Ch go 13-3 or 12-4, they’ll likely be a 2 or 3 seed.

So, if NC STate goes 11-5, wins the tourney, and if doing so, they beat Duke, I would see either a 3 or 4. Likely a 3 though, in spite of the fact that at 11-5 in conference, we’ll have 7 overall losses.

However, it’s a little early to get mad at fans for getting mad at our seating.

3. VaWolf82
01/31/2006 at 9:36 AM #

In 2003, we didnâ€™t have the marque team that we do this year.

What do you mean?

So, if NC STate goes 11-5, wins the tourney, and if doing so, they beat Duke,

Most of the projections (and posters) only assumed a second-place finish…with no mention of ACCT performance, RPI, or beating Duke. Those are the people that I had in mind.

4. Jim
01/31/2006 at 9:55 AM #

I’m far more concerned with how we are playing in mid-March than what exact seed number we get. From the 2nd round on everyone (with the possible exception of some #1s) has a very tough game and has to play well to win. There is so much parity these days that, IMO, too much emphasis is put on seed #s. If we are peaking we will do well. If we aren’t we won’t, regardless of seed.

5. newswolf
01/31/2006 at 10:00 AM #

^Jim I would agree with you EXCEPT I feel that this NC State teams needs to be a 4 seed (for a shot at a final four) so we can play in the Greensboro POD.

That would give us 2 big games in front of our own fans.

I never understood when we lost to Seton Hall people were saying (this game doesn’t matter) I think it cost us 1 line in tournament seeding.

6. RickJ
01/31/2006 at 10:23 AM #

VaWolf82 – awesome stuff & thanks. Just out of curiosity – who was the team with the RPI of 32 that got a 2 seed? I’m guessing a non-BCS conference team with a very good record.

Newswolf – The Greensboro POD issue is one that has not been discussed much. Duke is obviously set for one slot if nothing untoward happens to them. I’m not all that familiar with how POD’s are assigned but I think the principal is that, when possible, the highest seeded teams travel to the nearest location. It would be a big deal if we could finish strong and land in Greensboro but I would have to think with Duke getting one, this is not likely to happen.

7. J.R.
01/31/2006 at 10:24 AM #

It’s not that the Seton Hall didn’t matter. It’s that we lost and people have to find a way to rationalize it, so they just say, Oh no worries. I agree it did cost us a seed spot. I disagree that seeding doesn’t matter. Seeding can make or break a torney run. I see what you’re saying about peaking at the right time, but that is just 1 element of the perfect storm that it would take for NC State to make a Final 4. There is absolutely no excuse for the Pack to ever play as flat as they did against Seton Hall. We acted as if we were one of those teams that have won so much that we became bored with it. I love the way this team bounces back from losses, but unfortunately that is an asset that means nothing in March. I don’t even think that we have a 6 game winning streak this year, and we all know that is the first criteria of a national champion.

8. VaWolf82
01/31/2006 at 10:26 AM #

I think that seeding is a big deal. I think that there are some #1 seeds that State just can not beat, short of a performance like Villinova vs Georgetown in 1985. If State cannot get a 2-3 seed, I would rather have 6-7 seed than a 4-5 seed….especially if State is in the same regional as Duke.

I think it cost us 1 line in tournament seeding.

I’m not so sure….simply because I don’t know what the Selection Committee considers most important when they are seeding teams. I think that a team’s record over the last 10 games is used here…but have not taken the time to prove it.

Here’s the example that makes me think that the last 10 games are important in seeding. Last year, Kansas had the #1 RPI ranking, but got a 3 seed….why? Well, they finished the year on a “slide” at 3-5 with losses to #63 Iowa State and #72 Missouri. This may not be the type of resume that the Selection Committee is looking for when selecting the top seeds. Didn’t Kansas lose early in the tourney last year?

9. VaWolf82
01/31/2006 at 10:33 AM #

Just out of curiosity – who was the team with the RPI of 32 that got a 2 seed?

Oregon in 2002….22-8 (14-4) and lost in the second round of the Pac-10 tourney. I have no rememberance or explanation for this.

I got all of the brackets from ESPN and they only have four years worth of data available. I searched and found another site that broke down the match-ups in the S-16 that year…and that site also listed Oregon as a 2-seed.

10. VaWolf82
01/31/2006 at 10:36 AM #

….and the Pac-10 was the 5th rated conference in 2002. Surprising to say the least.

11. Jim
01/31/2006 at 10:54 AM #

You guys know way more than me about RPIs and historical seeds, etc., so I will defer to you cats on that.

Just in my uneducated opinion, I would rather be peaking and have a #6 or 7 seed than get a #2 or 3 but not be playing well. Aside from the 1s and *maybe* some 2s, the seed #s are not very relevant. Teams that are playing well win, those that aren’t lose.

Obviously I understand that we need to catch a break with the brackets to make a deep run. But IMO this isn’t necessarily a function of a high seed at all. We could be a #6 or 7 with an easier “road” than if we were a #3 in a very difficult bracket. Ease of matchups is fairly random, IMO, unless you are a 1 seed (or an 8 or below, obviously). And that’s not even considering upsets that may or may not happen in front of us. More random noise.

The Seton Hall game sucked the big one. But we bounced back and beat a tough Clemson team on the road. That’s where we are. Maybe Herb can turn that unbelievably crappy showing into some type of positive by using it as a cautionary tale and we can earn back the seed line it cost us.

All of this RPI and seed stuff is interesting to me and I am glad to read it, but I’m just saying let’s not miss the forest for the trees. If this team is playing to potential when we get to the Dance I will take my chances with whatever seeding we get (assuming it’s no lower than 7). To me that’s the far more important thing.

12. class of '74
01/31/2006 at 11:59 AM #

^As Coach V said it’s who you are playing and where you play them that counts. Seeds mean next to nothing after the first round, it’s all about match-ups. I’m presuming a decent finish to the season and easy entry to the NCAAT. I think anyone with good quickness will be our waterloo based on what I’ve seen so far.

13. Jim
01/31/2006 at 12:50 PM #

^ Quick, penetrating guards especially. I agree.

14. BJD95
01/31/2006 at 12:58 PM #

Which teams have that requisite quickness? Texas immediately comes to mind. Fortunately, nobody in the ACC does.

15. class of '74
01/31/2006 at 1:57 PM #

^Clemson does but they shoot free throws like eight yearolds at a YMCA game.

16. choppack
01/31/2006 at 2:02 PM #

Va Wolf – I agree w/ you on seedings. You don’t want 8-9 seeds or 4-5 seeds. Pretty much, the rest is close to a wash-out. It hasn’t been unusual for the NCAA to seed teams from power conferences in that 10 spot – and any team from a power conference will have the potential to do some damage. When we got our 10 seed vs. UNCC last year, I was thrilled – and I guarantee Calhoun wasn’t. Of course, there are no guarantees – I think talent-wise, there’s not a lot of difference between 3-10s in majors anyhoo.