Sunday, January 21, 2007

W-7 BP65

Pretty amazing that there are only seven weeks until Selection Sunday. The first conference tournaments are only about 5 weeks away. Anyway, as per usual, the BP65:

1. UCLA (PAC 10)
1. OHIO ST. (BIG 10)

2. Wisconsin
2. KANSAS (BIG 12)
2. Texas A&M

3. Arizona
3. Oklahoma State
3. Duke
3. Oregon

4. LSU
4. Alabama
4. Texas

5. Tennessee
5. Clemson

6. USC
6. Kentucky
6. Marquette
6. Georgetown

7. Washington State
7. Boston College
7. Illinois

8. Northern Iowa
8. West Virginia
8. Maryland
8. Michigan State

9. Indiana
9. Missouri State
9. Notre Dame

10. Georgia Tech
10. Creighton
10. California

11. UConn
11. Texas Tech

12. Florida State
12. Arkansas
12. Louisville
12. UNLV


14. PENN (IVY)




Anonymous said...

illinois a 7 seed? Come on Man. Wake up. They aren't even worthy for the Big Dance.

Jeff said...

Remember, this isn't what their seed would be if the season ended now, but what it will be at the end of the year. Illinois has a very good strength of schedule that has their RPI at 50 even though they only have a record of 14-7. That record should pick up, because they have a pretty easy schedule the rest of the way. They avoid Wisconsin and Ohio State again (only had to play each one once this year), and they only have to play Michigan State at home. Meanwhile, they get Northwestern twice, and Minnesota, Indiana, Michigan and Michigan State at home. They could easily end up 10-6, although I see them more likely at 9-7, with 22-10 overall record and an RPI in the 30's. That should get them a 7 seed or so.

Anonymous said...

Jeff, you must be the only guy that does there bracket based on what it will be in March.

Most everyone else I have seen does it based on TODAY. As if the season were to finish today, this is what the brackets would look like.

Illinois is now 2-7 against the top 50.

12 of their wins are vs. teams ranked 100 or higher.

Jeff said...

Well I don't think it takes much skill to go through the current Top 25 and divide by four. The only thing that really matters is what the brackets will look like when they actually exist, so that's what I care about.

D. Moore said...

I like the bracket being an actual prediction of what will be. It gives us more to think about, and there are already plenty of brackets out there that tell us where things stand so far.

And, we're getting to the meat of the season where the two will converge, anyway.

I've found it easy this season to find fault with teams that are picked. However, it just feels wrong to shred the picks without offering better alternatives, and finding worthy teams that aren't in the bracket is becoming more and more difficult for me. Would you consider adding a list of the teams that you considered, but didn't quite make the cut?

I suspect that the teams that replace those that falter will be teams that surge unexpectedly, but it would be interesting to note those that look like they might be starting to make a run to the NCAAs (Vanderbilt?).

D. Moore said...

OK, I found a missing team. Now I can complain :^).

Connecticut. I still don't see it. They have only one road win for the entire season, are 13-6 so far (2-4 in conference), and have 6 home games remaining and 5 away (4 in conference). They look to me like they will go 6-5 in the rest of the regular season (6-4 in conference, for 8-8 total, around 9th place). They won't have a single top 50 win at the end of the season, unless they can beat Villanova at home, or GA Tech or Georgetown on the road (and hope they stay in the top 50 -- Syracuse won't be there any more). Their RPI is 87, and I don't see how it will rise enough to compare well with other candidates.

The surging team to replace them with is Virginia Tech. They have bad early season losses, but have marquee wins to compensate, and should end up at 10-6 in conference. Their RPI is 32, and shouldn't slip much.

Jeff said...

Sure, I'll make an effort to put a short list of "the best of the rest". Vanderbilt is a surging team to consider, so is Virginia Tech, so is Stanford. There are several to consider.

As for UConn, I understand what you're saying. But I look at their schedule right now and say that they could easily win their next five conference games. Providence, at Depaul, Rutgers, Syracuse, Seton Hall (and throw in a potential resume-building game at Georgia Tech in the middle of it all). If they win those games, they're 7-4 in conference without even that hard of a schedule after that. Although I'll grant you that if they lose at home to Providence this weekend it will be their 5th loss in the Big East, and they'll join Washington on the "They have a lot of talent, but just too many conference losses to be in the field right now" list.

Jeff said...

Another thing to keep in mind for UConn are the Sagarin ratings. As most people know, he has an Elo-Chess rating which only cares about wins and losses and ends up mostly lining up with the RPI. He also has a Predictor rating that only cares about scores, and he considers this a much better predictor of future success (hence the name). The combination of these two averages gives his overall rankings, which are actually a better measure of likelihood of Tournament selection than the RPI. The selection committee will pass over teams with high RPIs and bad Sagarin ratings more than they will teams with high Sagarin ratings and questionable RPIs. You can see this over the last few years.

So, look at UConn. They have an Elo-Chess of 82nd in the country, which is right in line with their RPI. Their Predictor is an amazingly high 20. In other words, they have lost a LOT of very close games. Combine those ratings together and you have a total Sagarin rating of 54. Assuming a little improvement from here on out and UConn can easily expect a Top 40 Sagarin rating, which is usually enough to get in.

A good example of the opposite of this was Notre Dame football in the fall (sorry to you Notre Dame fans). They kept winning their games, so they kept moving up in the Elo-Chess ratings (and all BCS computer rankings, which aren't allowed take into account score). But the Predictor noticed that close wins over awful teams signified the likelihood of future failure. Right before their game against USC, their Elo-Chess was around 4th in the country while their Predictor was around 25th. I immediately put down money on USC to cover (I think I only had to give 13.5).

When in doubt, the Sagarin ratings a great objective indicator of a team whose on-field success has not yet correlated with its actual abilities.

For the record, the Sagarin ratings suggest that the following teams are really underperforming so far: Maryland, Virginia, Villanova, West Virginia, Purdue, Texas, Oklahoma, Mississippi State. These teams, like UConn, can be expected to improve their resumes in the future. The next set of teams, however, has over-performed and should be expected to cool off: Texas Tech, Air Force, UNLV, Oregon, Washington State, Stanford, Nevada.

It's not a perfect predictor, of course, but it's a good clue.

D. Moore said...

First, UConn. Yes, they could turn it around, and their remaining schedule is weak enough that they certainly could. Part of my point, though, is that their remaining schedule is so weak that if they do well against it, they can prevent those opponents from being top 50 teams, and will thus have no solid wins on their resume. And if they don't do well against that schedule, then they'll just suck. They desperately need wins against the few strong teams remaining on their schedule, and they need road wins.

The Sagarin rating is very interesting. I can see cases where this would be a great predictor, and cases where it would be very misleading. Teams that lose a lot of close games could be unlucky, or they could be teams that can't shoot free throws and have poor leadership. Teams that win a lot of close games could have been lucky (Oregon's last second shots?), or they could be teams that play a style that leads to low scoring games, and have senior leaders who know how to win at the end (Air Force?).

Anonymous said...

I still don't understand how you believe Ohio State is a better team than Wisconsin. If Wisconsin wins at Iowa and Indiana this week, will you bump them up to a #1 seed? Winning those 2 games would nearly guarantee a 27-1 record going into the @MSU, @OSU, MSU stretch, where they will likely win at least 2 games. Do you really think it will be possible for a team going, say, 31-3 with a top 5 RPI to be denied a #1 seed?

Jeff said...

I don't think Ohio State is better than Wisconsin. If I was on the Selection Committee, I'd give Wisconsin that #1 seed. But the way it's probably going down is that UNC, Florida & UCLA are getting three of the #1 seeds. The other one will go to the winner of the Big 10 Tournament. I think Wisconsin will win the regular season Big 10 crown, but my prediction is that Ohio State will upset them in the Big 10 Tourney finals. That will give them #1 seed, despite what happened during the regular season.