Sunday, February 11, 2018

Mock the selection committee mock

The 1 line:  Villanova, Virginia, Xavier, Kansas

The first 3 are easy. But man I don't know.  Can you leap Michigan St all the way up there?  I almost left Purdue on the 1 line.  Clemson, TTU, and Cincy (and Auburn) all have thin enough resumes.  In the end, Kansas has:  the #1 SoS, 9 Group 1 wins, and an average RPI win of 80.  These are magic bullets.

The 2 line:  Michigan St, Purdue, Clemson, Texas Tech

MSU has concerning metrics which caused them to drop, but winning cures all ailes.  And now with a signature win in tow, here they are.  I wouldn't argue against almost any order of these four teams.  The SoS numbers aren't great for MSU and Tech so I wouldn't be surprised if they wind up on the 3 line.

The 3 line:  Cincinnati, Auburn, Duke, Ohio St

I think these teams are pretty self explanatory.  I wouldn't be surprised if the committee cites SoS numbers and puts Duke on the 2 line.

The 4 line:  North Carolina, Arizona, Rhode Island, Tennessee

All of a sudden, we're running out of quality teams for the back end of this list.  WVU and Oklahoma are punting...Gonzaga and St Mary's aren't quite good enough.  Creighton has SoS issues and Seton Hall is losing games now.  I just don't know who else you can put here except for these teams.

1) Villanova @Pittsburgh
2) Clemson @Charlotte
3) Ohio St @Wichita
4) Rhode Island @Boise

1) Kansas @Wichita
2) Michigan St @Detroit
3) Duke @Pittsburgh
4) Arizona @San Diego

1) Xavier @Nashville
2) Purdue @Detroit
3) Auburn @Dallas
4) North Carolina @San Diego

1) Virginia @Charlotte
2) Texas Tech @Dallas
3) Cincinnati @Nashville
4) Tennessee @Boise

As you can see, a balancing issue persists.  There's no real way to solve it.  The problem?  The ACC has 4 teams in the top 4, meaning someone has to go west.  But Arizona is the only western team, which means North Carolina (sharing the same line as Arizona) can't go west. Therefore it's Duke who has to go, and having a weak 3 seed go in with a weak 1 and 2 seed makes for a soft region.  The other solution is to send Clemson to the west, which really isn't fair to them.

Don't be surprised if the selection committee in March artifically fudges the seeds of a couple ACC teams, to make the regions seem more balanced than they are.  I also wouldn't be surprised if they automatically put #4 and #5 overall in the west, especially if there's a lot of debate over who should be a 1 seed.


Bryan Wilson said...

Projections from me and my computer system:

1: Virginia, Villanova, Xavier, Kansas,
2: Auburn, Clemson, Cincinnati, North Carolina,
3: Texas Tech, Purdue, Duke, Tennessee,
4: Michigan State, Texas A&M, Ohio State, West Virginia

My Projections:
1: Villanova, Virginia, Xavier, Kansas,
2: Purdue, North Carolina, Michigan State, Duke,
3: Clemson, Auburn, Cincinnati, Texas Tech,
4: West Virginia, Tennessee, Ohio State, Arizona

HenryMuto said...

I think 14 of the 16 teams are locks then 2 out of 4 to claim the last 2 spots between (Rhode Island Oklahoma, Arizona and West Virginia). Rhode Island has great RPI but only 6 top 100 equiv wins. Overall 1/2 is a coin flip. #3 Xavier for sure in my book. #4 overall could be 1 of several teams just depends on what the committee looks for. This is why this really helps out for March selections you will know right now how they view teams like MSU, Cincy and Ohio State instead of pure guessing in March which leads to different people having them 2-3 seeds apart. Of course that was the fun in the past and how I have done so well compared to most people and now I lose that edge though last year it sure didn't hurt me.

Ok I didn't study this much so this is much of a gut feel list.

1 seeds - Virginia, Villanova, Xavier, Purdue
2 seeds - Michigan State, Kansas, Auburn, Texas Tech
3 seeds - Clemson, Duke, Cincinnati, Tennessee
4 seeds - North Carolina, Ohio State, Arizona, Rhode Island

Andrew said...

This is actually useful! They just said they moved teams on the 2 line to accommodate balanced regions. That's a big deal.

HenryMuto said...

Well I got the top 4 in exact order the reason I had Virginia over Nova is Virginia has 4 top 10 RPI wins and Nova only had 1 I let that be the difference. Almost everyone I looked at today had Nova over Virginia. I was one of the very few that had Purdue on the 1 line I just felt after looking at the teams despite back to back losses they still had best resume and I got that right.

Cincy a 2 ? Seriously come on committee they really are awarding the "quad 1/2 number wins" over the actual teams they beat another reason why Michigan State has not shown up yet.

I don't agree at all with that I think who you actually beat is more important then an RPI number.

Half way through the show we are

Andrew said...

Ha, they put Cincy on the 2 line.

Auburn at 5 is my one squabble. That feels like a "SEC championship so they must be there" kind of seed.

HenryMuto said...

Went 15/16 I missed the final team I did say 2 of 4 teams would via for last 2 spots got 1 right (Arizona) got 1 wrong (Had Rhode Island who I didn't feel should be a 4 but I didn't think WVU or Oklahoma should get there but I thought it was still possible)

I understand why Michigan State was so low I only pushed them up so high because I thought the committee would fall into the trap of the name and overall ranking and overall record but glad to be wrong as I really felt they didn't deserve a 2 seed based on resume.

Their resume isn't that great. Only 6 quad 1/2 wins which is the lowest of any of the 16 teams that made it.

HenryMuto said...

Michigan State actually has 5 quad 2 wins (they had 3 as of Friday) A couple teams must have played their way up into a quad 2 win for them.

So they had 8 quad 1/2 wins not the 6 I thought they had so that was not as few as I thought.

Their non conf SOS was awful though that might been why so low.

HenryMuto said...

I was 19 off in total number (Lunardi was 35 off ouch)

I got 8 teams exactly right (Lunardi got 2 exactly right)

Bryan Wilson said...

Just finished adjusting weights of metrics in my computer system to try to match the committee's seeding and order. Thought you might be interested to know the results.

The closest I could come to matching the committee basically involved cranking up heavily the computer weight (BPI/Sagarin/whatever computer) and significantly reducing the impact of RPI number and both regular and OOC SOS from previous years.

Some other interesting adjustments I needed to make:
- Road/Neutral record still means almost nothing
- Top 50 wins (or equivalently now, Quadrant 1 wins) are still just as important.
- Marquee wins have much reduced impact.
- OOC SOS means almost nothing, at least for seeding in the top 4 lines.

Interesting to note the first item since that is one of the one things the committee head talked about as being important.

Hard to read too much into this from the small sample size, but definitely my biggest takeaway is the computer metrics are without a doubt having an impact.

I was able to get the seeds correct except for Arizona and Oklahoma on the 4 line replaced by Texas A&M and Wichita. Texas A&M is playing some good basketball recently but I understand if it is taking the committee a bit to catch up. Wichita also is playing well but doesn't have wins to match it.

I also understand putting Arizona in because of leading the Pac 12 and all, but the Pac 12 is terrible and really does not warrant a high bid, especially for dropping 3 games in already. This is more of a historical trust in it being a "power" conference than anything else.

Oklahoma, on the other hand, makes no sense in any universe. After some trying, I was able to produce a contrived set of weights that got Oklahoma to the 4 line. It involved putting an insane emphasis on their marquee road wins (2-0 against top 30 on the road) but in the collateral damage it put Florida on the 3 line among other bad things. Oklahoma is just not good recently. Maybe the committee had a blackout on Oklahoma games over the last 3 weeks and did not watch any of them.