Bubble Watch - Week of February 25th to March 3rd

[jncuse]

SEE POST #9 for the Saturday Schedule by Time

Current bubble picture with change since last week. Based on 40 current brackets on the bracket matrix.

Moved In - Marquette, Providence
Moved Out - TCU, Clemson

Positive Movers

Marquette went from 35% to 98% by beating Xavier and St. John's at home. They were fairly close to the line last week (2nd last out). This week Marquette plays at Providence and at Xavier. Theoretically by next week they can be back out.

Providence made a major move from 15% to 73%, by getting a huge road win at Creighton. This week Providence plays Marquette and Depaul at home, they could potentially be at about 100% at the end of the week.

The other major mover was Vanderbilt from 0%to 38% based on beating South Carolina and then a road win at a fellow bubble team in Tennessee.

Negative Movers

TCU lost at Iowa St and at Kansas and moved down from 89% to 38%. Sort of a tough week that they can easily bounce back from by beating West Virginia and Kansas St at home this week.

Similar to TCU, Clemson fell from 54% to 8% after losing at Virginia and at Miami. These week they have 2 home games to move up including Florida St.

Syracuse - held stable this week after losing at Tech and beating Duke

 

[jncuse]

Primary Bubble - Games this week

I will track the overall record of these games over the weekend

 

[jncuse]

Secondary Bubble - Teams that are not getting any current mentions on the matrix, but could be in or right near the line if they win out.

Pitt
Indiana
Georgetown
Ole Miss
Tennessee

Teams that are clearly in now but probably need to win at least one more game (Arkansas being in the SEC may need more)

USC
Arkansas
Michigan
VCU
Miami
Northwestern

 

[pearl31]

I think we want Marquette to lose two, even if one of those benefits Provy...
it's nice (as long as we take care of business) to have the GT game serve as a sort of "two-birds-with-one-stone" game for us...
I'm betting MichSt goes 1-2 (home Wisc, at Ill and at Maryland)...
Cal has @ Utah and @ Colorado remaining - "should-win" games but still tricky (no real benefit in winning, sinks them if they lose)...
Vandy still has to play Fla and UK, likely losses which would put them at 15 losses before the conf tourney...
who do we pull for in that Kst-TCU game on Wed?

 

[pearl31]

Also, Michigan could easily finish the season 1-2 or even 0-3 as they have home Purdue, at NW, and at Nebraska - that would bring them right back to the cut line

 

[jncuse]

Also, Michigan could easily finish the season 1-2 or even 0-3 as they have home Purdue, at NW, and at Nebraska - that would bring them right back to the cut line

I have them in something I called the secondary bubble. Teams that probably need to win at least one more.

 

[Briancuse]

Secondary Bubble - Teams that are not getting any current mentions on the matrix, but could be in or right near the line if they win out.

Pitt
Indiana
Georgetown
Ole Miss
Tennessee

Teams that are clearly in now but probably need to win at least one more game (Arkansas being in the SEC may need more)

USC
Arkansas
Michigan
VCU
Miami
Northwestern

Not knocking you at all, I just don't get the Pitt thing, how many conference games do they have to lose before they are out of the discussion?

 

[jncuse]

Not knocking you at all, I just don't get the Pitt thing, how many conference games do they have to lose before they are out of the discussion?

None. I think they are down to their last chance. What will move them up quickly is their 3 last games would all be quality wins. Based on Sagarin their chance of winning out the regular season is about 1%... so we should be fine.

 

[jncuse]

Saturday Viewing Schedule

Primary - teams that are right around the bubble and should be the focus tomorrow
Secondary (more longshots) - teams that need to win out or lose out to be on the bubble, So much less likely to move in or move out.

Primary Games (want bold team lose)
12:00 Florida St at Clemson
2:00 VCU at Rhode Island
2:00 West Virginia at TCU
2:00 Illinois St at Northern Iowa
2:00 Seton Hall at Depaul
4:00 Miss St at Vanderbilt
4:00 Marquette at Providence
6:00 Kansas St at Oklahoma


Secondary Games (want bold team to lose)
12:00 UNC at Pitt
12:00 Georgetown at St. John's
1:00 Tennessee at South Carolina
2:00 Texas Tech at Oklahoma St
3:30 Missouri at Ole Miss
4:00 Purdue at Michigan
6:00 LSU at Georgia
8:00 Northwestern at Indiana
8:30 Arkansas at Auburn

 

[jncuse]

I think we want Marquette to lose two, even if one of those benefits Provy...
it's nice (as long as we take care of business) to have the GT game serve as a sort of "two-birds-with-one-stone" game for us...
I'm betting MichSt goes 1-2 (home Wisc, at Ill and at Maryland)...
Cal has @ Utah and @ Colorado remaining - "should-win" games but still tricky (no real benefit in winning, sinks them if they lose)...
Vandy still has to play Fla and UK, likely losses which would put them at 15 losses before the conf tourney...
who do we pull for in that Kst-TCU game on Wed?

I went for Marquette losing as well. I think they have more to gain by winning. I think the Kansas St vs TCU game is a total toss up as well. Probably would prefer the home team winning.

I think the game of the day tomorrow in terms of the bubble and it's impact on us is West Virginia at TCU. A winnable top 25 game for them at home.

 

[Niastri]

Thank you for these posts, jncuse. I find they are the highlight of this board for me right now. It is a lot of work and your attention to detail is very good. Nice job

 

[longislandcuse]

These are awesome breakdowns. Thank you for doing this each week!

 

[orangenirvana]

Cal is not making the Tourney with a 1-7 record against the Top 50.

Looking at these lists, it's obvious that too many of these (lazy) "bracketologists" are still putting way too much weight on RPI.

USC, for instance, has Sagarin/KenPom/BPI ranks of 58/63/63 with a measly 2-6 record vs. the Top 50. Both wins coming at home. Why would they ever be considered a lock?? The only explanation is that they have an RPI of 30 (for some reason).

Exactly why I don't trust the bracket matrix.

 

[orangenirvana]

Not knocking you at all, I just don't get the Pitt thing, how many conference games do they have to lose before they are out of the discussion?

Pitt's three wins vs. the Top 25 (which includes a road win at Maryland) keeps them in the hunt. Conference record is not relevant.

 

[RMH44]

Cal is not making the Tourney with a 1-7 record against the Top 50.

Looking at these lists, it's obvious that too many of these (lazy) "bracketologists" are still putting way too much weight on RPI.

USC, for instance, has Sagarin/KenPom/BPI ranks of 58/63/63 with a measly 2-6 record vs. the Top 50. Both wins coming at home. Why would they ever be considered a lock?? The only explanation is that they have an RPI of 30 (for some reason).

Exactly why I don't trust the bracket matrix.

Serious question: does the committee still use how a team is playing down the stretch (last 10 games?) as part of their selection process?

 

[omniorange]

jncuse, just want to say thank you for these excellent posts/threads. I know a lot work went into them and appreciate your thoughts on this.

Cheers,
Neil

 

[orangenirvana]

Serious question: does the committee still use how a team is playing down the stretch (last 10 games?) as part of their selection process?

I think they got rid of that criteria. Which is the right thing, imo.

 

[orangenirvana]

Going into today, I'd rank the 30 bubble teams like this...

Cumulative computer ranking / Record vs. RPI Top 50 (Top 25)

1. Miami - 30 / 3-7 (2-5)
2. Marquette - 32/31/29/69 = 40 / 5-5 (2-3)
3. Michigan - 27/27/25/53 = 33 / 3-7 (1-3)
4. Syracuse - 48 / 6-6 (3-3)
5. Providence - 59/56/53/55 = 56 / 6-7 (2-5)
6. Seton Hall - 52/57/58/48 = 54 / 4-6 (1-5)
7. Michigan State - 42/52/50/43 = 47 / 4-6 (2-6)
8. Kansas State - 35/30/39/61 = 41 / 3-8 (1-4)
9. VCU - 44/42/38/25 = 37 / 2-1 (1-1)
10. Northwestern - 39/39/44/45 = 42 / 3-6 (1-5)
11. Vanderbilt - 53/51/54/46 = 51 / 4-6 (1-4)
12. Clemson - 44 / 4-10 (0-7)
13. Arkansas - 45/49/49/35 = 45 / 3-5 (0-3)
14. Virginia Tech - 45 / 2-6 (2-5)
15. TCU - 45/43/51/54 = 48 / 2-9 (0-5)
16. Georgia Tech - 86 / 4-6 (4-3)
17. Pittsburgh - 62 / 3-9 (3-6)
18. Illinois - 67/69/69/57 = 66 / 3-7 (1-5)
19. Houston - 47/40/37/58 = 46 / 2-4 (0-3)
20. Rhode Island - 54/55/41/47 = 49 / 1-3 (1-3)
21. Indiana - 31/46/30/101 = 52 / 3-8 (2-5)
22. USC - 58/63/63/30 = 54 / 2-6 (2-5)
23. Georgetown - 51/58/52/79 = 60 / 3-9 (2-5)
24. Illinois State - 56/54/66/33 = 52 / 1-1 (0-0)
25. Texas Tech - 37/35/35/100 = 52 / 2-7 (1-3)
26. Tennessee - 55/48/56/64 = 56 / 2-9 (1-5)
27. California - 52/44/45/51 = 48 / 1-7 (0-6)
28. Georgia - 49/45/64/50 = 52 / 1-7 (0-5)
29. Wake Forest - 37 / 1-9 (0-8)
30. Mississippi - 70/74/78/66 = 72 / 1-8 (0-3)

 

[jncuse]

Cal is not making the Tourney with a 1-7 record against the Top 50.

Looking at these lists, it's obvious that too many of these (lazy) "bracketologists" are still putting way too much weight on RPI.

USC, for instance, has Sagarin/KenPom/BPI ranks of 58/63/63 with a measly 2-6 record vs. the Top 50. Both wins coming at home. Why would they ever be considered a lock?? The only explanation is that they have an RPI of 30 (for some reason).

Exactly why I don't trust the bracket matrix.

I use it as a guide it's great to assess general position. In my view it's worth tracking but at the same time should always be open to questioning like you are doing especially with the openness to analytics.

I tend to agree there is a bit an over-reliance on RPI on a few teams and I am not making those teams locks. But I do think they are both in as of now, due the underwhelming nature of the last teams out. They need to be careful however.

At the same time there is a tonne of P5 teams with terrible RPI's that are getting some appreciation on the matrix. And nobody is putting the mid majors with high RPI's and empty brackets in the tournament. So I don't think it is as RPI crazy as you mention.

 

[jncuse]

Going into today, I'd rank the 30 bubble teams like this...

Cumulative computer ranking / Record vs. RPI Top 50 (Top 25)

1. Miami - 30 / 3-7 (2-5)
2. Marquette - 32/31/29/69 = 40 / 5-5 (2-3)
3. Michigan - 27/27/25/53 = 33 / 3-7 (1-3)
4. Syracuse - 48 / 6-6 (3-3)
5. Providence - 59/56/53/55 = 56 / 6-7 (2-5)
6. Seton Hall - 52/57/58/48 = 54 / 4-6 (1-5)
7. Michigan State - 42/52/50/43 = 47 / 4-6 (2-6)
8. Kansas State - 35/30/39/61 = 41 / 3-8 (1-4)
9. VCU - 44/42/38/25 = 37 / 2-1 (1-1)
10. Northwestern - 39/39/44/45 = 42 / 3-6 (1-5)
11. Vanderbilt - 53/51/54/46 = 51 / 4-6 (1-4)
12. Clemson - 44 / 4-10 (0-7)
13. Arkansas - 45/49/49/35 = 45 / 3-5 (0-3)
14. Virginia Tech - 45 / 2-6 (2-5)
15. TCU - 45/43/51/54 = 48 / 2-9 (0-5)
16. Georgia Tech - 86 / 4-6 (4-3)
17. Pittsburgh - 62 / 3-9 (3-6)
18. Illinois - 67/69/69/57 = 66 / 3-7 (1-5)
19. Houston - 47/40/37/58 = 46 / 2-4 (0-3)
20. Rhode Island - 54/55/41/47 = 49 / 1-3 (1-3)
21. Indiana - 31/46/30/101 = 52 / 3-8 (2-5)
22. USC - 58/63/63/30 = 54 / 2-6 (2-5)
23. Georgetown - 51/58/52/79 = 60 / 3-9 (2-5)
24. Illinois State - 56/54/66/33 = 52 / 1-1 (0-0)
25. Texas Tech - 37/35/35/100 = 52 / 2-7 (1-3)
26. Tennessee - 55/48/56/64 = 56 / 2-9 (1-5)
27. California - 52/44/45/51 = 48 / 1-7 (0-6)
28. Georgia - 49/45/64/50 = 52 / 1-7 (0-5)
29. Wake Forest - 37 / 1-9 (0-8)
30. Mississippi - 70/74/78/66 = 72 / 1-8 (0-3)

You should add an in/out line to add some more context to the post, which you put a lot of effort into.

You nailed the biggest primary factor "Quality Wins" that the committee uses but two other primary factors of their list of five include
1) Performance away from Home
2) Non-Conference SOS

I am working off memory which may be flawed but in order the five were shown as follows I believe:

1) Quality Wins
2) Performance away from home
3) or 4) Non-Conference SOS / Individual team Rankings
5) Head to Head Matchuos of Bubble Teams.

 

[nycfan]

Thank you for these posts, jncuse. I find they are the highlight of this board for me right now. It is a lot of work and your attention to detail is very good. Nice job

Yep, was just gonna post the same thing---it's definitely the highlight for me and the first thread I always go to. And then these new weekly ones on Saturday are a bit like Christmas morning--if these aren't the first thing you begin your weekend with then I don't know what to tell you. And I also really love the Saturday afternoon time schedule (with teams in bold) for viewing purposes. My day is now set, so thanks again jncuse! (I think it may actually be the first time I have thanked you this season, but have appreciated all the time you put into this the past 2 years).

 

[jncuse]

I created an in/out line to add some context to Orange Nirvana's post. While I don't think it fully captures all factors, it does weigh the biggest factor they have used, and also incorporates team rankings quite highly... and as I have mentioned a few times if they get rid of the RPI for other numbers, it inherently means they didn't trust the old individual numbers but trust the new numbers. So those rankings could mean much more this year.

It's a relevant concept with supportable data.

100% Created by OrangeNirvana (simply added an In/Out Line)

Cumulative computer ranking / Record vs. RPI Top 50 (Top 25)

1. Miami - 30 / 3-7 (2-5)
2. Marquette - 32/31/29/69 = 40 / 5-5 (2-3)
3. Michigan - 27/27/25/53 = 33 / 3-7 (1-3)
4. Syracuse - 48 / 6-6 (3-3)
5. Providence - 59/56/53/55 = 56 / 6-7 (2-5)
6. Seton Hall - 52/57/58/48 = 54 / 4-6 (1-5)
7. Michigan State - 42/52/50/43 = 47 / 4-6 (2-6)
8. Kansas State - 35/30/39/61 = 41 / 3-8 (1-4)
9. VCU - 44/42/38/25 = 37 / 2-1 (1-1)
10. Northwestern - 39/39/44/45 = 42 / 3-6 (1-5)
11. Vanderbilt - 53/51/54/46 = 51 / 4-6 (1-4)
12. Clemson - 44 / 4-10 (0-7)
13. Arkansas - 45/49/49/35 = 45 / 3-5 (0-3)
14. Virginia Tech - 45 / 2-6 (2-5)

IN/OUT LINE

15. TCU - 45/43/51/54 = 48 / 2-9 (0-5)
16. Georgia Tech - 86 / 4-6 (4-3)
17. Pittsburgh - 62 / 3-9 (3-6)
18. Illinois - 67/69/69/57 = 66 / 3-7 (1-5)
19. Houston - 47/40/37/58 = 46 / 2-4 (0-3)
20. Rhode Island - 54/55/41/47 = 49 / 1-3 (1-3)
21. Indiana - 31/46/30/101 = 52 / 3-8 (2-5)
22. USC - 58/63/63/30 = 54 / 2-6 (2-5)
23. Georgetown - 51/58/52/79 = 60 / 3-9 (2-5)
24. Illinois State - 56/54/66/33 = 52 / 1-1 (0-0)
25. Texas Tech - 37/35/35/100 = 52 / 2-7 (1-3)
26. Tennessee - 55/48/56/64 = 56 / 2-9 (1-5)
27. California - 52/44/45/51 = 48 / 1-7 (0-6)
28. Georgia - 49/45/64/50 = 52 / 1-7 (0-5)
29. Wake Forest - 37 / 1-9 (0-8)
30. Mississippi - 70/74/78/66 = 72 / 1-8 (0-3)

My Analysis

Compared to the Matrix he has Vanderbilt and Clemson in instead of California and USC. As he mentioned above the big reason is P12 ranking and he thinks people are overrating their RPI.

Teams like Pitt and Illinois are all much closer to being in his view. And Northwestern and Virginia Tech are much closer to being out than portrayed.

Your conclusions seem reasonable above the line which matters the most. It does seem to break down a bit after the line in my view because most teams below the line have crappy top 50 records and are hard to separate,. But above the line is what matters the most.

Let's see how this bears out as we get closer to Selection Sunday.

 

[Briancuse]

Pitt's three wins vs. the Top 25 (which includes a road win at Maryland) keeps them in the hunt. Conference record is not relevant.

Of course its relevant.How is it not relevant?

 

[jncuse]

Of course its relevant.How is it not relevant?

It's relevant and not relevant. Let me explain.

How you perform in the ACC is relevant to your overall results of 30-35 games. But it is not a subset of 18 games that matters to the committee. They don't give two craps about that 18 game record in comparing teams.

Consider your results as about 32 individual games. They use subsets of those 32 games for analysis such as top 50 wins, top 50 performance, top 100 road wins, road wins, OOC SOS. They also give your team an individual ranking based on 32 games -- not 18 games. They are extremely explicit that conference record does not matter as a metric or subset of information.

Of course if you do well in the ACC or poorly in the ACC you are more likely to do better or worse in the relevant metrics or subsets. But it is not a 100% correlation either. Which is why you see teams like Clemson and Pitt linger.

 

[Louie and Bouie]

Of course its relevant.How is it not relevant?

It's relevance is extremely diminished with an unbalanced conference schedule. 8-8 does not equal 8-8 in this conference.

We shouldn't forget Syracuse had one of the easiest ACC conference schedules.