Objective Quadrant Analysis

[Consigliere]

Thrilled to get win number 17 yesterday. Got to be honest - I've feared at a few points this season that our streak of 48 straight winning seasons (second in D1 history only to UCLA's 54)was in jeopardy. Yesterday's win clinches number 49.

Time to start looking toward the next goals for this year's Orange - winning 20 and making the tournament. Lots of bracketology projections out there and while there is a lot of subjectivity in the analyses of Lunardi and others, the new quadrant system provides the opportunity to take a more objective approach. I've taken the 16 teams in Lunardi's bubble (Last 4 byes, Last 4 in, first four out, next four out) and applied just a little math. Award 4 points for a Q1 win, 3 for a Q2, 2 for a Qr and 1 for a Q1. Similarly subtract 1 point for a Q1 loss down to -4 for a Q4 loss. Now obviously this can't be an absolute. Any system that equates a home win against Buffalo with a road win at Villanova certainly has flaws, but it provides a starting point.

Using this methodology, Syracuse ranks sixth of the sixteen teams. According to Lunardi's bracket that's the second biggest snub behind Mississippi State. Washington, conversely gets the biggest bump. With the recent big wins Lunardi has the Huskies as a last four bye while they rank no better than ninth by the numbers.

Nothing more than a conversation starter at this point but here is the complete analysis:



Lunardi Team Total
LFB Arkansas 30
LFB Virginia Tech 27
LFI Kansas State 27
LFI UCLA 24
NFO Mississippi State 24
FFO Syracuse 22
LFB Providence 22
FFO Boise State 21
LFB Washington 20
LFI USC 20
FFO Nebraska 20
LFI NC State 18
FFO St. Bonaventure 17
FFO Utah 16
NFO Temple 14
NFO Western Kentucky 9

 

[Consigliere]

BTW, I'm not completely sure of my data source but it seems more closely aligned to the road / home view of quadrant ratings than others:

College Men's Basketball Standings (D1)

 

[pearl31]

interesting - thanks!

 

[jncuse]

An interesting way to try to turn the quad records into one number, but I don't think it work nor would it be the way they look at it.

Here is the major flaw with your analysis.
A team that is 1-3 in Quad 1 is viewed equally to a team that is 5-7 in Quad One. And the committee will certainly value the team at 5-7 much more.

 

[Consigliere]

An interesting way to try to turn the quad records into one number, but I don't think it work nor would it be the way they look at it.

Here is the major flaw with your analysis.
A team that is 1-3 in Quad 1 is viewed equally to a team that is 5-7 in Quad One. And the committee will certainly value the team at 5-7 much more.

Absolutely true. Hence my disclaimer - "Any system that equates a home win against Buffalo with a road win at Villanova certainly has flaws, but it provides a starting point. "

But to be clear - A team that is 1 - 3 in Q1 would receive 4 points for the win and lose 3 for the losses (in Q1 games only) netting 1 point. A team that is 5-7 would get 20 for the wins (5 wins * 4 points) and lose 7 (7 losses * 1 point) netting 13.

 

[NineOneSeven]

An interesting way to try to turn the quad records into one number, but I don't think it work nor would it be the way they look at it.

Here is the major flaw with your analysis.
A team that is 1-3 in Quad 1 is viewed equally to a team that is 5-7 in Quad One. And the committee will certainly value the team at 5-7 much more.

there is a way to work around this, and we should all take the rest of the week off from work and figure it out.

 

[Knicks411]

I think the quad system is an improvement than the old way of just using top 50 wins or whatever, by adjusting for home/road/neutral. It's a start, it's better than what we had, no doubt about that. But it will still treat a road win against #75 as the same as a road win against #1. Granted, you need to draw the line somewhere, I get it, I just wanted to emphasize that point.

 

[Trueblue25]

Time to give nc st le boot

 

[jncuse]

Absolutely true. Hence my disclaimer - "Any system that equates a home win against Buffalo with a road win at Villanova certainly has flaws, but it provides a starting point. "

But to be clear - A team that is 1 - 3 in Q1 would receive 4 points for the win and lose 3 for the losses (in Q1 games only) netting 1 point. A team that is 5-7 would get 20 for the wins (5 wins * 4 points) and lose 7 (7 losses * 1 point) netting 13.

Actually, I misread your post. I thought in Quad 1, you were giving 4 for a win, and -4 for a loss. I was going to suggest inverting the loss value to make it 4 for a W, and -1 for a L.

But now that I re-read your post that is what you are doing. I think it's a good predictive metric at a high level.

May I suggest one tweak to the Consigliere Tracking Model. Elite Wins. The Committee use to also emphasize top 25 wins. Maybe add 3 bonus points for each Top 15 home or Top 30 road win.

 

[Knicks411]

Actually, I misread your post. I thought in Quad 1, you were giving 4 for a win, and -4 for a loss. I was going to suggest inverting the loss value to make it 4 for a W, and -1 for a L.

But now that I re-read your post that is what you are doing. I think it's a good predictive metric at a high level.

May I suggest one tweak to the Consigliere Tracking Model. Elite Wins. The Committee use to also emphasize top 25 wins. Maybe add 3 bonus points for each Top 15 home or Top 30 road win.

Yeah i think this is a good idea, anyway to differentiate a little at the top level with wins.

 

[Consigliere]

Actually, I misread your post. I thought in Quad 1, you were giving 4 for a win, and -4 for a loss. I was going to suggest inverting the loss value to make it 4 for a W, and -1 for a L.

But now that I re-read your post that is what you are doing. I think it's a good predictive metric at a high level.

May I suggest one tweak to the Consigliere Tracking Model. Elite Wins. The Committee use to also emphasize top 25 wins. Maybe add 3 bonus points for each Top 15 home or Top 30 road win.

I like that tweak. If I update next week I made roll that in. My initial thought was that the quadrants were not granular enough - 4 categories for 330+ teams doesn't really give you the differentiation you need - but I balanced that with the easy availability of data.

 

[SBU72]

What is a QUAD.?

 

[Dave85]

An interesting way to try to turn the quad records into one number, but I don't think it work nor would it be the way they look at it.

Here is the major flaw with your analysis.
A team that is 1-3 in Quad 1 is viewed equally to a team that is 5-7 in Quad One. And the committee will certainly value the team at 5-7 much more.

You have some bandwidth in your brain's processor!

 

[Dave85]

What is a QUAD.?

And you don't.


{{{just kidding}}}

 

[Dave85]

Time to give nc st le boot

18!

 

[BuffOrange]

I think the quad system is an improvement than the old way of just using top 50 wins or whatever, by adjusting for home/road/neutral. It's a start, it's better than what we had, no doubt about that. But it will still treat a road win against #75 as the same as a road win against #1. Granted, you need to draw the line somewhere, I get it, I just wanted to emphasize that point.

Better than the crappy old way notwithstanding, it's ridiculous we still have these discussions in 2018.
Gary Parrish who I have ripped before, made so much sense on the cbs podcast today my ears almost fell off (about resumes, schedules, rankings, everything). I say put him in charge of seeding and move on with our lives.

 

[Knicks411]

Better than the crappy old way notwithstanding, it's ridiculous we still have these discussions in 2018.
Gary Parrish who I have ripped before, made so much sense on the cbs podcast today my ears almost fell off (about resumes, schedules, rankings, everything). I say put him in charge of seeding and move on with our lives.

I usually think Parrish has pretty good takes about this kind of stuff. I haven't listened to the podcast in a while, maybe I'll fire it up

 

[orange79]

What is a QUAD.?

In an attempt to improve on the Top 25/Top 50/Top 100/etc. breakdown of wins based on RPI, the NCAA Selection Committee has developed quadrants that vary by the site of the game (home/away/neutral).

This article explains it pretty well.

NCAA selection committee adjusts team sheets, emphasizing road/neutral games more than ever

Quadrant 1: Home 1-30; Neutral 1-50; Away 1-75
Quadrant 2: Home 31-75; Neutral 51-100; Away 76-135
Quadrant 3: Home 76-160; Neutral 101-200; Away 136-240
Quadrant 4: Home 161-plus; Neutral 201-plus; Away 241-plus

 

[pearl31]

quadrants seem like a step in the right direction... what about bad losses, though - have they stated how those are quantified now? Are they quantified?

 

[orange79]

quadrants seem like a step in the right direction... what about bad losses, though - have they stated how those are quantified now? Are they quantified?

You can quantify them with the same criteria.

For example, after yesterday's games, here are SU's losses by quadrant:
Q1 losses = 4
Q2 losses = 3
Q3 losses = 1
Q4 losses = 0

So, no really bad losses (Q4), and only 1 'bad' loss (Q3 - Ga Tech).

Unless I misinterpreted your question.

 

[pearl31]

You can quantify them with the same criteria.

For example, after yesterday's games, here are SU's losses by quadrant:
Q1 losses = 4
Q2 losses = 3
Q3 losses = 1
Q4 losses = 0

So, no really bad losses (Q4), and only 1 'bad' loss (Q3 - Ga Tech).

Unless I misinterpreted your question.

Nope, you didn't misinterpret - just that I hadn't heard/seen anything stating they're definitely going to treat losses in an exact converse manner to wins

 

[jncuse]

quadrants seem like a step in the right direction... what about bad losses, though - have they stated how those are quantified now? Are they quantified?

I would think bad losses are Quad 3 and Quad 4 bad losses... which was 101-200, 200+ in the past. Not many teams on the bubble have quad 4 losses... Providence does though.

 

[jncuse]

You can quantify them with the same criteria.

For example, after yesterday's games, here are SU's losses by quadrant:
Q1 losses = 4
Q2 losses = 3
Q3 losses = 1
Q4 losses = 0

So, no really bad losses (Q4), and only 1 'bad' loss (Q3 - Ga Tech).

Unless I misinterpreted your question.

I agree with this, at the same never heard it confirmed. But it would seem odd not to use this.

 

[orange79]

Nope, you didn't misinterpret - just that I hadn't heard/seen anything stating they're definitely going to treat losses in an exact converse manner to wins

Fair question. I haven't seen it stated explicitly either, but if you're going to use it for wins, you might as well use it for losses. But I think looking at bad losses probably only comes into play when you are looking at teams that are pretty much the same in terms of Q wins.

 

[pearl31]

just seems like there's been a fuzzy, undefined realm of how a teams wins are weighted against their losses - seems like they're in no hurry to make that aspect more objective or transparent...
(i.e., that remains a behind-closed-doors "judgement call" by certain louder/more-influential committee members)