Bracketology (Week of Feb. 15th)

Status
Not open for further replies.
#5      
Is putting them 6th in the country not liking them?

Let's be honest, no one is even close to Baylor and Gonzaga, all the teams that have been in the 3-7ish range this year have been overrated, and the results have reflected that.

I think outside of the top 4 for them at this moment, particularly with a win over us and Michigan not playing for 3 weeks, is not liking them.
 
#6      
I think outside of the top 4 for them at this moment, particularly with a win over us and Michigan not playing for 3 weeks, is not liking them.
Well, a computer ranking doesn't know Michigan hasn't played.

If you want to complain about something, what the heck is Colgate doing up there?

NET is better than the RPI, but you're still fighting with basically two arms behind your back if you're not taking margin of victory into account.
 
#7      
Well, a computer ranking doesn't know Michigan hasn't played.

If you want to complain about something, what the heck is Colgate doing up there?

NET is better than the RPI, but you're still fighting with basically two arms behind your back if you're not taking margin of victory into account.

I get that first part, and it might just be that I’m too busy to have seen it, but has Ohio state lost since they played us? Or any bad losses?

I just think having good wins over that timeframe would be enough to put them ahead of Michigan having no wins.

Tag. You’re it.
 
#8      
We really need OSU and/or Michigan to start losing cause if we end up a 2 seed, I'm pretty sure we'll get paired up with Baylor or the Zags b/c of the conference rules thing.
 
#9      
I get that first part, and it might just be that I’m too busy to have seen it, but has Ohio state lost since they played us? Or any bad losses?

I just think having good wins over that timeframe would be enough to put them ahead of Michigan having no wins.

Tag. You’re it.
Well, the team it's tough to argue should be in front of OSU isn't really Michigan, it's us.

Gotta be honest, I am as perplexed by the current conventional wisdom about the quality and rest of season outlook of this team as I was when people were jumping from rooftops after we lost to Maryland.
 
#11      
NET is better than the RPI, but you're still fighting with basically two arms behind your back if you're not taking margin of victory into account.
I disagree with this. I think a "win-is-a-win-is-a-win" is all that really matters. Now, if you have common opponents, maybe margin of victory could be used as a "tie breaker" of sorts. Otherwise, who cares by how much?

Imagine if in an NCAA tournament a win was disregarded because "you only won by 2 points".
 
#12      
I disagree with this. I think a "win-is-a-win-is-a-win" is all that really matters. Now, if you have common opponents, maybe margin of victory could be used as a "tie breaker" of sorts. Otherwise, who cares by how much?

Imagine if in an NCAA tournament a win was disregarded because "you only won by 2 points".

That makes zero sense. A win vs Nebraska is not the same as a win vs Gonzaga.
 
#14      
That makes zero sense. A win vs Nebraska is not the same as a win vs Gonzaga.
Right, that's not what I'm saying.

What I'm saying is a 7-point Illinois win in overtime against Nebraska is the same as a 34-point regulation win over Nebraska.

My argument is that "margin of victory", in reality, means very little. Winning, however, means a lot.
 
#15      
Right, that's not what I'm saying.

What I'm saying is a 7-point Illinois win in overtime against Nebraska is the same as a 34-point regulation win over Nebraska.

My argument is that "margin of victory", in reality, means very little. Winning, however, means a lot.

Sure, but at issue is comparing teams with similar records. You can't just go by wins and losses. If you want to come up with some ranking of team strength, you have to take into account who a team beat, where, and by how much.
 
#16      
Baltimore, MD
That makes zero sense. A win vs Nebraska is not the same as a win vs Gonzaga.
Quality of opponent is large part of NET rankings. I think his point is more about margin of victory- who cares if you win by 1 or 10+. I disagree with his point, though, as I think margin of victory should be taken into account. I also think it could be argued that it shouldn't be capped at 10 like it is currently, even though you could argue that larger margin of victory is captured in the NET's "net efficiency" rating. We blew the doors off Minnesota, but there's little difference in NET between beating them by 9 or 27 (our margin of victory).

Edit - I was too slow, he clarified his point himself. Ugh.
 
#17      
Missing last year's tournament was a total bummer, but man, how cool is it that we have a really good chance of going from not making the tournament for 7 years in a row to being a #1 seed?? That is just insane! I have still not fully grasped the ridiculousness of the turnaround Underwood has orchestrated. And Ayo might just win the Illinois GOAT argument solely given the context of where this program was before he got here vs. where it is now. Monumental!!
 
#18      
Right, that's not what I'm saying.

What I'm saying is a 7-point Illinois win in overtime against Nebraska is the same as a 34-point regulation win over Nebraska.

My argument is that "margin of victory", in reality, means very little. Winning, however, means a lot.
These metrics are there to try and answer the question "who is better" across a large pool of teams that lack head-to-head games and in many instances lack common opponents. I don't understand how less information is a helpful solution.

Also, if I give you two teams and the only info you have to go off is "Team A beat Nebraska by 7 in overtime. Team B beat Nebraska by 34," and I ask you, based on that limited info, which is better, the answer is absolutely Team B.
 
Last edited:
#19      
Quality of opponent is large part of NET rankings. I think his point is more about margin of victory- who cares if you win by 1 or 10+. I disagree with his point, though, as I think margin of victory should be taken into account. I also think it could be argued that it shouldn't be capped at 10 like it is currently, even though you could argue that larger margin of victory is captured in the NET's "net efficiency" rating. We blew the doors off Minnesota, but there's little difference in NET between beating them by 9 or 27 (our margin of victory).

Edit - I was too slow, he clarified his point himself. Ugh.

I think capping at 10 makes some sense. Maybe you could argue for one more level or something. But really, if the game is not particularly close at the end, it's not unusual for teams to coast or one team to get demoralized and the score to balloon. At a certain point, additional margin is meaningless.
 
#20      
Missing last year's tournament was a total bummer, but man, how cool is it that we have a really good chance of going from not making the tournament for 7 years in a row to being a #1 seed?? That is just insane! I have still not fully grasped the ridiculousness of the turnaround Underwood has orchestrated. And Ayo might just win the Illinois GOAT argument solely given the context of where this program was before he got here vs. where it is now. Monumental!!
Yeah. If no COVID we would have likely been somewhere around a 6 or 7 seed last year, though, so it didn't come completely out of nowhere. I do like the trajectory of the program since BU started.
 
#21      
These metrics are there to try and answer the question "who is better" across a large pool of teams that lack head-to-head games and in many instances lack common opponents. I don't understand how less information is a helpful solution.

Also, if I give you two teams and the only info you have to go off is "Team A beat Nebraska by 7 in overtime. Team B beat Nebraska by 34," and I ask you, based on that limited info, which is better, the answer is absolutely Team B.

More information is always better, but only if it actually matters. Which is why I also said:
...if you have common opponents, maybe margin of victory could be used as a "tie breaker" of sorts...

Honestly, you took one line of what I said, misrepresented it, and then completely ignored the rest of my post.

Long story short, here's what I'm saying:

Let's say you have Team A and Team B.
Team A has lost to the #1 team by 20 points, beaten the #5 team by 20 points, beaten the #6 team by 2 points, and lost to an unranked team by 30 points.
Team B has lost to the #2 team by 5 points, beat the #4 team by 25, beat the #8 team by 10, and lost to an unranked team by 10.

Which team is better?

I see two teams that each lost an expected game, each won two top tier teams, and both lost a game they should have won. Who cares what the margin of victory is in this scenario. Truthfully, for most common opponent games, unless the margin of victory is something like a 30 to 1 disparity, the margin of victory is a worthless stat without looking into a plethora of other scenarios that lead to those margins.

Now, in the grand scheme of things, really we're talking about stats and figures to sort two teams that are basically comparable. I don't think you get it wrong either way.
 
#22      
More information is always better, but only if it actually matters. Which is why I also said:


Honestly, you took one line of what I said, misrepresented it, and then completely ignored the rest of my post.

Long story short, here's what I'm saying:

Let's say you have Team A and Team B.
Team A has lost to the #1 team by 20 points, beaten the #5 team by 20 points, beaten the #6 team by 2 points, and lost to an unranked team by 30 points.
Team B has lost to the #2 team by 5 points, beat the #4 team by 25, beat the #8 team by 10, and lost to an unranked team by 10.

Which team is better?

I see two teams that each lost an expected game, each won two top tier teams, and both lost a game they should have won. Who cares what the margin of victory is in this scenario. Truthfully, for most common opponent games, unless the margin of victory is something like a 30 to 1 disparity, the margin of victory is a worthless stat without looking into a plethora of other scenarios that lead to those margins.

Now, in the grand scheme of things, really we're talking about stats and figures to sort two teams that are basically comparable. I don't think you get it wrong either way.
I get what you're saying, but fundamentally disagree. If you can't take manner of victory into account, you ultimately can't make the fine-grain distinctions that separate say a 4 seed from a 5 seed. May as well just throw all the teams into 4 pots/tiers and then draw them randomly come tournament time (like they do in the Champion's League or World Cup). We could just as easily be a 1 seed as we could a 4 seed under that kind of metric.

At the end of the day the Team A and Team B in your scenario get seeded for the tourney and maybe one is going to have a 1 seed and the other a 2 seed. I personally wouldn't want the committee to only look at who they beat and who they lost to, and not how they won or lost those games.

From a selfish standpoint, for the Illini the efficiency stats, Nebraska game notwithstanding, have been helping our ranking.
 
#23      
Baltimore, MD
I think capping at 10 makes some sense. Maybe you could argue for one more level or something. But really, if the game is not particularly close at the end, it's not unusual for teams to coast or one team to get demoralized and the score to balloon. At a certain point, additional margin is meaningless.
I understand where you're coming from, but the issue is a close game can just as easily balloon and not show how close the game was wire to wire. Take the Texas Tech/WVU game a week or two ago as an example. TT was down 71-69 with just 1:30 left. After a bunch of fouls and a tech on Beard, the final was 82-71, so WVU gets as much "credit" for a 10+ point win in the NET as we did in the Minnesota game where we got up by like 30 and just cruised. It's a hard thing to quantify and I don't think margin necessarily is the best way to do it. I'd like to see some kind of statistic that captures the average lead a team has through a game. Did a team get up and coast or was it close and then a team pull it out at the end and the score not show how close it was.
 
#24      
I understand where you're coming from, but the issue is a close game can just as easily balloon and not show how close the game was wire to wire. Take the Texas Tech/WVU game a week or two ago as an example. TT was down 71-69 with just 1:30 left. After a bunch of fouls and a tech on Beard, the final was 82-71, so WVU gets as much "credit" for a 10+ point win in the NET as we did in the Minnesota game where we got up by like 30 and just cruised. It's a hard thing to quantify and I don't think margin necessarily is the best way to do it. I'd like to see some kind of statistic that captures the average lead a team has through a game. Did a team get up and coast or was it close and then a team pull it out at the end and the score not show how close it was.
I think this is why efficiency stats are more useful over time. Can the last 2 minutes of a game skew the efficiency states for 40 minutes of play? Sure. Can the last 2 minutes of a game meaningfully skew the efficiency stats of an entire season? Doubtful.
 
#25      
Orange Krush Class of 2013
Rochester, MN
I get what you're saying, but fundamentally disagree. If you can't take manner of victory into account, you ultimately can't make the fine-grain distinctions that separate say a 4 seed from a 5 seed. May as well just throw all the teams into 4 pots/tiers and then draw them randomly come tournament time (like they do in the Champion's League or World Cup). We could just as easily be a 1 seed as we could a 4 seed under that kind of metric.

At the end of the day the Team A and Team B in your scenario get seeded for the tourney and maybe one is going to have a 1 seed and the other a 2 seed. I personally wouldn't want the committee to only look at who they beat and who they lost to, and not how they won or lost those games.

From a selfish standpoint, for the Illini the efficiency stats, Nebraska game notwithstanding, have been helping our ranking.

Totally agree. Every possession is a game in itself; margin of victory (after adjusting for tempo) is a basic way to sum up what fraction of those "games" were won, within the larger 40 minutes. Obviously basketball is a 40 minute game, and winning that 40 minutes (by whatever margin) is the goal as a player or coach. But as an analyst trying to decide which team is better, especially two teams like A and B in the example above that look similar by the "eye test" and have the same record, you should:
1) not throw out data (i.e., margin of victory)
2) quantify your uncertainty (e.g., give win probabilities rather than "picks")

As an aside, it would be really cool to have a NET-like ranking that gave ratings as a range, rather than a single value. A team that loses to good teams and beats bad ones might have a narrow range, while a team that whips Wisconsin and then goes to OT versus Nebraska might have a wider one -- what some might call "consistency". Has anyone seen any rating systems that try to quantify this?
 
Status
Not open for further replies.