Does the Unbalanced Schedule Impact the Championship

[uh_no]

This is an oft cited fact, and a discussion was spawned in another thread. Posts by KP were referenced and other alluded to that claim that no, the unbalanced schedule does not have an impact. I find those analyses totally bogus for a couple reasons that I mentioned in the other thread. To put my money where my mouth was, though, I wanted to perform my own.

Of the factors in an unbalanced schedule that impact a teams record, I'll claim that the two most important are:

1) the teams against which you play a home and away. Obviously playing BC twice is far easier than playing UVA twice. Obviously if you play BC twice, you're not guaranteed to win both, but if we believe win probabilities, then a significant % of the time, you would win that extra game. For teams like duke, this would account for ~.5 extra wins (.999 - .5)
2) Of the teams you play twice, which you play at home and away. The impact of home court is real and measured. The impact is about 3.5 points from a neutral court, or 7 total from home->away. This means if a championship-caliber team plays BC and the like on the road, and another championhsip team at home, they are at a significant advantage. The reasoning is they already have a big enough buffer to BC that the away handicap doesn't mean much. It means a huge amount, however, in a game that is likely to be close. Given the win probability graph has a very steep slope when teams are evenly matched, small changes in expected points have major impact on the win probability. a 7 point swing in a game which should be close is about a 25% swing in win probability, but a 7 point swing in a game expected to be closer is expected to be ~1-2%. So each flip flop of location in a game which is expected you to be tight nets you ~.2 more wins.

My goal is to look back at the last 10 years of ACC results and see if it's likely the championship may have been different in a balanced schedule. The first impact is trivial to account for. Each opponent only counts for 1 game, regardless of how many times you played. you either go 1-0, .5-.5, or 0-1. The second is a bit tougher. IF you were within 7 points of an opponent on their home floor, AND you believe the game was representative of the two team's intrinsic abilities, you would expect to win on your own home floor. In the best case, you count that series as "even" and in the worst, you compute a that you would win at home based on the difference. For now, lets assume that if you lost by <7 on the road, you'd win at home (and vice versa).

I'll follow up with the "data" in case I made a mistake, but here's a summary:

2019, while the #1 seed was not affected, the unbalanced schedule helped UNC gain a share of the regular season
2016, no impact, though the unbalanced schedule did allow miami to be far closer than they should have been
2015, unbalanced schedule cost duke a share, as well as the 1 seed
2013, it's EXTREMELY likely the unbalanced schedule cost duke both the sole regular season and the #1 seed.
2012, no impact, though duke was closer than perhaps they ought to have been
2011, cost duke a share of the title, and possibly a 1 seed (though i'm not going to resolve the tiebreaker on that...)
2010, the result was unaffected

So in 10 years, the ACC race was within a game 7 times. Of those 7 times, 4 times the regular season crown was affected (2019, 2015, 2013, 2011), twice was the 1 seed flipped (2015, 2013...maybe 2011) and in the most egregious case, once was the outright champion flipped (2013).

So the argument that the unbalanced schedule doesn't really have an impact on the outcome on the regular season? That argument ain't worth a ticket to the dean dome. The claim was that the difference was <.5 games...welp, turns out ACC regular seasons are usually pretty close at the top, and as I guessed in the other thread, it would be about half of them....as you would expect.


I intentionally didn't get into ifs (for instance, what if zion were not injured!). Those aren't the fault of the balanced schedule. There will ALWAYS be a ton of variability. My argument is only that the ACC is close enough at the top that an unbalanced schedule has a pretty big chance of impacting the outcome, and the data 100% backs this up.

[uh_no]

Last season, UVA and UNC were tied at 16-2. UVA claimed the 1 seed by virtue of a tiebreaker.

Lets adjust UVA:
they lost to duke twice, so get a loss back 16-1
They beat ND twice, so lose a win 15-1
They beat VT twice, so lose a win 14-1
All their home wins were large enough to not warrant second guessing.

Now UNC:
They beat duke twice, so lose a win 15-2
They beat miami twice, so lose a win 14-2
They split with louisville, so lose half of each 13.5-1.5
All their home wins were significant enough.

So in this case, the unbalanced schedules, likely didn't impact the winner.

In 2018, UVA ran away with the title. Also UMBC.

In 2017, UNC won by 2 games.
In 2016, UNC won by 1 game game over each uva and miami with a record of 14-4.

UNC
split with duke 13.5-3.5
2x beat BC 12.5-3.5
2x beat cuse 11.5-3.5
3 close away losses, and a close home win over syracuse, giving them an extra win on average
12.5-2.5

Virginia
split miami 12.5-4.5
split VT 12-4
2x beat UL 11-4
3 close away losses, 2 close home wins, netting them an extra .5
11.5-3.5

Miami
split with UVA 12.5-4.5
2x beat ND 11.5-4.5
2x beat FSU 10.5-4.5
2 close home wins, losing them an extra win
9.5-5.5

So while the scheduling helped miami significantly, it didn't impact the regular season winner itself.

In 2015, uva beat duke by 1 game
Virginia
2x beat ncsu 15-2
2x beat VT 14-2
split with UL 13.5-1.5
They had a close home win
13-2

Duke
2x beat UNC 14-3
2x beat wake 13-3
split with ND 12.5-2.5
We had a close away loss
13-2

So in 2015, it's likely the regular season was impacted by the unbalanced schedule

In 2014, UVA won by 2 games
In 2013, Miami beat duke by a game. Note that in this year, there were only 12 ACC teams, so 7 home and aways. In order to simplify this, we'll only look at 4 non-home-and-away games.

Miami
beat maryland in a close home game 15-4
beat n.c. state in an away game 16-4
beat uva in a close home game 16-5
lost to wake in an away game 16-6

Duke
beat clemson at home 15-4
beat GT at home 16-4
beat FSU away 17-4
lost in a close game to UVA on the road 18-4

It's extremely likely the unbalanced schedule cost duke the regular season in 2013.

In 2012, UNC beat Duke by a game This was a 16 game schedule. We'll still just look at the 6 non-double games

UNC
beat BC at home 15-2
lost to FSU away 15-3
beat VT away 16-3
beat gt at home 17-3
beat wake away 18-3
beat clemson at home 19-3

Duke
beat GT away 14-3
beat UVA close at home 14-4
beat clemson away 15-4
lost to miami at home 15-5
beat NCSU close at home 15-6
beat BC away 16-6

It seems the unbalanced schedule made this closer than it might have otherwise been

In 2011, UNC beat duke by 1 game

UNC
beat uva away 15-2
beat vt close at home 15-3
lost big away to GT 15-4
beat miami away 16-4
beat wake at home 17-4
beat maryland at home 18-4

Duke
lost closet to FSU away 14-3
lost away to wake big 14-4
beat bc at home 15-4
beat GT at home 16-4
lost close to VT away 17-4
beat clemson at home 18-4

It's likely the 2011 regular season was affected by the unbalanced schedule.

in 2010 duke and maryland "tied" with duke taking the 1 seed

Duke
beat wake at home 14-3
lost to state away 14-4
beat FSU at home 15-4
beat miami away 16-4
beat vt at home 17-4
beat uva away 18-4

Maryland
lost to wake close away 14-3
beat BC away 15-3
beat miamia at home 16-3
beat UNC at thome 17-3
beat GT at home close 17-4
beat VT close away 18-4

So it's unlikely the 2010 result was affected

[Bluedog]

So, if I'm reading it right, the unbalanced schedule either has no impact or hurts Duke and NEVER helps Duke?? 😉 Did I read that right? If so, why? Is it because we always play UNC twice and they're typically highly ranked. You'd think that given the somewhat random rotation (I get each team had partners) sometimes it would make our schedule easier.

[uh_no]

So, if I'm reading it right, the unbalanced schedule either has no impact or hurts Duke and NEVER helps Duke?? 😉 Did I read that right? If so, why? Is it because we always play UNC twice and they're typically highly ranked. You'd think that given the somewhat random rotation (I get each team had partners) sometimes it would make our schedule easier.

in 2012 it significantly helped duke relative to UNC, though not enough to get them a tie.

[OldPhiKap]

To me, if it helped or hurt a team then it is unfair and wrong. Whether that meant it “cost” a team or not is irrelevant.

Wrong is wrong. Period.

[MarkD83]

Does the improved strength of schedule help with NCAA seeding

[OldPhiKap]

Does the improved strength of schedule help with NCAA seeding

Not if you lose, or if everyone just beats each other up into mediocrity.

[Acymetric]

My suspicion when we started having this discussion in whichever thread we were discussing it before is that the teams more likely to be impacted are teams in the middle, where the unbalanced schedule puts them on one side of the bubble or the other, or changes their seed line.

[HereBeforeCoachK]

My suspicion when we started having this discussion in whichever thread we were discussing it before is that the teams more likely to be impacted are teams in the middle, where the unbalanced schedule puts them on one side of the bubble or the other, or changes their seed line.

The unbalanced schedule has often helped UVa and hurt Duke and the Cheats....as both are guaranteed to have 2 games with their toughest foe.

But the biggest imbalance is something that cannot be scheduled.....and that is Duke faces 20 opponent "Super Bowls" - and no one else does.

[uh_no]

The unbalanced schedule has often helped UVa and hurt Duke and the Cheats...as both are guaranteed to have 2 games with their toughest foe.

But the biggest imbalance is something that cannot be scheduled....and that is Duke faces 20 opponent "Super Bowls" - and no one else does.

This is "common wisdom" but I'm skeptical the performance of teams against us is appreciably better than one would expect given the quality of the teams. I would suspect it's more that when duke DOES lose, it's blown way out of proportion, both because duke is duke, and because duke is consistently highly ranked, it's always a big upset. I'm not sure, though. Would be an interesting thing to look at.

[TampaDuke]

Not if you lose, or if everyone just beats each other up into mediocrity.

Except for SEC football.

[elvis14]

It's often been noted that teams tend to lose after playing Duke (win or lose). I know last year a football coach was saying he didn't want Clemson to beat his team twice (once on the field and then again the next week as they mentally and physically recover). I think that this can come into play with an unbalanced schedule. Trap games (game before a 'big' game) and hangover games (game after a 'big' game) can happen.

[HereBeforeCoachK]

This is "common wisdom" but I'm skeptical the performance of teams against us is appreciably better than one would expect given the quality of the teams. I would suspect it's more that when duke DOES lose, it's blown way out of proportion, both because duke is duke, and because duke is consistently highly ranked, it's always a big upset. I'm not sure, though. Would be an interesting thing to look at.

Human nature is pretty clear that when you point to one thing a year out of 35 other like things, that your performance in that one you pointed to will likely be your best. Human nature also indicates that home court advantages are real, and if that's the case (it is) then home court advantages with your largest and loudest crowd all year will even be more advantageous. And this does not mean "big upsets" all the time....often it's a mild upset or no upset at all, but it's still the other team's Super Bowl.

This is absolutely a factor.

[DarkstarWahoo]

Except for SEC football.

The SEC Vortex of Domination is a delightful annual tradition! This year, I think the feedback loop started with Auburn's loss to Florida in October. Florida is looked upon as a great team, thus Auburn is credited with a good loss, meaning that Auburn will also be considered a great team, and Florida will get more credit for beating a team that, if it hadn't had an off day, would have almost been good enough to beat Florida.

It's a fun phenomenon that applies to the SEC and no other conference. Every team is great, and thus every win is against a great team and every loss is against a great team. So each game, no matter the result, bolsters the resume of both teams.

[chris13]

The SEC Vortex of Domination is a delightful annual tradition! This year, I think the feedback loop started with Auburn's loss to Florida in October. Florida is looked upon as a great team, thus Auburn is credited with a good loss, meaning that Auburn will also be considered a great team, and Florida will get more credit for beating a team that, if it hadn't had an off day, would have almost been good enough to beat Florida.

It's a fun phenomenon that applies to the SEC and no other conference. Every team is great, and thus every win is against a great team and every loss is against a great team. So each game, no matter the result, bolsters the resume of both teams.

And then when the SEC teams lose a bowl, like Georgia last year losing to Texas, it doesn't really count because Georgia was too sad about blowing the SEC title game and being left out of the playoff.

[DarkstarWahoo]

And then when the SEC teams lose a bowl, like Georgia last year losing to Texas, it doesn't really count because Georgia was too sad about blowing the SEC title game and being left out of the playoff.

While you're absolutely correct, the Alabama-Utah Didn't Want to Be There Corollary is in a completely different section of code.

[HereBeforeCoachK]

And then when the SEC teams lose a bowl, like Georgia last year losing to Texas, it doesn't really count because Georgia was too sad about blowing the SEC title game and being left out of the playoff.

I don't know who says it doesn't count....and I certainly don't toot the horn for the SEC.....but dude, this kind of thing is a factor. Watch the Sharps in Vegas take advantage of it every bowl season. Now playing a program like Texas should've gotten Georgia psyched, but the fact is it did not, and that was a factor in the outcome. There are several of these every bowl season, and they are often predictable.

[uh_no]

continuing the analysis for 2020

Duke: 15-5
2x win over UNC: 14-5
split with NCSU: 13.5 - 4.5
split with wake: 13 - 4
swept miami: 12-4
swept VT: 11-4
swept BC: 10-4

close loss on the road to clemson: 10.5-3.5
close loss on the road to UVA: 11-3
Close win at home vs FSU: 10.5-3.5

FSU: 16-4
split pitt: 15.5-3.5
split clemson: 15-3
swept UL: 14-3
split virginia: 13.5-2.5
swept miami: 12.5-2.5
swept ND: 11.5-2.5

close loss away to duke: 12-2
close win at home over UNC: 11.5-2.5
close win at home over Syracuse: 11-3


Louisville: 15-5
swept miami: 14-5
swept pitt: 13-5
swept by FSU: 13-4
split GT: 12.5-3.5
split clemson: 12-3
split virigina: 11.5-2.5



UVA: 15-5
split cuse 14.5-4.5
swept carolina: 13.5-4.5
swept VT: 12.5-4.5
split bC: 12-4
split FSU: 11.5-3.5
split UL: 11-3

close home win vs duke: 10.5-3.5
close home win over ND: 10-4
close home win over clemson: 9.5-4.5


Despite the fact that Louisville technically had .5 more adjusted wins, given that FSU swept them, I don't think I can conclude that an unbalanced schedule was a driving factor in the regular season title or 1 seed. I expect that to be the outcome more often moving forward as the expanded schedule means the teams you play twice regress towards the mean of "average ACC team," and further, you're slightly less likely to get burned by playing an especially good or bad only on the road or at home.

The biggest conclusion this year was UVA getting to play duke at home likely had a hand in pulling them into the mix when they otherwise wouldn't have been.

[jv001]

continuing the analysis for 2020

Duke: 15-5
2x win over UNC: 14-5
split with NCSU: 13.5 - 4.5
split with wake: 13 - 4
swept miami: 12-4
swept VT: 11-4
swept BC: 10-4

close loss on the road to clemson: 10.5-3.5
close loss on the road to UVA: 11-3
Close win at home vs FSU: 10.5-3.5

FSU: 16-4
split pitt: 15.5-3.5
split clemson: 15-3
swept UL: 14-3
split virginia: 13.5-2.5
swept miami: 12.5-2.5
swept ND: 11.5-2.5

close loss away to duke: 12-2
close win at home over UNC: 11.5-2.5
close win at home over Syracuse: 11-3


Louisville: 15-5
swept miami: 14-5
swept pitt: 13-5
swept by FSU: 13-4
split GT: 12.5-3.5
split clemson: 12-3
split virigina: 11.5-2.5



UVA: 15-5
split cuse 14.5-4.5
swept carolina: 13.5-4.5
swept VT: 12.5-4.5
split bC: 12-4
split FSU: 11.5-3.5
split UL: 11-3

close home win vs duke: 10.5-3.5
close home win over ND: 10-4
close home win over clemson: 9.5-4.5

can
Despite the fact that Louisville technically had .5 more adjusted wins, given that FSU swept them, I don't think I can conclude that an unbalanced schedule was a driving factor in the regular season title or 1 seed. I expect that to be the outcome more often moving forward as the expanded schedule means the teams you play twice regress towards the mean of "average ACC team," and further, you're slightly less likely to get burned by playing an especially good or bad only on the road or at home.

The biggest conclusion this year was UVA getting to play duke at home likely had a hand in pulling them into the mix when they otherwise wouldn't have been.

That and J-Rob not playing one minute in the Virginia game after having a very good game against Wake. Baker, Jack and Hurt had these combined stats in that game:
32 minutes, 2 points, 4 rebounds, 1-11 from the field, 0-5 on threes and 2 turnovers. No one will convince me, don't even try, that J-Rob would have made a difference in that game. Javin played well in his 23 minutes and contributed in a positive way but those other 32 minutes could have made a difference. I'll just have to look in the rafters at Cameron. I guess. :cool:

GoDuke!

[uh_no]

That and J-Rob not playing one minute in the Virginia game after having a very good game against Wake. Baker, Jack and Hurt had these combined stats in that game:
32 minutes, 2 points, 4 rebounds, 1-11 from the field, 0-5 on threes and 2 turnovers. No one will convince me, don't even try, that J-Rob would have made a difference in that game. Javin played well in his 23 minutes and contributed in a positive way but those other 32 minutes could have made a difference. I'll just have to look in the rafters at Cameron. I guess. :cool:

GoDuke!

this isn't the "Does underutilizing JRob impact the championship" thread :D perhaps such a thread doesn't exist because the answer is obvious.

[DarkstarWahoo]

The SEC Vortex of Domination is a delightful annual tradition! This year, I think the feedback loop started with Auburn's loss to Florida in October. Florida is looked upon as a great team, thus Auburn is credited with a good loss, meaning that Auburn will also be considered a great team, and Florida will get more credit for beating a team that, if it hadn't had an off day, would have almost been good enough to beat Florida.

It's a fun phenomenon that applies to the SEC and no other conference. Every team is great, and thus every win is against a great team and every loss is against a great team. So each game, no matter the result, bolsters the resume of both teams.

Fun to see this thread pop back up, because it turns out I was wrong about the bolded portion! It's been tons of fun to see the NET-Enabled B1G Hoops Vortex of Domination will itself into existence this year.

[uh_no]

Fun to see this thread pop back up, because it turns out I was wrong about the bolded portion! It's been tons of fun to see the NET-Enabled B1G Hoops Vortex of Domination will itself into existence this year.

the B10 is ACTUALLY good this year though. 10 teams in KP top 30 is a huge concentration. Objectively they are the best conference by far, and the SEC the worst.

[jhmoss1812]

continuing the analysis for 2020

Duke: 15-5
2x win over UNC: 14-5
split with NCSU: 13.5 - 4.5
split with wake: 13 - 4
swept miami: 12-4
swept VT: 11-4
swept BC: 10-4

close loss on the road to clemson: 10.5-3.5
close loss on the road to UVA: 11-3
Close win at home vs FSU: 10.5-3.5

FSU: 16-4
split pitt: 15.5-3.5
split clemson: 15-3
swept UL: 14-3
split virginia: 13.5-2.5
swept miami: 12.5-2.5
swept ND: 11.5-2.5

close loss away to duke: 12-2
close win at home over UNC: 11.5-2.5
close win at home over Syracuse: 11-3


Louisville: 15-5
swept miami: 14-5
swept pitt: 13-5
swept by FSU: 13-4
split GT: 12.5-3.5
split clemson: 12-3
split virigina: 11.5-2.5



UVA: 15-5
split cuse 14.5-4.5
swept carolina: 13.5-4.5
swept VT: 12.5-4.5
split bC: 12-4
split FSU: 11.5-3.5
split UL: 11-3

close home win vs duke: 10.5-3.5
close home win over ND: 10-4
close home win over clemson: 9.5-4.5


Despite the fact that Louisville technically had .5 more adjusted wins, given that FSU swept them, I don't think I can conclude that an unbalanced schedule was a driving factor in the regular season title or 1 seed. I expect that to be the outcome more often moving forward as the expanded schedule means the teams you play twice regress towards the mean of "average ACC team," and further, you're slightly less likely to get burned by playing an especially good or bad only on the road or at home.

The biggest conclusion this year was UVA getting to play duke at home likely had a hand in pulling them into the mix when they otherwise wouldn't have been.

This is not intended to come across as snarky but do you think Duke would have been in the mix if they had to travel to FSU and Louisville?

[uh_no]

This is not intended to come across as snarky but do you think Duke would have been in the mix if they had to travel to FSU and Louisville?

that's accounted for in the analysis, which attempts to normalize the unbalanced-ness, and duke came out a game better than virginia, and only a half game behind louisville.

So yes.

Remember it's not just "if duke had to travel to louisville" it's "if everyone had to travel everywhere"

[sagegrouse]

the B10 is ACTUALLY good this year though. 10 teams in KP top 30 is a huge concentration. Objectively they are the best conference by far, and the SEC the worst.

Very hard to make comparisons across conferences, when there are only ten interconference games in the top leagues after January 1 (SEC-Big 12 challenge). Even if you are genius like KenPom -- there's no good data.

[Wahoo2000]

This is an oft cited fact, and a discussion was spawned in another thread. Posts by KP were referenced and other alluded to that claim that no, the unbalanced schedule does not have an impact. I find those analyses totally bogus for a couple reasons that I mentioned in the other thread. To put my money where my mouth was, though, I wanted to perform my own.

Of the factors in an unbalanced schedule that impact a teams record, I'll claim that the two most important are:

1) the teams against which you play a home and away. Obviously playing BC twice is far easier than playing UVA twice. Obviously if you play BC twice, you're not guaranteed to win both, but if we believe win probabilities, then a significant % of the time, you would win that extra game. For teams like duke, this would account for ~.5 extra wins (.999 - .5)
2) Of the teams you play twice, which you play at home and away. The impact of home court is real and measured. The impact is about 3.5 points from a neutral court, or 7 total from home->away. This means if a championship-caliber team plays BC and the like on the road, and another championhsip team at home, they are at a significant advantage. The reasoning is they already have a big enough buffer to BC that the away handicap doesn't mean much. It means a huge amount, however, in a game that is likely to be close. Given the win probability graph has a very steep slope when teams are evenly matched, small changes in expected points have major impact on the win probability. a 7 point swing in a game which should be close is about a 25% swing in win probability, but a 7 point swing in a game expected to be closer is expected to be ~1-2%. So each flip flop of location in a game which is expected you to be tight nets you ~.2 more wins.

My goal is to look back at the last 10 years of ACC results and see if it's likely the championship may have been different in a balanced schedule. The first impact is trivial to account for. Each opponent only counts for 1 game, regardless of how many times you played. you either go 1-0, .5-.5, or 0-1. The second is a bit tougher. IF you were within 7 points of an opponent on their home floor, AND you believe the game was representative of the two team's intrinsic abilities, you would expect to win on your own home floor. In the best case, you count that series as "even" and in the worst, you compute a that you would win at home based on the difference. For now, lets assume that if you lost by <7 on the road, you'd win at home (and vice versa).

I'll follow up with the "data" in case I made a mistake, but here's a summary:

2019, while the #1 seed was not affected, the unbalanced schedule helped UNC gain a share of the regular season
2016, no impact, though the unbalanced schedule did allow miami to be far closer than they should have been
2015, unbalanced schedule cost duke a share, as well as the 1 seed
2013, it's EXTREMELY likely the unbalanced schedule cost duke both the sole regular season and the #1 seed.
2012, no impact, though duke was closer than perhaps they ought to have been
2011, cost duke a share of the title, and possibly a 1 seed (though i'm not going to resolve the tiebreaker on that...)
2010, the result was unaffected

So in 10 years, the ACC race was within a game 7 times. Of those 7 times, 4 times the regular season crown was affected (2019, 2015, 2013, 2011), twice was the 1 seed flipped (2015, 2013...maybe 2011) and in the most egregious case, once was the outright champion flipped (2013).

So the argument that the unbalanced schedule doesn't really have an impact on the outcome on the regular season? That argument ain't worth a ticket to the dean dome. The claim was that the difference was <.5 games...welp, turns out ACC regular seasons are usually pretty close at the top, and as I guessed in the other thread, it would be about half of them...as you would expect.


I intentionally didn't get into ifs (for instance, what if zion were not injured!). Those aren't the fault of the balanced schedule. There will ALWAYS be a ton of variability. My argument is only that the ACC is close enough at the top that an unbalanced schedule has a pretty big chance of impacting the outcome, and the data 100% backs this up.

This seems like a lot of sketchy math to base the entire thread on. I think I'd rather base conference SoS judgements on those professional stat/math-heads like kenpom, whose eval is probably much more nuanced (honestly no offense meant to you personally - you obv put some time and thought into this, much more than me). Conference SoS is available in his conference breakdown. As far as figuring how much that SoS difference impacts actual wins and losses? You got me - I'll say that any year the outright champ is only 1 game ahead of 2nd, and 2nd has a significantly higher conference SoS, you can make an argument.

Observationally, it also seems how a team performs in their "wtf" games (or rather, how many "wtf" games they end up HAVING) has a WAAAAAAAAAAY larger effect on the title than the SOS. In other words, among top teams, it's usually going to come down more to consistency than quality w/r/t who walks away with the regular season "belt". For instance - any of Louisville/Duke/UVa upset about their finish should be looking more at games vs GT/Wake/BC as to why they finished where they did rather than what their schedules were vs each other. I'd only end up with worrying about these schedule differences when the top teams all pretty much took care of business and won all the "should win" games. If that's the case, and it comes down to a big imbalance in scheduling within your tier, then you can gripe a little.

Just my opinion, which is FAAAAR from an expert one.

[Wahoo2000]

Very hard to make comparisons across conferences, when there are only ten interconference games in the top leagues after January 1 (SEC-Big 12 challenge). Even if you are genius like KenPom -- there's no good data.

This is correct - there's no real accounting for teams within a conference making a large leap in improvement vs the level of improvement made in other conferences as a whole.

I kind of dream of a cbb world where we can either:
a) go back to 14 or 16 game conference schedules, or
b) sprinkle noncon all throughout the year for all teams. Play 2-3 noncon games each and every month from nov-feb.

[House P]

This is correct - there's no real accounting for teams within a conference making a large leap in improvement vs the level of improvement made in other conferences as a whole.

I kind of dream of a cbb world where we can either:
a) go back to 14 or 16 game conference schedules, or
b) sprinkle noncon all throughout the year for all teams. Play 2-3 noncon games each and every month from nov-feb.

I am with you. My proposal would be for the NCAA tourney selection committee to institute a rule that at-large bids will only be given to teams who play multiple Quad 2 or better non-conference games after Jan 15.

[uh_no]

This seems like a lot of sketchy math to base the entire thread on. I think I'd rather base conference SoS judgements on those professional stat/math-heads like kenpom, whose eval is probably much more nuanced (honestly no offense meant to you personally - you obv put some time and thought into this, much more than me). Conference SoS is available in his conference breakdown. As far as figuring how much that SoS difference impacts actual wins and losses? You got me - I'll say that any year the outright champ is only 1 game ahead of 2nd, and 2nd has a significantly higher conference SoS, you can make an argument.

Observationally, it also seems how a team performs in their "wtf" games (or rather, how many "wtf" games they end up HAVING) has a WAAAAAAAAAAY larger effect on the title than the SOS. In other words, among top teams, it's usually going to come down more to consistency than quality w/r/t who walks away with the regular season "belt". For instance - any of Louisville/Duke/UVa upset about their finish should be looking more at games vs GT/Wake/BC as to why they finished where they did rather than what their schedules were vs each other. I'd only end up with worrying about these schedule differences when the top teams all pretty much took care of business and won all the "should win" games. If that's the case, and it comes down to a big imbalance in scheduling within your tier, then you can gripe a little.

Just my opinion, which is FAAAAR from an expert one.

The question this thread sought to answer was "is the ACC regular season regularly close enough that the unbalanced schedule likely imacts the outcome?" and the answer is overwhelmingly "yes".

That doesn't mean that things like "my team should have stunk less" isn't also something that would have made a difference. If you think that was my point, you've far misunderstood what I'm trying to do.

The beauty of this is it doesn't depend on any compiled SoS judgement value, even if that value is somewhat objective. It simply looks at how an individual team actually performed game by game. The point isn't to say that X or Y would have or should have won with a balanced schedule, only that there is a good chance it made a difference. The latter is a much weaker statement, but one that expressly contradicts the analyses KP has done based on the SoS values, which at worst, indicates that KP is making conclusions on his analysis perhaps beyond what the data would suggest.


Probably correct: natural variance has a pretty big impact on the regular season
Probably correct: the variable impact of the unbalanced schedule isn't usually that big between any two teams
KP conclusion: the unbalanced schedule can be ignored because the natural variance is big
My conclusion: the ACC is close enough that even the small variations need to be considered.


It's the difference between saying "a bad call in the last minute doesn't matter because the team could have made some shot earlier" and "the game is close, so a single bad call can swing the balance"

[Wahoo2000]

The question this thread sought to answer was "is the ACC regular season regularly close enough that the unbalanced schedule likely imacts the outcome?" and the answer is overwhelmingly "yes".

That doesn't mean that things like "my team should have stunk less" isn't also something that would have made a difference. If you think that was my point, you've far misunderstood what I'm trying to do.

The beauty of this is it doesn't depend on any compiled SoS judgement value, even if that value is somewhat objective. It simply looks at how an individual team actually performed game by game. The point isn't to say that X or Y would have or should have won with a balanced schedule, only that there is a good chance it made a difference. The latter is a much weaker statement, but one that expressly contradicts the analyses KP has done based on the SoS values, which at worst, indicates that KP is making conclusions on his analysis perhaps beyond what the data would suggest.


Probably correct: natural variance has a pretty big impact on the regular season
Probably correct: the variable impact of the unbalanced schedule isn't usually that big between any two teams
KP conclusion: the unbalanced schedule can be ignored because the natural variance is big
My conclusion: the ACC is close enough that even the small variations need to be considered.


It's the difference between saying "a bad call in the last minute doesn't matter because the team could have made some shot earlier" and "the game is close, so a single bad call can swing the balance"

I like that last sentence. Intended or not, I'm interpreting this as "it's more about your point of view than really proving the point either way" (FWIW, I'm totally one of those guys who, when my team is screwed or helped by a late call, I'm more likely to say "meh - we/they shoulda played better if we/they didn't want it to come down to an official's mistake")

Edited to add - for me it's also a question not of WHETHER the schedule makes an impact (I mean, living in a colder weather climate as opposed to a warmer one could have an impact, but how much?) - it's "what is the significance of that impact"? I tend to think there are so many BIGGER factors affecting the title race, that conf SoS, while maybe not completely negligible, is a pretty small deal. Others are free to disagree - and if someone can make a REALLY in-depth, high end statistical model that takes a lot more factors into account, maybe I'd be convinced it's a bigger deal.

[uh_no]

I like that last sentence. Intended or not, I'm interpreting this as "it's more about your point of view than really proving the point either way" (FWIW, I'm totally one of those guys who, when my team is screwed or helped by a late call, I'm more likely to say "meh - we/they shoulda played better if we/they didn't want it to come down to an official's mistake")

Edited to add - for me it's also a question not of WHETHER the schedule makes an impact (I mean, living in a colder weather climate as opposed to a warmer one could have an impact, but how much?) - it's "what is the significance of that impact"? I tend to think there are so many BIGGER factors affecting the title race, that conf SoS, while maybe not completely negligible, is a pretty small deal. Others are free to disagree - and if someone can make a REALLY in-depth, high end statistical model that takes a lot more factors into account, maybe I'd be convinced it's a bigger deal.

totally agree.

The unbalanced schedule is what it is. Bad calls are what they are. I would love a double round robin, and I would love perfectly balanced reffing, but I'm not going to get it....so the best thing to do is go out and play better. Duke sure as heck didn't do that, especially this year.

[Wahoo2000]

totally agree.

The unbalanced schedule is what it is. Bad calls are what they are. I would love a double round robin, and I would love perfectly balanced reffing, but I'm not going to get it...so the best thing to do is go out and play better. Duke sure as heck didn't do that, especially this year.

Double agree - the reg season title was there for the taking by any of the top 4 - kudos to FSU for having the least hiccups and taking the crown (though I sure would have liked to see ND not choke that game away last week. Grrrrr)

[Faustus]

Eh... I just can't worry about all this either, not just the unbalanced schedules, at the mercy of the conference office drawn up last summer, but the time in the season these teams play, and that's always been a complicating point whether teams played full round-robin schedules or not. I think Duke would easily have beaten UVa even in Cville had the game been in January, when UVa still couldn't hit anything and Duke's offense was somehow more effective, but conversely I have no confidence Duke would have beaten Michigan State or Kansas if they'd played in the last couple of weeks either, instead of before Christmas. And how can a committee scientifically correctly rank or seed this year's Duke team that both beat a bad Wake time by over 25 points, yet inexplicably lost to them a few weeks later? Which team do they rank? Just hoping the team that enters the NCAA's this month is healthy and playing at a high level vs. whomever the NCAA committee places in front of them.

[MarkD83]

With all of this great analysis I bet unc still finishes last

[uh_no]

With all of this great analysis I bet unc still finishes last

UNC could have the best record and they'd still finish last in my book.

[Toolie92]

Wahoo here. Figured this was a safe thread to post in since it mostly deals with Duke v. Virginia schedule but this is my first post so feel free to ignore me.

As I understand the approach, you are taking post-facto game outcomes then changing them based on H/A venue to subtract out wins and losses. This uses game outcomes (in particular, only the outcomes of very close games) to judge the toughness of a schedule rather than looking at the schedule as a whole. By that method a close home win over the toughest team could be worth the same as a close road loss to the worst team (both get erased). This seems to reward MOV without actually considering whether the opponent was good or weak, and whether you are getting double games vs. the weakest teams.

For example, your method applied to 2020 dinged Virginia for its close wins. But Torvik and KenPom both say that Duke had the easiest ACC schedule of all 15 teams, while Virginia's was harder than all the Top 8 finishers except GT (top teams always tend to have easier schedules since they do not play themselves). According to them, the biggest beneficiary of unbalanced scheduling this year was actually Duke (2xBC, 2xWF, 2xMiami, 2xUNC, 2xVT - that is 5 of the bottom 6 ACC teams). (I would add none of the differences were very large and I think unbalanced scheduling is a minor factor - I am not suggesting Duke should have been 3rd). The schedule should be fair or unfair regardless of the actual game outcomes - either you had an easy slate of games or a hard slate of games, and this year Duke undeniably had an easy slate of games.

The way advanced stats normally approaches this is to take each teams schedule and compute how many wins a median league team would win against that particular schedule (without regard to the actual game results, by adding up the win probabilities for each game). This would represent the handicap each team faced with its particular schedule. If the relative handicaps of various teams would have altered the league final standings, then you could say the schedule might have mattered. I am not bringing any math to the table today, but when I have done the calculations in past seasons (looking at expected KenPom win probabilities for each ACC schedule) it was generally less than 0.5 games, and in cases where a team had a more substantial (say 1 game) advantage or disadvantage it was almost always just a random team down the standings, not the champ.

[devildeac]

Wahoo here. Figured this was a safe thread to post in since it mostly deals with Duke v. Virginia schedule but this is my first post so feel free to ignore me.

As I understand the approach, you are taking post-facto game outcomes then changing them based on H/A venue to subtract out wins and losses. This uses game outcomes (in particular, only the outcomes of very close games) to judge the toughness of a schedule rather than looking at the schedule as a whole. By that method a close home win over the toughest team could be worth the same as a close road loss to the worst team (both get erased). This seems to reward MOV without actually considering whether the opponent was good or weak, and whether you are getting double games vs. the weakest teams.

For example, your method applied to 2020 dinged Virginia for its close wins. But Torvik and KenPom both say that Duke had the easiest ACC schedule of all 15 teams, while Virginia's was harder than all the Top 8 finishers except GT (top teams always tend to have easier schedules since they do not play themselves). According to them, the biggest beneficiary of unbalanced scheduling this year was actually Duke (2xBC, 2xWF, 2xMiami, 2xUNC, 2xVT - that is 5 of the bottom 6 ACC teams). (I would add none of the differences were very large and I think unbalanced scheduling is a minor factor - I am not suggesting Duke should have been 3rd). The schedule should be fair or unfair regardless of the actual game outcomes - either you had an easy slate of games or a hard slate of games, and this year Duke undeniably had an easy slate of games.

The way advanced stats normally approaches this is to take each teams schedule and compute how many wins a median league team would win against that particular schedule (without regard to the actual game results, by adding up the win probabilities for each game). This would represent the handicap each team faced with its particular schedule. If the relative handicaps of various teams would have altered the league final standings, then you could say the schedule might have mattered. I am not bringing any math to the table today, but when I have done the calculations in past seasons (looking at expected KenPom win probabilities for each ACC schedule) it was generally less than 0.5 games, and in cases where a team had a more substantial (say 1 game) advantage or disadvantage it was almost always just a random team down the standings, not the champ.

Nice start/first post. Good analysis.

I've been hesitant to wade in on this thread (not a stats geek), but, Duke had the "advantage" with favorable scheduling this year and we did well but didn't "get the job done." Nor did we get the job done in alternating years when we've had favorable scheduling. UVa and the cheats (sadly) did well when presented with those opportunities.

[jv001]

Wahoo here. Figured this was a safe thread to post in since it mostly deals with Duke v. Virginia schedule but this is my first post so feel free to ignore me.

As I understand the approach, you are taking post-facto game outcomes then changing them based on H/A venue to subtract out wins and losses. This uses game outcomes (in particular, only the outcomes of very close games) to judge the toughness of a schedule rather than looking at the schedule as a whole. By that method a close home win over the toughest team could be worth the same as a close road loss to the worst team (both get erased). This seems to reward MOV without actually considering whether the opponent was good or weak, and whether you are getting double games vs. the weakest teams.

For example, your method applied to 2020 dinged Virginia for its close wins. But Torvik and KenPom both say that Duke had the easiest ACC schedule of all 15 teams, while Virginia's was harder than all the Top 8 finishers except GT (top teams always tend to have easier schedules since they do not play themselves). According to them, the biggest beneficiary of unbalanced scheduling this year was actually Duke (2xBC, 2xWF, 2xMiami, 2xUNC, 2xVT - that is 5 of the bottom 6 ACC teams). (I would add none of the differences were very large and I think unbalanced scheduling is a minor factor - I am not suggesting Duke should have been 3rd). The schedule should be fair or unfair regardless of the actual game outcomes - either you had an easy slate of games or a hard slate of games, and this year Duke undeniably had an easy slate of games.

The way advanced stats normally approaches this is to take each teams schedule and compute how many wins a median league team would win against that particular schedule (without regard to the actual game results, by adding up the win probabilities for each game). This would represent the handicap each team faced with its particular schedule. If the relative handicaps of various teams would have altered the league final standings, then you could say the schedule might have mattered. I am not bringing any math to the table today, but when I have done the calculations in past seasons (looking at expected KenPom win probabilities for each ACC schedule) it was generally less than 0.5 games, and in cases where a team had a more substantial (say 1 game) advantage or disadvantage it was almost always just a random team down the standings, not the champ.

Good post and welcome to DBR. Most Duke fans agreed that the 2020 conference schedule favored Duke and we were hoping this would be the year we would win the regular season title. But as devildeac said, we didn't get the job done. We had home against Lville, home against FSU and away at the Cavs and we went 1-2. Then we lose at Clemson which was tough at home and at NCSU. We have no one to blame but ourselves and I'm sure the Duke kids are not happy in how they played in those losses. But when you have 4 freshmen playing major minutes, that can happen.

GoDuke!

[HereBeforeCoachK]

Good post and welcome to DBR. Most Duke fans agreed that the 2020 conference schedule favored Duke and we were hoping this would be the year we would win the regular season title. But as devildeac said, we didn't get the job done. We had home against Lville, home against FSU and away at the Cavs and we went 1-2. Then we lose at Clemson which was tough at home and at NCSU. We have no one to blame but ourselves and I'm sure the Duke kids are not happy in how they played in those losses. But when you have 4 freshmen playing major minutes, that can happen.

GoDuke!

I would agree, but the only reason we had an advantage this year is the two games against the Heels were not nearly as tough as normal. With a guaranteed two spot in that match up every season, UVa has been the team to most normally have the advantage. But clearly Duke did this season, and did not take advantage. FSU worthy reg season champ. Really nice push by UVa late.

[uh_no]

I would agree, but the only reason we had an advantage this year is the two games against the Heels were not nearly as tough as normal. With a guaranteed two spot in that match up every season, UVa has been the team to most normally have the advantage. But clearly Duke did this season, and did not take advantage. FSU worthy reg season champ. Really nice push by UVa late.

It was much worse than that.

NCSU was pretty good, but other than that, the teams we played twice have the following ACCT seeds:
9
10
11
12
14

So aside from swapping pitt for NCSU, our H+H games were against the worst teams in the league. Of the other top teams in the league, we had 2 at home and one on the road.

You couldn't have dreamt a more ideal schedule. It wasn't just UNC...it was the entire thing.

[DarkstarWahoo]

I would agree, but the only reason we had an advantage this year is the two games against the Heels were not nearly as tough as normal. With a guaranteed two spot in that match up every season, UVa has been the team to most normally have the advantage. But clearly Duke did this season, and did not take advantage. FSU worthy reg season champ. Really nice push by UVa late.

Without getting too deep into the schedule weeds, I would like to point out that UVA's counterpart in this comparison is Louisville, and their NCAA seeds since they joined the league have been:

4
Probation, but 23-8 record
2
NIT
7
And whatever they wind up with this season.

UNC has gone 4-1-1-2-1-LOL in that timespan. They've been tougher than Louisville, but it's not like UVA is getting 2018 Pitt twice a year. (The other two permanent partners - VT for UVA and Wake for Duke - seem fairly comparable to me.)