Weekly Polls Thread Week 8

[bearcreekfan]

Almost all the metrics don’t like us. There was a blurb about it today in The Athletic where they compared us to similar teams over the past decade and in general, those teams have not ended up doing all that well over the course of the season.

Men’s college basketball Top 25: Four elite defenses to watch as conference play heats up

Marquette, Maryland, Duke and Cincinnati have four of the best defenses in college basketball, and they get some love this week.

www.nytimes.com

blurb on OU...

"No. 17 Oklahoma​

The Sooners are now 12-0 with six wins over high-majors, yet the computers don’t seem to like them. Oklahoma currently ranks 37th at KenPom, 45th at Bart Torvik and 30th at Evan Miya. I went back through the last 10 years to find some similar comps, using Torvik’s sorting tool to search for any high-majors that were undefeated on the Sunday closest to Christmas but didn’t rank in the top 25 at that site.

USC (16-17)	54	22	26-10	62	Lost in Round of 32
Creighton (16-17)	30	10	25-10	28	Lost in Round of 64
TCU (17-18)	29	10	21-12	24	Lost in Round of 64
St. John’s (18-19)	67	NR	21-13	84	Lost play-in game
Ole Miss (23-24)	61	24	20-12	79	Missed
Oklahoma (24-25)	45	14 (last week)	??	??	??

I’m higher on the Sooners than the computers, but if history is any indication, the computers might be onto something."

The column headings above wouldn't copy, but they are: Torvik Rank Near Xmas, AP Rank Near Xmas, Final Record, Final Torvik Rank, NCAA Tournament

 

[CoachTalk]

What an unnecessary comment with the only possible goal of creating non-constructive arguments and/or derailing a thread. A bit hypocritical how you’re doing the very thing here that many - yourself included - complain that other posters on this board do in creating unnecessary dissension, don’t you think? Can we all not agree to just be above this type of sarcastic, petty behavior?

I mean joking isn’t ok?

We all know undefeated gets us in…

I won’t agree. I agreed awhile back to be better but you aren’t taking h humor away.

Merry Christmas

 

[schoonerdriver]

We all saw last year. Major divergence was last year we were smoked by UNC and this season we outlasted Michigan. The problem is it appears that we may have a more difficult conference schedule this year over last season in the Big 12. We have lots of work yet to do.

 

[snydrosooner]

Men’s college basketball Top 25: Four elite defenses to watch as conference play heats up

Marquette, Maryland, Duke and Cincinnati have four of the best defenses in college basketball, and they get some love this week.

www.nytimes.com

blurb on OU...

"No. 17 Oklahoma​

The Sooners are now 12-0 with six wins over high-majors, yet the computers don’t seem to like them. Oklahoma currently ranks 37th at KenPom, 45th at Bart Torvik and 30th at Evan Miya. I went back through the last 10 years to find some similar comps, using Torvik’s sorting tool to search for any high-majors that were undefeated on the Sunday closest to Christmas but didn’t rank in the top 25 at that site.

USC (16-17)	54	22	26-10	62	Lost in Round of 32
Creighton (16-17)	30	10	25-10	28	Lost in Round of 64
TCU (17-18)	29	10	21-12	24	Lost in Round of 64
St. John’s (18-19)	67	NR	21-13	84	Lost play-in game
Ole Miss (23-24)	61	24	20-12	79	Missed
Oklahoma (24-25)	45	14 (last week)	??	??	??

I’m higher on the Sooners than the computers, but if history is any indication, the computers might be onto something."

The column headings above wouldn't copy, but they are: Torvik Rank Near Xmas, AP Rank Near Xmas, Final Record, Final Torvik Rank, NCAA Tournament

Unfortunately, I think computers take too much credit in how big you win. I have always been of the opinion that you are up by 15 or so with 10 minutes left in a game you give some of your backups more minutes. This can limit the spread of victory. Leaving the burners on against bad teams does not prove how good of a team you are, but computers seem to like it.

 

[BoulderSooner]

Men’s college basketball Top 25: Four elite defenses to watch as conference play heats up

Marquette, Maryland, Duke and Cincinnati have four of the best defenses in college basketball, and they get some love this week.

www.nytimes.com

blurb on OU...

"No. 17 Oklahoma​

The Sooners are now 12-0 with six wins over high-majors, yet the computers don’t seem to like them. Oklahoma currently ranks 37th at KenPom, 45th at Bart Torvik and 30th at Evan Miya. I went back through the last 10 years to find some similar comps, using Torvik’s sorting tool to search for any high-majors that were undefeated on the Sunday closest to Christmas but didn’t rank in the top 25 at that site.

USC (16-17)	54	22	26-10	62	Lost in Round of 32
Creighton (16-17)	30	10	25-10	28	Lost in Round of 64
TCU (17-18)	29	10	21-12	24	Lost in Round of 64
St. John’s (18-19)	67	NR	21-13	84	Lost play-in game
Ole Miss (23-24)	61	24	20-12	79	Missed
Oklahoma (24-25)	45	14 (last week)	??	??	??

I’m higher on the Sooners than the computers, but if history is any indication, the computers might be onto something."

The column headings above wouldn't copy, but they are: Torvik Rank Near Xmas, AP Rank Near Xmas, Final Record, Final Torvik Rank, NCAA Tournament

the key to me is that 4 of the 5 made the Dance ..

i wish i could get those teams SOR at the same point in time ..

 

[stormspencer]

Unfortunately, I think computers take too much credit in how big you win. I have always been of the opinion that you are up by 15 or so with 10 minutes left in a game you give some of your backups more minutes. This can limit the spread of victory. Leaving the burners on against bad teams does not prove how good of a team you are, but computers seem to like it.

You are correct, but the problem is that our backups haven't necessarily played a ton against the bad opponents. Computers/advanced metrics would prefer that you absolutely dominate craptastic opponents. Unfortunately, in far too many games against the sub-300 teams this year, we've played fairly iffy and disinterested causing the games to be closer than the predictive scores would calculate.....thus affecting our overall numbers in the analytics. And again based on recent history, it doesn't necessarily bode well for OU moving forward. It also doesn't mean we are going to flop either, but we've seen this up close and personal the last several years and we got the short end of the stick.

Computers rely heavily on probabilities and simulations. For example, we beat Michigan by 1, but there is still a gap between them and us. This is primarily because if we played Michigan 10 times, the metrics tell us that Michigan would probably win 7 or 8 of those matchups. What computers tend to somewhat negate is the outcome of one game sample-sizes and instead focus on overall efficiency in all of the games that you play.....weighted by SOS.

 

[Sooner04]

What an unnecessary comment with the only possible goal of creating non-constructive arguments and/or derailing a thread. A bit hypocritical how you’re doing the very thing here that many - yourself included - complain that other posters on this board do in creating unnecessary dissension, don’t you think? Can we all not agree to just be above this type of sarcastic, petty behavior?

I would use the ignore list, cbertus. It's made a big difference for me.

 

[stormspencer]

It is frustrating, bc the NET-Eff rating has us so low. Hopefully there will be some adjustments made.

For example, no reason to have Maryland above us. MD is above us in both barttorvik and KenPom, which is silly.

#16 MD - 9-2, SOS 344
#36 OU - 11-0, SOS 258

Why are they ahead of us? Bc of Off-eff rating...they have basically been curb-stomping weaker opponents with winning margins of 50, 34, 31, 71 and so on.

They also have 7 games against 250+ KP rated opponents, OU only has 5.

It will even itself out eventually, but definitely frustrating to say the least.

NET as it stands for OU:

Q1: 2-0
Q2: 2-0
Q3: 1-0
Q4: 6-0

You nailed it. In summation, you really need to beat the crap out of the gawd-awful opponents on you schedule for the computers to give you some grace. That means winning by 30+ against sub-300 teams.....not farting around for the majority of the game and winning by 15-20. It may or may not be fair....but it is what it is.

 

[SoonerBounce13]

Relying on computers will be not just college basketballs downfall but also humanity's downfall.

Hell, we already have AI telling kids to kill themselves and AI maliciously lying and copying itself to survive.

 

[WichitaSooner]

You are correct, but the problem is that our backups haven't necessarily played a ton against the bad opponents. Computers/advanced metrics would prefer that you absolutely dominate craptastic opponents. Unfortunately, in far too many games against the sub-300 teams this year, we've played fairly iffy and disinterested causing the games to be closer than the predictive scores would calculate.....thus affecting our overall numbers in the analytics. And again based on recent history, it doesn't necessarily bode well for OU moving forward. It also doesn't mean we are going to flop either, but we've seen this up close and personal the last several years and we got the short end of the stick.

Computers rely heavily on probabilities and simulations. For example, we beat Michigan by 1, but there is still a gap between them and us. This is primarily because if we played Michigan 10 times, the metrics tell us that Michigan would probably win 7 or 8 of those matchups. What computers tend to somewhat negate is the outcome of one game sample-sizes and instead focus on overall efficiency in all of the games that you play.....weighted by SOS.

Well said and a good explanation of how metrics and analytics work. And you’re right … our issue hasn’t been caused by walkons or even backups giving up points late. Heck, Cole didn’t even get it till the last couple minutes yesterday. Sadly, it looks like he has been relegated to basically playing with the walkons, which probably tells us that he isn’t likely to be much of a factor at OU.

I think some people generally dislike analytics, but sometimes they don’t even know why. I think they are necessary in college sports because teams play such wildly different schedules. For people who think results are all that should matter, that argument would lead to a bunch of teams from some terrible conferences being ranked in the top 15 or 20, because each year there will be a few teams that end up going 29-3 or 30-2 against bad competition.

The analytics take every possession of every game into account, and adjust for level of competition. As you point out, if you are going to load up with games against the worst of the worst, you need to absolutely pound them.

 

[daqman]

Week 7

Team	AP	Coaches	KenPom	BPI	SOR	SOS	NET
Tennessee	1​	1​	2​	4​	2​	37​	2​
Auburn	2​	2​	1​	1​	1​	8​	1​
Kentucky	4​	4​	11​	11​	4​	77​	7​
Alabama	6​	7​	8​	7​	3​	2​	10​
Florida	7​	6​	7​	16​	9​	91​	5​
Texas A&M	12​	11​	18​	20​	13​	10​	28​
Oklahoma	14​	15​	37​	39​	5​	186​	41​
Ole Miss	17​	16​	24​	26​	20​	171​	22​
Miss St	26​	25​	30​	29​	15​	62​	24​
Arkansas	27​	32​	40​	42​	31​	112​	47​
Missouri	28​	37​	49​	48​	32​	340​	29​
Georgia	38​	34​	39​	38​	17​	116​	26​
Texas	45​	NR​	38​	17​	44​	195​	44​
LSU	NR	NR​	59​	55​	47​	75​	64​
Vanderbilt	NR	NR​	62​	63​	33​	133​	54​
South Carolina	NR	NR​	68​	67​	99​	142​	99​

this is the week 7 poll.
we are 12/11 this week.

 

[SoonerBounce13]

Well said and a good explanation of how metrics and analytics work. And you’re right … our issue hasn’t been caused by walkons or even backups giving up points late. Heck, Cole didn’t even get it till the last couple minutes yesterday. Sadly, it looks like he has been relegated to basically playing with the walkons, which probably tells us that he isn’t likely to be much of a factor at OU.

I think some people generally dislike analytics, but sometimes they don’t even know why. I think they are necessary in college sports because teams play such wildly different schedules. For people who think results are all that should matter, that argument would lead to a bunch of teams from some terrible conferences being ranked in the top 15 or 20, because each year there will be a few teams that end up going 29-3 or 30-2 against bad competition.

The analytics take every possession of every game into account, and adjust for level of competition. As you point out, if you are going to load up with games against the worst of the worst, you need to absolutely pound them.

There was a time when analytics were not used in college basketball and it was just fine

 

[WichitaSooner]

There was a time when analytics were not used in college basketball and it was just fine

In your opinion. Which is fine, by the way. But there were also a lot of people who disagreed.

 

[TEvans4Three]

Relying on computers will be not just college basketballs downfall but also humanity's downfall.

Hell, we already have AI telling kids to kill themselves and AI maliciously lying and copying itself to survive.

There was a time when analytics were not used in college basketball and it was just fine

As someone who has made their career in analytics and data, I couldn't disagree with you more. Which is fine, I still respect your opinion on the matter.

I feel the same way with going to NBA games, I miss going to games with no virtual entertainment, no audio telling me how to cheer every 6 seconds and every possession. Every timeout something with activities...but the game isn't for me anymore, it is for entertainment, selling tickets, and the experience. While I disagree with it bc I love the purity of watching a game with shoes squeaking and yelling, I still get it. I just know that I am the grumpy old man in the stands, until I see my kid light up that the dancing bear is trying to dunk or seeing Red Panda. (to be fair, I love me some Red Panda)

Analytics are great for the game.

 

[skyvue]

OU's up to #12 in the AP poll (I guess Stephen Means got his stuff figured out).

 

[stormspencer]

Well said and a good explanation of how metrics and analytics work. And you’re right … our issue hasn’t been caused by walkons or even backups giving up points late. Heck, Cole didn’t even get it till the last couple minutes yesterday. Sadly, it looks like he has been relegated to basically playing with the walkons, which probably tells us that he isn’t likely to be much of a factor at OU.

I think some people generally dislike analytics, but sometimes they don’t even know why. I think they are necessary in college sports because teams play such wildly different schedules. For people who think results are all that should matter, that argument would lead to a bunch of teams from some terrible conferences being ranked in the top 15 or 20, because each year there will be a few teams that end up going 29-3 or 30-2 against bad competition.

The analytics take every possession of every game into account, and adjust for level of competition. As you point out, if you are going to load up with games against the worst of the worst, you need to absolutely pound them.

This is a parallel, but look no further than the initial 12-team college football playoff. You had two teams (Indiana and SMU) that had gaudy records and had beaten NO ONE. Neither team belonged in a playoff. Now neither of those teams were from "terrible" conferences, but they had the benefit of not having to play the better teams in their conference because of an unbalanced schedule. In Indiana's case, the one team they did play (Ohio State) ended up hammering them. I think analytics always have a place and should be given considerable weight when determining/comparing teams for the tournament. Having said that, analytics are NOT the "be all and end all", but no one can deny how effective of a tool that they can be when utilized appropriately.

 

[snydrosooner]

You are correct, but the problem is that our backups haven't necessarily played a ton against the bad opponents. Computers/advanced metrics would prefer that you absolutely dominate craptastic opponents. Unfortunately, in far too many games against the sub-300 teams this year, we've played fairly iffy and disinterested causing the games to be closer than the predictive scores would calculate.....thus affecting our overall numbers in the analytics. And again based on recent history, it doesn't necessarily bode well for OU moving forward. It also doesn't mean we are going to flop either, but we've seen this up close and personal the last several years and we got the short end of the stick.

Computers rely heavily on probabilities and simulations. For example, we beat Michigan by 1, but there is still a gap between them and us. This is primarily because if we played Michigan 10 times, the metrics tell us that Michigan would probably win 7 or 8 of those matchups. What computers tend to somewhat negate is the outcome of one game sample-sizes and instead focus on overall efficiency in all of the games that you play.....weighted by SOS.

But it also takes away from teams working on offensive sets in games. I saw a few times yesterday when ou could have attacked in the second half and pulled it out and ran a set play. If this is on purpose, then OU was getting practice against a live team running sets. That is worthwhile, analytics don’t like that though.

 

[snydrosooner]

You are correct, but the problem is that our backups haven't necessarily played a ton against the bad opponents. Computers/advanced metrics would prefer that you absolutely dominate craptastic opponents. Unfortunately, in far too many games against the sub-300 teams this year, we've played fairly iffy and disinterested causing the games to be closer than the predictive scores would calculate.....thus affecting our overall numbers in the analytics. And again based on recent history, it doesn't necessarily bode well for OU moving forward. It also doesn't mean we are going to flop either, but we've seen this up close and personal the last several years and we got the short end of the stick.

Computers rely heavily on probabilities and simulations. For example, we beat Michigan by 1, but there is still a gap between them and us. This is primarily because if we played Michigan 10 times, the metrics tell us that Michigan would probably win 7 or 8 of those matchups. What computers tend to somewhat negate is the outcome of one game sample-sizes and instead focus on overall efficiency in all of the games that you play.....weighted by SOS.

I have watched other teams in 30 point wins have starters in till the end. I get it, I just think you should give backups more time. Was not commenting on OU but more of the fact computers take spread into to much consideration. 15 point win or more is a good win, especially if backups get to play. And I don’t mean your walk ons.

 

[skyvue]

Look at the oswho game. We won by 15 but it simply was not that close. But the computers would never know that.

There's definitely a place and a purpose for analytics, but margin of victory should be way down the list of important stats--especially once you're past a 15-point margin.

 

[snydrosooner]

Well said and a good explanation of how metrics and analytics work. And you’re right … our issue hasn’t been caused by walkons or even backups giving up points late. Heck, Cole didn’t even get it till the last couple minutes yesterday. Sadly, it looks like he has been relegated to basically playing with the walkons, which probably tells us that he isn’t likely to be much of a factor at OU.

I think some people generally dislike analytics, but sometimes they don’t even know why. I think they are necessary in college sports because teams play such wildly different schedules. For people who think results are all that should matter, that argument would lead to a bunch of teams from some terrible conferences being ranked in the top 15 or 20, because each year there will be a few teams that end up going 29-3 or 30-2 against bad competition.

The analytics take every possession of every game into account, and adjust for level of competition. As you point out, if you are going to load up with games against the worst of the worst, you need to absolutely pound them.

They limited spread in the old BCS so teams did not feel they had to run it up. So should they do that in basketball? Say. You get credit for 20 point spread max in calculations? Just thoughts