I like the explanation from Tony, but one of my pet peeves about all the analytics is that they don't account for actual performance on the floor. A team may not be playing to its potential, but all that matters at the end of the day is the W or the L. I know, I know, that is heresy to those of us who grew up with Bob Knight, but you are ultimately judged by your wins and losses. As Bill Parcells famously said, "you are what your record says you are". Even when I coached, I would rather win ugly than lose pretty.
Well, ironically, they do actually account for performance on the floor. That's the whole point: how did the team play, independent of the result? That way, a one-point loss to a great team can be weighted appropriately next to a 5-point win over an awful team where you might have played like crap but still escaped with the W. The record is, of course, the most important thing at the end of the day, but the analytics allow a different way for comparing teams when strengths of schedule are so different across the country.
Head to head does matter in standings and tie-breakers, but that's not what KenPom is attempting to judge. This is where it is really important to judge a rating system based on its goal. KenPom is trying to compare teams' relative performance on a per-possession basis. I agree it seems odds for IU to be ranked lower than a team it beat in their gym, but that's due in part to Michigan opening the season with a higher ranking and those preseason priors still being baked in.
It seems like the first part of that calculation related to previous year minutes played needs to be adjusted. Guys moving in the transfer portal has changed the game.
Perhaps, but I also think there's some merit to it. Take, for instance, a team like Purdue. They return most of their minutes and they look like one of the best teams in the country in November. Same can be said for teams like Houston, UConn, and Marquette (all in the top 5 on KenPom).
Then, take Indiana. Large roster turnover and their performances at the beginning of the year were pretty ugly. In fact, their performances were actually worse than what KenPom was predicting.
Pomeroy has tweaked his algorithm several times, so I'm sure he's looking at trends to see if/where things need adjusted year to year.
I review KenPom after every IU game. I was wondering about IU’s low score but also knew we’ve only played one solid opponent, UConn. Great breakdown of how these rankings work. I predict IU will finish at 35 on KenPom,
I never check KenPom, in answer to your query! It may be generational, I'm old, but I concentrate on the play on the floor, the team on the floor and the results. Results matter to me. I enjoy all of this analytics discussion and kerfluffle, but in a relatively busy life, I usually skip charts, grafts and predictive narrative. Maybe I'm just lazy.
It's to each their own. I know a lot of people who find all of the advanced metrics trying to explain underlying factors in sports to be tiresome and remove some of the "magic" of just watching the game and focusing on the result. I enjoy the challenge of trying to incorporate it all while the finding the right balance and context.
Thanks for that breakdown. I always knew it was more than win/loss, and the team's progression actually mirrors my personal intuition regarding the team. They've shown poor, but improving, play as games progress. If they get all the moving pieces together they'll be a pretty good team.
Thanks Tony. I love this stuff. I tend to focus on the things that we are clearly an outlier on (good and bad) and try to understand whether they will continue. At some point in the season, you usually are who you are and what the analytics are pointing out but early in the season there is still some chance to shape things. For example, the often discussed three point shooting / rate / percentage of offense, etc. IU is not going to suddenly become Iowa, but I do hold out some hope that we can shape those numbers somewhat. By being a fan of the team you are more aware than computers on what might transpire. X could get healthy and give us a boost, CJ could start to make a few, Galloway might regain some confidence. All are still possible I think but won't show up in the analytics until we demonstrated them with some regularity. If this stuff was truly the end all there wouldn't be a Vegas. Thanks again.
You're exactly right. An example of something that was a huge outlier was Indiana's 3-point defense. At one point, opponents were shooting almost 40% from three and the Hoosiers were almost last in the country in that regard. Now, opponents are shooting 32.3% from three, which actually has Indiana better than average.
With these early-season trends, just a few games can swing the data so much, so it's always important to remember that.
I started looking at KenPom with some regularity in the past couple of years (mostly because my analytics-minded husband got interested in the numbers). Mostly I looked at KenPom out of curiosity, understanding only peripherally how the ranking are calculated. Thanks for a clear explanation of KenPom, Tony!
Tony, great analysis. Didn’t know most of this. Hypothetical question... Should IU beat both Auburn and Kansas, we’d be 9-1 and certainly in the Top 25 in polls. Where would that put us in KenPom?
The answer will depend on how much Indiana outperforms the analytical expectation.
Currently, IU is a 7-point underdog on a neutral court to Auburn, and a 6-point underdog at home to Kansas. If the Hoosiers outperform those results, even in losses, their KenPom ranking is likely to increase. If we WIN both of those games, I would expect a jump of somewhere between 15-20 spots at a minimum.
Thanks, Michael! Jerod explained it very well. If Indiana can outpace those predictions (which they would pretty heavily with wins), they'd definitely shoot up KenPom, as well. How far? Not entirely sure, but I'd guess they'd be in the top 40.
Dec 8, 2023·edited Dec 8, 2023Liked by Tony Adragna
I’m not too into the analytics, but I do like how he gradually weeds out the past over time. It makes sense why we’re lower right now. The eye test, I think we can probably all agree on, shows us this team is improving and feels better than the current ranking. If we continue to improve the rankings will take care of themselves. It’s so nice to see our team improving compared to before Woody when it seemed like we never showed growth. Nice explanation. Thank you & Go Hoosiers.
A great explanation and analysis Tony. I think as the season rolls on, more teams get into conference play and the sample size gets into this year, IU will be more top 10% of programs than top 20%.
KenPom is worth my $20 a year, if for no other reason, because it's a great place for aggregated data on specific team and player performance spanning 25 years.When I'm looking at a team playing this season, the site tracks overall strengths and weaknesses. The individualized player data and lineup frequency tracking helps deep dives into a team's identity.
The overall rankings are important and it's a good addition to Torvik, NET and polling on evaluating how teams are actually doing. Still, at this point in the season and throughout, it's a good guide to how teams play, who they have on the court and what their resume is to date.
Good explainer, Tony. Because the rating services have radically different comp weightings on current vs past performance, I generally shift between Torvik and KenPom depending on the time of the year.
KenPom seems to hate Indiana beginning of every season that I have been looking at it. Any analytics that put anything over winning I'm not a fan of. The committee seems to care more about that then analytics so seems like that should be the main factor in any ranking systems, no?
Pols are just pols, I have no faith in them. Look at what the score is when the buzzer sounds.
THESE KIDS are finding ways to win, also faith in the TEAM. The bench is now scoring , learning ther roll
to be part of the TEAM, getting there minutes and growing for FUTURE. It will not happen in a couple of years. Let us be happy they are finding the way to be IU, our fans will learn to enjoy this season.
I like the explanation from Tony, but one of my pet peeves about all the analytics is that they don't account for actual performance on the floor. A team may not be playing to its potential, but all that matters at the end of the day is the W or the L. I know, I know, that is heresy to those of us who grew up with Bob Knight, but you are ultimately judged by your wins and losses. As Bill Parcells famously said, "you are what your record says you are". Even when I coached, I would rather win ugly than lose pretty.
Well, ironically, they do actually account for performance on the floor. That's the whole point: how did the team play, independent of the result? That way, a one-point loss to a great team can be weighted appropriately next to a 5-point win over an awful team where you might have played like crap but still escaped with the W. The record is, of course, the most important thing at the end of the day, but the analytics allow a different way for comparing teams when strengths of schedule are so different across the country.
But at the end of the day, head-to-head has to matter. For example, an IU win at Michigan and we are still below them.
Head to head does matter in standings and tie-breakers, but that's not what KenPom is attempting to judge. This is where it is really important to judge a rating system based on its goal. KenPom is trying to compare teams' relative performance on a per-possession basis. I agree it seems odds for IU to be ranked lower than a team it beat in their gym, but that's due in part to Michigan opening the season with a higher ranking and those preseason priors still being baked in.
Thanks for a great, clear explanation. I look at KenPom occasionally, but now understand it better. IU will be top 20, but I’m an optimist.
Absolutely - thanks for reading!
It seems like the first part of that calculation related to previous year minutes played needs to be adjusted. Guys moving in the transfer portal has changed the game.
Perhaps, but I also think there's some merit to it. Take, for instance, a team like Purdue. They return most of their minutes and they look like one of the best teams in the country in November. Same can be said for teams like Houston, UConn, and Marquette (all in the top 5 on KenPom).
Then, take Indiana. Large roster turnover and their performances at the beginning of the year were pretty ugly. In fact, their performances were actually worse than what KenPom was predicting.
Pomeroy has tweaked his algorithm several times, so I'm sure he's looking at trends to see if/where things need adjusted year to year.
In reality though, all I care about is 7 W and 1 L. Next game, survive and advance. Stay healthy. Go Hoosiers.
Very possible. This is why some people prefer T-Rank or Haslametrics. They each have their different biases.
I review KenPom after every IU game. I was wondering about IU’s low score but also knew we’ve only played one solid opponent, UConn. Great breakdown of how these rankings work. I predict IU will finish at 35 on KenPom,
Thanks for reading, Brian! I'm right in line with your thinking on where they finish.
I never check KenPom, in answer to your query! It may be generational, I'm old, but I concentrate on the play on the floor, the team on the floor and the results. Results matter to me. I enjoy all of this analytics discussion and kerfluffle, but in a relatively busy life, I usually skip charts, grafts and predictive narrative. Maybe I'm just lazy.
It's to each their own. I know a lot of people who find all of the advanced metrics trying to explain underlying factors in sports to be tiresome and remove some of the "magic" of just watching the game and focusing on the result. I enjoy the challenge of trying to incorporate it all while the finding the right balance and context.
Thanks for that breakdown. I always knew it was more than win/loss, and the team's progression actually mirrors my personal intuition regarding the team. They've shown poor, but improving, play as games progress. If they get all the moving pieces together they'll be a pretty good team.
Agreed! It's how I felt this team would look, as well. It made a lot of sense given the roster turnover and all of the new pieces.
Thanks Tony. I love this stuff. I tend to focus on the things that we are clearly an outlier on (good and bad) and try to understand whether they will continue. At some point in the season, you usually are who you are and what the analytics are pointing out but early in the season there is still some chance to shape things. For example, the often discussed three point shooting / rate / percentage of offense, etc. IU is not going to suddenly become Iowa, but I do hold out some hope that we can shape those numbers somewhat. By being a fan of the team you are more aware than computers on what might transpire. X could get healthy and give us a boost, CJ could start to make a few, Galloway might regain some confidence. All are still possible I think but won't show up in the analytics until we demonstrated them with some regularity. If this stuff was truly the end all there wouldn't be a Vegas. Thanks again.
Thanks for reading, Mark!
You're exactly right. An example of something that was a huge outlier was Indiana's 3-point defense. At one point, opponents were shooting almost 40% from three and the Hoosiers were almost last in the country in that regard. Now, opponents are shooting 32.3% from three, which actually has Indiana better than average.
With these early-season trends, just a few games can swing the data so much, so it's always important to remember that.
I started looking at KenPom with some regularity in the past couple of years (mostly because my analytics-minded husband got interested in the numbers). Mostly I looked at KenPom out of curiosity, understanding only peripherally how the ranking are calculated. Thanks for a clear explanation of KenPom, Tony!
Sure thing, Lisa! Thanks for reading!
^good summary of where IU sits but also good to check Haslam and EvanMiya too
https://bracketologists.com/team/indiana-hoosiers
Tony, great analysis. Didn’t know most of this. Hypothetical question... Should IU beat both Auburn and Kansas, we’d be 9-1 and certainly in the Top 25 in polls. Where would that put us in KenPom?
The answer will depend on how much Indiana outperforms the analytical expectation.
Currently, IU is a 7-point underdog on a neutral court to Auburn, and a 6-point underdog at home to Kansas. If the Hoosiers outperform those results, even in losses, their KenPom ranking is likely to increase. If we WIN both of those games, I would expect a jump of somewhere between 15-20 spots at a minimum.
Thanks, Michael! Jerod explained it very well. If Indiana can outpace those predictions (which they would pretty heavily with wins), they'd definitely shoot up KenPom, as well. How far? Not entirely sure, but I'd guess they'd be in the top 40.
He's gonna tell you to refer to the formulas. All those efficiency formulas are the factors, not whether they win or lose.
But I'm hoping. 😉
I’m not too into the analytics, but I do like how he gradually weeds out the past over time. It makes sense why we’re lower right now. The eye test, I think we can probably all agree on, shows us this team is improving and feels better than the current ranking. If we continue to improve the rankings will take care of themselves. It’s so nice to see our team improving compared to before Woody when it seemed like we never showed growth. Nice explanation. Thank you & Go Hoosiers.
Thanks for reading, Michael!
A great explanation and analysis Tony. I think as the season rolls on, more teams get into conference play and the sample size gets into this year, IU will be more top 10% of programs than top 20%.
KenPom is worth my $20 a year, if for no other reason, because it's a great place for aggregated data on specific team and player performance spanning 25 years.When I'm looking at a team playing this season, the site tracks overall strengths and weaknesses. The individualized player data and lineup frequency tracking helps deep dives into a team's identity.
The overall rankings are important and it's a good addition to Torvik, NET and polling on evaluating how teams are actually doing. Still, at this point in the season and throughout, it's a good guide to how teams play, who they have on the court and what their resume is to date.
Thanks, Bob! Totally agree on all points.
Good explainer, Tony. Because the rating services have radically different comp weightings on current vs past performance, I generally shift between Torvik and KenPom depending on the time of the year.
You and I, both!
KenPom seems to hate Indiana beginning of every season that I have been looking at it. Any analytics that put anything over winning I'm not a fan of. The committee seems to care more about that then analytics so seems like that should be the main factor in any ranking systems, no?
Pols are just pols, I have no faith in them. Look at what the score is when the buzzer sounds.
THESE KIDS are finding ways to win, also faith in the TEAM. The bench is now scoring , learning ther roll
to be part of the TEAM, getting there minutes and growing for FUTURE. It will not happen in a couple of years. Let us be happy they are finding the way to be IU, our fans will learn to enjoy this season.