[IUFR] How Indiana Botched the Metrics So Badly in 2023-24

Winning games will always be the most important thing. But in college hoops, metrics matter. Tony explains how IU botched the metrics this season, and why it's so important this doesn't happen again.

[Editor’s note: Our weekly IU Film Room posts are usually published on Fridays and reserved only for our paid subscribers. We are sending this bonus midweek edition out to everyone. Subscribe now to make sure you never miss a post from Tony!]

The Indiana Hoosiers sit at 14-10 overall with a 6-7 record in the Big Ten.

Not bad right? I mean, Michigan State is 15-9 with a 7-6 record in the Big Ten — one game better overall — and sit at a cozy #15 in KenPom.

The Hoosiers can’t be far behind, right?

RIGHT?

Well …

On paper, Indiana has won most of the games you’d expect them to win and lost most of the games you’d expect them to lose. In terms of wins and losses, the outliers are really only the home loss to Penn State and the road win at Ohio State.

Sure, you could argue for a couple of other games, but Indiana is largely winning games they’re projected to win and losing games they’re projected to lose.

The HOW is what is absolutely killing Indiana in the metrics that matter.

Currently, here’s where the Hoosiers sit in metrics that matter:

• KenPom: 96th … 81 spots below Michigan State!

• Bart Torvik: 88th

• NET: 100th

When you’re looking at KenPom rankings, you may wonder how does a team like Indiana (14-10) rank so far below Michigan State or only rank four spots higher than a team like Michigan (8-16).

Heck, Maryland is just 13-11 and lost to Indiana … yet the Terps are ranked 57th while the Hoosiers sit at 96th.

It all comes down to adjusted efficiency margin.

What is Adjusted Efficiency Margin, Why Does it Matter, and How Has Indiana Botched It?

A team’s adjusted efficiency margin is the difference between how many points per 100 possessions they score and how many points per 100 possessions they give up, adjusted for opponent strength.

The adjusted efficiency margin (or AdjEM) is what KenPom uses to rank teams.

Essentially, that AdjEM is how much better or worse you’re expected to be against an average opponent on a neutral floor if you played 100 possessions.

For Indiana this year, their AdjEM currently sits at 7.84. What this number tells us is that the Hoosiers would be expected to score 7.84 more points than an average team on an average floor if they both possessed the ball 100 times.

For context, Purdue’s AdjEM is 30.96 and they are 2nd in KenPom.

A 7.84 AdjEM is 96th best in the country, therefore Indiana is ranked 96th in KenPom.

Why Does AdjEM Matter?

As it was explained above, AdjEM is the sole driving force behind the KenPom rankings.

Adjusted efficiency is also a factor in the NET rankings and Bart Torvik.

While things like quad records, non-con records, and road/neutral performance are going to be a more important factor come Selection Sunday, adjusted efficiency still plays a part.

For me personally, IU’s KenPom ranking plays a big part in the psychology of how I think Indiana is doing, especially in a sport where schedule strengths can differ so much from one team to the next. Because it adjusts for the level of competition, a ranking like KenPom paints a pretty solid picture of how a team is actually playing beyond their simple won/loss record, especially in mid-February.

While a team’s KenPom ranking isn’t the end-all, be-all, it’s a pretty strong indicator.

How Has Indiana Botched Their AdjEM?

What’s important to remember about KenPom/AdjEM is that it’s a tempo-free, predictive efficiency metric.

In layman’s terms: it doesn’t care about wins and losses. It looks at how you perform versus how you were expected to perform and makes adjustments accordingly. And how you perform isn’t just a win or loss. It’s looking at the margin of the win/loss versus what was expected.

Let’s look at IU’s first game against Florida Gulf Coast. The Hoosiers won the game 69-63. All good, right? Meh.

Looking at that single game sample size, adjusting for FGCU’s offensive and defensive efficiencies throughout the year, IU actually performed at a level of -0.33. Essentially, just a hair less than average.

Same thing with Army. Hoosiers won 72-64. Army is one of the worst teams in the country, so factoring in their efficiencies, IU actually performed at a level of -7.99. That’s performing at a level of a team that is ~270th in KenPom, currently.

In other words: it was a win. Woo-hoo. But KenPom projects that 270ish other college basketball teams probably would have won by more on a neutral court than Indiana did at home. If it was just a one-off performance, fine, you write it off and forget about it. But when it becomes a trend, now your metrics really get weighed down.

A single individual performance isn’t usually going to skew the metrics a ton on its own. But over time, performances below expectation start to add up.

Here are the games just this season in which IU has performed with an AdjEM worse than a college hoops team in the KenPom top 100:

• Florida Gulf Coast

• Army

• Wright State

• UConn

• Auburn

• Morehead State

• Kennesaw State

• Nebraska

• Rutgers

• Purdue (2x)

• Penn State

You’ll notice on here that five of these games were wins against weak, non-conference opponents.

Several others were the blowout losses that have been an alarming trend for Indiana.

So how did Indiana botch their AdjEM?

They struggled with lowly opponents, oftentimes barely squeaking out wins, and then got blown out in several of their losses.

The easy retort to this is to say, “Well hey Numbers Guy, only wins and losses really matter.” And certainly the result is the most important thing. But comparing team records is very difficult when relative strength of schedules can be so different.

Let’s use Florida Gulf Coast for a deeper example.

They are currently ranked #252 in KenPom. They’ve played four non-conference opponents ranked in the top-100, losing to all of them. Here are the margins of defeat and the current KenPom ranking of each team:

• #33 Cincinnati: won by 27

• #76 Minnesota: won by 20

• #63 Pittsburgh: won by 12

• #96 Indiana: won by 6

Cincinnati is the 33rd-best team in the country and blew FGCU out by 27. Indiana is ranked 63 spots lower and only won by 6. Not every comparison will be this neat, as we all know college basketball produces plenty of outlier performances. But as a general rule, this is what you’ll see across college hoops.

One more example.

Say Indiana got a better performance from Xavier Johnson against UConn and made a few 3s, ultimately keeping it close and losing by just 4 instead of by 20.

Given how good UConn is, this would have really said something positive about Indiana. While that effort on that day couldn’t beat the #4 in the country, it might have beaten anyone outside the top-10. Thus, the Hoosiers likely would have moved up in the adjusted efficiency margin rankings instead tumbling down them.

What If Things Were Different?

I get it. ‘What ifs’ are frustrating to look at and don’t mean a lot because you can’t change the past.

HOWEVER …

I went through IU’s schedule and did some tinkering with the scores of games categorized as ‘stinkers’ to see what difference better performances in those games might have on the metrics, and in turn, how the season is perceived.

I essentially made it to where IU got sizable wins (10-20 point victories) against the teams they were supposed to dominate (FGCU, Army, Morehead St., Wright St., and Kennesaw State) and respectable losses (10-15 points) in the games they were blown out.

Again, none of these shifts change IU’s record at all. They still sit at 14-10 with these shifts.

Here’s a look at the shifts I made. The green columns are the adjusted scores, and the bold scores are the ones I actually adjusted. The real score is to the right in the last two columns.

Where would Indiana be sitting in KenPom if these were the scores of their games, still sitting at 14-10?

53rd.

A far cry from the 96th spot they currently sit at.

So why bring up this ‘what if’ if it can’t be changed? Because this is something that next year’s staff has to be cognizant of and make a priority.

Winning games against good teams will always be the most important thing you can do, there’s no question about it.

But a 53rd-ranked KenPom team at 14-10 would still have some hope down the stretch. If IU was hovering around 50th in KenPom with a road win at Ohio State, there might even be some bubble talk brewing about this team after getting their first quad 1 win.

These metrics matter, not only from a selection standpoint but the psychology of folks surrounding your team, as well.

Next season, these things should be taken in to consideration from Game One.

Example #1: Beat the brakes off of the teams you should.

All coaches will do some tinkering early in the season, which often includes trying different lineup combinations and giving young bench players more time on the court to develop.

But games in November count the same as games in March for NCAA Tournament resumes and for the metrics.

So IU might consider not using all-bench lineups early in the season as a way to get those guys experience together. Those decisions hurt you in the long-term if that lineup is consistently getting killed during its minutes. Besides, there is no real payoff because the all-bench (or even mostly-bench) lineup is never going to play later in the season in games that matter.

Example #2: Don’t let a loss snowball into a 20-point blowout.

This isn’t the NBA, where only wins and losses matter during the long, arduous 82-game regular season.

In the NBA, when it’s not a team’s night, and the result has been decided early on, the coach will often just call of the dogs, rest all the key players, and get ready for the next one. Losing by 10 or by 30 doesn’t really matter.

In college basketball, it does — whether any of us likes it or not.

And that is a key point to understand here: I’m not saying any of this is how it should be; I’m merely pointing out that it is.

And if Mike Woodson wants to put his IU program in the best position to earn NCAA Tournament berths, he has to show a greater understanding of college basketball math than he has through three seasons as IU’s coach. (Obviously there are lots of things we hope Woody shifts his thinking on. This is just one of ones he has the most control over through his decision making.)

The bottom line is that Indiana’s 2023-24 season has been a perfect storm of metric assassination. Playing close games against bad teams and getting blown out by good teams is a recipe for a team to take a meteoric fall from their preseason ranking, which is what Indiana has done this season.

With the heat slowly turning up on his seat here in Year Three, Mike Woodson can’t afford to let this happen again in Year Four.

Now over to you:

• Would you feel better about this Indiana team if they were sitting around 50th in KenPom?

• What do you think about the role analytics play in college basketball?

Until next time,

Coach Adragna

Founder, IU Film Room