Quantifying the Role of Experience in College Basketball Roster Building
We analyzed nearly 1,500 rosters from the past four years to see what the data reveals about freshmen, transfers, and the “universal truths” of building a winning NCAA team.
For a while, I’ve been obsessing over roster building in the new era of college basketball. In a market where the only constant has been change, identifying where competitive advantages can be found consistently should, in theory, allow teams to level the playing field in a sport where team budgets (and, therefore, NIL budgets) vary more wildly than in any other sports league in the US.
In March, I wrote about the underrecruited areas of both the US and the rest of the world and proposed that recruiting (a) internationally, (b) in certain regions of the States not considered high school basketball hotbeds, and despite what certain TV pundits will tell you, (c) public high school programs, can lead to more recruiting steals than simply targeting the same states and the same high-level high school programs that everyone else scouts.
Right around the time I was putting that article together, St. John’s basketball coach Rick Pitino also talked about high school recruiting in the new era of college basketball, proposing something completely opposite to what my article described: don’t even attempt to get high school recruiting steals, simply stop recruiting high school players altogether.
Tiktok failed to load.Enable 3rd party cookies or use another browser
In the 2024-25 season, transfer players accounted for 42% of all available minutes in Division I basketball, a sharp increase from 27% just four years earlier.
So Pitino has perhaps been the most open in acknowledging the new reality of college basketball, where many programs share a similar vision that seems to be based on two arguments widely accepted as universal truths in college sports.
Two Universal Truths in College Sports
(1) You can’t win relying on freshmen
(2) You need experience on the roster to win big
Those two cornerstone beliefs seem to drive this transfer-centric vision of NCAA recruiting. If your experienced seniors are graduating, you need to get that experience from somewhere. With transfers developing into the college equivalent of NBA free agency, you can simply recruit experienced players from other schools, which, based on those two universal truths, should translate into more wins.
But then, on the other hand, there’s Dabo Swinney.
For those not familiar with college football, Swinney is one of the winningest active coaches in the sport. He’s won nine conference titles and two National Championships in his 16 full seasons at Clemson, famously doing so without recruiting transfers. The Tigers ranked 82nd in the transfer portal rankings this year according to 247Sports, which is actually their best ranking since this list has been compiled (unranked in 2024, 102nd in 2023, and 156th in 2022).
This, however, has brought him a fair share of criticism, especially during seasons like this year when Clemson doesn’t look like a National Championship contender and every game becomes a referendum on Swinney’s approach to roster building.
While it would be easy to look at Pitino’s and Swinney’s approaches as opposites—the former argues for not recruiting high school players, while the latter only recruits high school players—I see them as both sides of the same coin. Both prioritize experience, but one approach obtains it from external sources while the other develops it in-house.
The question now is: what do the numbers say about either approach? Is there truly a right way to build a college basketball roster?
The Method
To do that, I took all 1,450 rosters in college basketball between the 2021-22 and the 2024-25 seasons, and I divided the players on those rosters into three categories: returners, freshmen, and transfers1.
From there, I calculated the percentage of total minutes each team gave to players in each category to build a profile of every roster and see if those distributions correlated with winning.
To cross-reference each profile with team performance without making the exercise unnecessarily complex, I had to condense each team’s performance into a single number. While I considered going with a simple Win/Loss Record, I ultimately chose the Simple Rating System (SRS) from Sports Reference, as it also incorporates strength of schedule.
Here’s the table of all the rosters (the data is downloadable, too; you’re welcome).
With the data on full display, now it’s time to answer questions by looking at the correlation between performance and minute share by freshmen, returners, and transfers. Given that in statistics you rarely find a direct, one-to-one relationship between variables, we’ll be using three methods of evaluation.
Linear Correlations
This is the simplest way to look at things. Let’s put all 1,450 teams on a scatter plot with two axes—minute share by category and SRS—and see what the linear correlations are.
While there’s a slight positive correlation between returner minutes and SRS, and an even slighter negative correlation between both transfer and freshmen minutes and SRS, it’s certainly not enough to derive overarching conclusions about the pros and cons for each category, given how weak the correlations are.
Distribution Tails
Considering that there’s no strong linear correlation between minute shares and winning, we could look at how teams at the extremes of each category’s minute distribution performed to identify potential patterns.
Let’s look at the top and bottom 72 teams in minute share for each category, which represents the upper and lower 5% of cases.
While again, there’s not an overwhelming answer, it’s yet another evidence, as slight as it is, that more returning players tend to lead to better SRS, and that more freshmen tend to lead to worse SRS. When it comes to transfers, the effects were negligible.
Elite Teams
The final approach is to just focus on the cream of the crop: find out how the top 5% teams in college basketball built their rosters in comparison with the national average, in order to find minute-share patterns worth emulating.
As we can see, in the top 72 teams of the past four seasons (which includes all four national champions), freshmen played 14% of all available minutes, compared to just 11% for all college basketball teams. There’s also a slight positive differential for returners, who played 57.5% of all minutes for elite teams vs. 55.5% of minutes for all teams.
That leaves transfers as the only category with a negative differential: 33.3% of all minutes in college basketball vs. 28.3% minutes for elite teams.
The final tally
Returners: Positive linear correlation, positive distribution-tail differential, positive elite-team share.
Freshmen: Negative linear correlation, negative distribution-tail differential, positive elite-team share.
Transfers: Negative linear correlation, neutral distribution-tail differential, negative elite-team share.
Two Universal Truths, Revisited.
(1) You can’t win relying on freshmen.
I would say the evidence here is rather inconclusive. The linear correlation is weak at best. The success cases show that elite teams actually play their freshmen more than the average team. And when looking at the extreme cases, nearly 60% (70 teams out of 117) of teams where freshmen didn’t play a single minute had a negative SRS, so barring freshmen from playing didn’t exactly work out for those squads.
At the same time, some of the most freshmen-reliant teams over the past four years have also been among the absolute best. The top team in SRS last season was a Duke squad that reached the Final Four, with freshmen accounting for 52.4% of all minutes, the fifth-highest share in the past four years. Similarly, despite an early Sweet Sixteen exit to San Diego State, 2022-23 Alabama was the top team in SRS that season, with freshmen playing 45.6% of their minutes, a top 15 mark in this exercise.
The rebuttal here is rather easy: you’re obviously going to win with freshmen when your freshmen are named Cooper Flagg and Brandon Miller. The evidence supports that—only four of the teams with the highest freshmen minute share had a positive SRS without having a single RSCI Top 100 freshman: 2024-25 Kennesaw State, 2023-24 UMass, 2022-23 Oregon St., and 2021-22 College of Charleston.
Ultimately, this exercise shows that lack of experience doesn’t imply a lack of greatness, and teams that are able to recruit quality freshmen should continue to do so; unlike what the universal truth suggests, giving them minutes won’t (necessarily) lead to disaster.
(2) You need experience on the roster to win big.
Looking at all three tests for both returning-player minutes and transfer minutes, let me ask a question about this universal truth: will any type of experience do?
You will always need experience to win. However, considering that all three tests were positive for returner minutes and negative for transfer minutes, your energy (and NIL money!) is likely better spent on creating homegrown returners than trying to poach them from somewhere else.
Creating those homegrown returners goes hand in hand with recruiting high school prospects. Remember those teams in the distribution tails we listed earlier? The teams that gave freshmen the most minutes improved their SRS by an average of 1.8 points the next season, while the teams that gave freshmen no minutes saw their SRS decrease by an average of 1.4 points.
Does this mean that I foresee teams leaning more into traditional high school recruiting and ditching the transfer portal? Well, considering that in 2025 the transfer portal hit record numbers yet again, I’d say we’re far from it. In fact, I’d bet that this upcoming season might be the first where transfers play more minutes than returners.
If there’s any truth in the emotional cycles of the stock market, the rising number of transfers being recruited, combined with the lack of evidence that playing more transfers leads to more wins, suggests that the market for transfers might be approaching the “euphoria” stage. On the other end of the spectrum, other forms of recruiting—namely returners and, especially, freshmen—appear to be near the “capitulation” stage, despite the numbers proving them as viable ways to build a winning roster.
In short, I don’t see transfer recruiting going away anytime soon, but I’m willing to bet that a smart team, sooner rather than later, is going to “buy the dip” and find tremendous value by going back to basics, investing on returners, and betting on the freshmen market.
Do the Universal Truths of College Basketball Roster Building Stand?
Despite me spending the last 2,000 words trying to show how data challenges the conventional wisdom behind college basketball roster building in 2025, I do think the core premises still hold up with a few key qualifiers. I think that (1) you can win with key quality freshmen and that (2) you need returning players to win big.
However, at the end of the day, if there’s a concept from this exercise that I think could develop into something remotely similar to a universal truth, it’s this: you need greatness to win, and greatness can come from anywhere. Returners can help you win. Quality freshmen can help you win. Key transfers can help you win.
The transfer portal is certainly the shiny new thing, but the successful teams of recent years show that it’s not wise to close the door to any avenue for obtaining talent.
Some definitions are needed here, as there are a few grey areas. “Returners” are players who stayed with the same team from their previous season or, if they didn’t play the prior year, from two seasons ago. “Freshmen” are players in their first year of college basketball. “Transfers” are players who were on a different team the previous season or who returned after two or more years of inactivity.