What Do College Numbers Tell Us About Rookie-Year Numbers?

 In Draft Strategy, Rankings, Rookies, Sleepers, Statistical Analysis

Most rookies struggle to be more than late-round fantasy assets. Usually, there’s only two or three who are worth owning all year long and then a handful more become relevant late in the season when teams start playing for ping pong balls. There are always obvious rookie targets for fantasy owners. You don’t have to be a fantasy basketball prodigy to figure out that Zion Williamson and Ja Morant are probably going to perform well in their maiden campaigns. Deandre Ayton, Luka Doncic, and Trae Young didn’t exactly come out of nowhere. Rookies who are not obvious studs are much harder to predict. A player’s college numbers can give us an idea of what he will look like in his rookie year, but only if we know which numbers to look at. Not all college numbers are equally useful and some tell us almost nothing about a player’s rookie-year performance.

To help us understand how college performance translates, I put together a spreadsheet that examines how lottery picks from the past five drafts (excluding international players and Markelle Fultz) performed in their final year in college and in their rookie year. The statistics that I chose to look at (AST%, STL%, BLK%, etc.) are all percentage statistics. I chose these because they are not impacted by minutes played or pace. I then ran a series of simple regression analyses to determine if a player’s performance in a statistic in his final year in college would be a strong indicator of his performance in that same statistic in his rookie year. The R^2 generated from the regression analyses are as follows:

 

*For those with math skills that are a little rusty, here is a quick rundown on how R^2 works. R^2 tells us what proportion of the variance in the dependent variable can be explained by the independent variable. So in our case, it tells us what proportion of the variance in rookie-year numbers can be explained by college numbers. In other words, it tells us how useful a college statistic is at predicting rookie-year performance in that same statistic. For example, there is a strong relationship between college BLK% and rookie-year BLK% percentage. The R^2 of .7826 means that 78.26 percent of a player’s rookie-year BLK% can be explained by his college BLK%. Therefore, a player with a strong BLK% in college will likely have a strong BLK% as a rookie. What R^2 does NOT tell us is that if a player blocks X amount of shots in college then he will block X amount of shots as a rookie. Here is a breakdown of what is considered a strong and weak relationship between two variables:

0-.30 = weak relationship

.31-.50 = low relationship

.51-.70 = moderate relationship

.71-1.0 = strong relationship

 

And now a closer look at each R^2 number and what they tell us about this year’s rookie crop:

Field goal percentage

Field goal percentage: .2794

It should not be a surprise that college field goal percentage numbers do not tell us a lot about rookie-year field goal percentage numbers. It’s rare for a player’s role to stay the same after he makes the jump to the pros. Rookie field goal percentage is also going to be impacted by the team that a player is selected by. A player is going to get easier looks if he plays for the Celtics instead of the Hornets. This weak relationship tells us that De’Andre Hunter is not a lock to shoot well from the field as a rookie even though he hit 52.0 percent of his shots during his final year at Virginia.

Notable rookies with high college FG%: Zion Williamson, De’Andre Hunter, Jaxson Hayes, Rui Hachimura, Cameron Johnson, Brandon Clarke, Grant Williams, Dylan Windler, Bruno Fernando

Notable rookies with low college FG%: Coby White, Cam Reddish, Matisse Thybulle

Free throw percentage

Free throw percentage: .3261

Free throw rate: .3934

College free throw percentage not having a strong relationship with rookie-year free throw percentage is likely a sample-size issue. Most rookies do not shoot a ton of free throws due to their decreased role on offense which leads to a high amount of variance. The relationship is likely stronger for players who get to the line regularly in their rookie years. This means that Zion Williamson’s college free throw percentage is probably more predictive of his rookie-year free throw percentage than this R^2 suggests.

Notable rookies with high college FT%: Ja Morant, Tyler Herro, Matisse Thybulle, Grant Williams, Dylan Windler, Jordan Poole

Notable rookies with low college FT%: Zion Williamson, R.J. Barrett, Jarrett Culver, Brandon Clarke, P.J. Washington

Points

Points Per 100 Possessions: .0321

Usage percentage: .0503

This is another obvious result. Most NBA-level college prospects are prolific scorers at the college level but few become high-volume scorers in their rookie year.

Notable rookies with high college USG%: Zion Williamson, Ja Morant, R.J. Barrett, Jarrett Culver

Notable rookies with low college USG%: Jaxson Hayes, Matisse Thybulle

Three-pointers

Three-point percentage (excludes zero/minimal attempt players): .0012

Three-point rate (includes zero/minimal attempt players): .5049

Three-point rate (excludes zero/minimal attempt players): .3284

College FT% to Rookie-year 3p%: .3286

College three-point percentage tells us almost nothing about how a player will shoot from deep as a rookie. College free throw percentage is a better, but not great, indicator of success or failure from three at the NBA level. This is part of the reason why I am bullish on Tyler Herro’s potential from deep even though the former Wildcat only hit 35.5 percent of his attempts from three in college. While at Kentucky, Herro connected on a very impressive 93.5 percent of his free throw attempts. Brandon Ingram is the best recent example of the importance of keeping college free throw percentage in mind when judging a player’s potential from three. At Duke, Ingram averaged 2.2 3PG while shooting 41.0 percent from three. He also hit a paltry 68.2 percent of his attempts from the charity stripe. Ingram then went on to average only 0.7 3PG as a rookie while shooting 29.4 percent from three.

Notable rookies with high college 3P%: De’Andre Hunter, Cameron Johnson, Dylan Windler, Ty Jerome

Notable rookies with low college 3P%: R.J. Barrett, Jarrett Culver, Romeo Langford, Matisse Thybulle

Rebounds

Offensive rebounding percentage: .6966

Defensive rebounding percentage: .6237

Total rebounding percentage: .7602

If you’re a strong rebounder in college, then there’s a good chance you’ll be a strong rebounder at the NBA level. If you’re a weak rebounder in college, that probably won’t change once you reach the pros. Only BLK% comes with a higher R^2 than TRB%.

Notable rookies with high college TRB%: Zion Williamson, Jaxson Hayes, Rui Hachimura, P.J. Washington, Chuma Okeke, Brandon Clarke, Grant Williams, Dylan Windler, Mfiondu Kabengele, Bruno Fernando

Notable rookies with low college TRB%: Coby White, Cam Reddish, Tyler Herro, Matisse Thybulle, Ty Jerome, Jordan Poole, Kevin Porter Jr.

Assists

Assist percentage: .5294

An R^2 of .5294 suggests a moderate, but not strong, relationship between college assist numbers and rookie-year assist numbers. This moderate relationship likely has something to do with point guards’ styles usually not changing when they hit the NBA. Pass-first point guards rarely turn into shoot-first guards in the pros and point guards who functioned more as combo guards in college usually aren’t asked to be pass-first point guards once they hit the NBA.

Notable rookies with high college AST%: Ja Morant, R.J. Barrett, Jarrett Culver, Coby White, Nickeil Alexander-Walker, Ty Jerome

Notable rookies with low college AST%: All first-round big men, Cam Reddish, Keldon Johnson, Kevin Porter Jr.

Steals & Blocks

Steal percentage: .6300

Block percentage: .7826

Defensive numbers translate better than offensive numbers when a player moves from the college game to the pro game. Both steal percentage and block percentage are moderate-to-strong predictors of NBA performance. This means that Zion Williamson (3.9 STL%, 5.8 BLK%) is a good bet to be a monster source of defensive numbers right out of the gate and that De’Andre Hunter (1.2 STL%, 2.4 BLK%) is likely going to drag down both your steals and blocks if you select him.

Notable rookies with high college STL%: Zion Williamson, Ja Morant, Jarrett Culver, Cam Reddish, Chuma Okeke, Nickeil Alexander-Walker, Matisse Thybulle, Ty Jerome

Notable rookies with low college STL%: R.J. Barrett, De’Andre Hunter, Romeo Langford

Notable rookies with high college BLK%: Zion Williamson, Jaxson Hayes, P.J. Washington, Chuma Okeke, Matisse Thybulle, Brandon Clarke, Grant Williams, Mfiondu Kabengele, Bruno Fernando

Notable rookies with low college BLK%: R.J. Barrett, Cameron Johnson, Tyler Herro, Ty Jerome, Jordan Poole, Keldon Johnson

Turnovers

Turnover percentage: .2956

There is a weak relationship between college turnover rate and rookie-year turnover rate for the same reason there is a weak relationship between college field goal percentage and college scoring rates and their NBA equivalents. Turnover rate is going to be dependent on the size of a player’s role and a player’s role almost always changes when he goes pro.

Notable rookies with high college TOV%: Zion Williamson, Ja Morant, R.J. Barrett, Jarrett Culver, Coby White, Cam Reddish, Nickeil Alexander-Walker

Notable rookies with low college TOV%: De’Andre Hunter, Rui Hachimura, Cameron Johnson, Brandon Clarke, Ty Jerome

 

So what does this all mean and how does this apply to fantasy basketball?

Players with strong rebounding, assist, steal, and block numbers in college are safer picks in fantasy basketball drafts than players whose lines look good because they excel in the scoring and shooting categories. If a player posts strong block numbers in college, he is a good bet to post strong block numbers at the next level. If a player is an accurate shooter from deep in college, then maybe he will be a good shooter in the pros, but his college success rate cannot be relied upon to determine that. This is consistent with what types of players tend to be valuable in their rookie years. Point guards and big men, the types of players that usually post strong assist, rebounding, and defensive numbers, are almost always better bets to be solid fantasy assets in their rookie year than wings. Wings are not always asked to create in college, are usually forgettable rebounders, and trail behind point guards in the steals department and bigs in the blocks category. From the 2012-2013 season to the 2016-2017 season this was especially true. Over that span, not a single wing finished in the top-100 in nine-category leagues in their rookie year.

This analysis suggests that players like Brandon Clarke, Matisse Thybulle, and Grant Williams are probably going to be picked too late in deeper re-draft leagues this year and players like R.J. Barrett and De’Andre Hunter will be picked too high. This analysis also suggests that Zion Williamson and Ja Morant are safe picks as far as rookies go. Both excel in the areas that tend to translate to the pro game right away. For those hunting for deep rookie sleepers, keep an eye on Bruno Fernando, a second-round pick of the Hawks who could win the backup center job in camp. Fernando posted some monster TRB% and BLK% numbers in college and should be an excellent source of both boards and blocks whenever he receives extended run.

Follow me on Twitter @AdamGStock for the latest fantasy basketball news and analysis. I will gladly answer any fantasy basketball questions that you may have. 

 

 

 

 

 

Recommended Posts
Showing 2 comments
  • Jonathan Chown

    “The statistics that I chose to look at (AST%, STL%, BLK%, etc.) are all percentage statistics. I chose these because they are not impacted by minutes played or pace.”

    But they are affected by strength of schedule, which could be a concern with Ja Morant.

    “College three-point percentage tells us almost nothing about how a player will shoot from deep as a rookie. College free throw percentage is a better indicator of success or failure from three at the NBA level.”

    You see this said quite often in fantasy circles, but the R-squared isn’t terribly convincing. Simply being better than another indicator doesn’t make it a good one by itself.

    “An R^2 of .5294 suggests a moderate, but not strong, relationship between college assist numbers and rookie-year assist numbers. This moderate relationship likely has something to do with point guards’ styles usually not changing when they hit the NBA.”

    This doesn’t really make sense to me. If styles don’t change, shouldn’t the R-squared be higher?

    And finally, can this information be worked into any kind of credible multivariate regression analysis to try to estimate players’ overall value, or is it solely for hunting categories?

    • Adam Stock

      Great questions/critiques.

      Re: Morant – Agreed. Unfortunately, the number of high-lottery picks from minor conferences is never going to be high so we’re stuck throwing guys like Morant into the bigger pool.

      Re College FT% – Agreed. That’s why I said it’s something to keep in mind when projecting rather than saying it was a great indicator. The moral of the story is that rookie 3P% is a crapshoot.

      Re: Assists – I should have made that argument a little more clear. I’m guessing it’s moderate and not higher because of the non-point guard positions. I should do the math but by eye-balling my data it seems like there are bigger swings in college wing AST%, which makes sense since best player on the team = more creation responsibilities regardless of position. Those creation duties then disappear when they hit the next level.

      Re: Regression – That would be an absolute beast of an analysis and probably above my paygrade but I don’t see why it couldn’t be. I’m sure teams all have their models for how college numbers translate to rookie/career performance.

Leave a Comment

Start typing and press Enter to search