Illini Basketball 2026-2027

Fly Illini Fly! said:

Is it crazy to say Vaaks could be a version of 2022 Alfonso Plummer? Just from a shooting and scoring standpoint. People forget just how good that guy was.

Click to expand...

Every transfer we've had over the last 2 season has experienced a significant 3pt% dip in their first season after arriving.

Ben Humrichous - 41% down to 34%
Kylan Bowell - 38% down to 24%
Andrej Stojakovic - 32% down to 24%
Zvonimir Ivisic - 38% down to 28%
Even Jake Davis was 39% at Mercer and down to 34% his first year at IL.

*The lone exception to this is Tre White

Seeing Vaaks at 35% while shooting 8.4 threes per game, I absolutely thought 37% on 6.5 would be a realistic yardstick. Better percentage = less volume and better shots.

But our recent history makes me temper expectations.

PleaseMakeABowlGame said:

Every transfer we've had over the last 2 season has experienced a significant 3pt% dip in their first season after arriving.

Ben Humrichous - 41% down to 34%
Kylan Bowell - 38% down to 24%
Andrej Stojakovic - 32% down to 24%
Zvonimir Ivisic - 38% down to 28%
Even Jake Davis was 39% at Mercer and down to 34% his first year at IL.

*The lone exception to this is Tre White

Seeing Vaaks at 35% while shooting 8.4 threes per game, I absolutely thought 37% on 6.5 would be a realistic yardstick. Better percentage = less volume and better shots.

But our recent history makes me temper expectations.

Click to expand...

Here's the full sample:



I think with nearly half of them jumping up a level in play, might explain the 1% difference overall

Editing to add that actually none of these are reliable samples. Its widely accepted that 750 attempts is where a 3 point shooter's percentage fully stabilizes statistically.

NarrowJ said:

Here's the full sample:

View attachment 50478

I think with nearly half of them jumping up a level in play, might explain the 1% difference overall

Editing to add that actually none of these are reliable samples. Its widely accepted that 750 attempts is where a 3 point shooter's percentage fully stabilizes statistically.

Click to expand...

I said over the last two seasons - their first year at Illinois compared to their last year at said school they transferred from.

There's undeniably been a major dip for a lot of guys.

PleaseMakeABowlGame said:

I said over the last two seasons - their first year at Illinois compared to their last year at said school they transferred from.

There's undeniably been a major dip for a lot of guys.

Click to expand...

2 seasons is a completely worthless sample size

6 seasons (what I posted), while far better than 2 seasons, is still not 100% usable given that none of the players had reached 750 attempts

The thing that's undeniable is that none of this really tells us a whole lot due to the size of the dataset

NarrowJ said:

2 seasons is a completely worthless sample size

5-6 seasons is nearly worthless as well anyway given that none of the players had reached 750 attempts

The thing that's undeniable is that none of this really tells us much

Click to expand...

Not sure why they'd need to hit 750 attempts. We're measuring production; not someone's 100% absolute shooting prowess.

Fly Illini Fly! said:

Not sure why they'd need to hit 750 attempts. We're measuring production; not someone's 100% absolute shooting prowess.

Click to expand...

High variance + binomial nature of shooting. There's only two outcomes: make or miss. In stats, small samples of binary events suffer from massive swings. Its volatile not only due to the small number of attempts but also the small number of outcomes for each attempt.

Here's an example of how a real world situation can influence a player's shooting data at a small sample size:

A player's raw percentage over, lets say, 100 attempts can be artificially inflated by getting mostly wide-open uncontested looks in that player's most efficient shooting spot on the floor due to having a great playmaker. A player's raw percentage can also be artificially deflated by being placed in a role where they need to manufacture 3 point looks in end-of-clock situations.

There's just way more context than "there's a curse" -- that's frankly a weird thing to think (unless you actually believe in curses, I dunno? I don't believe that's a real, actual thing that affects college basketball players).

NarrowJ said:

High variance + binomial nature of shooting. There's only two outcomes: make or miss. In stats, small samples of binary events suffer from massive swings. Its volatile not only due to the small number of attempts but also the small number of outcomes for each attempt.

Here's an example of how a real world situation can influence a player's shooting data at a small sample size:

A player's raw percentage over, lets say, 100 attempts can be artificially inflated by getting mostly wide-open uncontested looks in that player's most efficient shooting spot on the floor due to having a great playmaker. A player's raw percentage can also be artificially deflated by being placed in a role where they need to manufacture 3 point looks in end-of-clock situations.

There's just way more context than "there's a curse" -- that's frankly a weird thing to think (unless you actually believe in curses, I dunno? I don't believe that's a real, actual thing that affects college basketball players).

Click to expand...

You make good points in the time/length of the sample size. However, that still doesn’t clearly define anything more than a smaller sample size.

The findings of a longer sample size could be flawed more than small one. There is no clear or definitive qualifying measure other than being behind the 3 point line. For example both sample sizes of multiple seasons or just a couple seasons include early season or late season shots, wide open shots, hand in the face shots, transition 3s, set 3s, catch-and-shoot 3s, half court buzzer beaters, etc.

That’s a bunch of mumbo jumbo, but I say that to ask does a high sample size tell more than a smaller sample size in terms of recency in this instance?

NarrowJ said:

High variance + binomial nature of shooting. There's only two outcomes: make or miss. In stats, small samples of binary events suffer from massive swings. Its volatile not only due to the small number of attempts but also the small number of outcomes for each attempt.

Here's an example of how a real world situation can influence a player's shooting data at a small sample size:

A player's raw percentage over, lets say, 100 attempts can be artificially inflated by getting mostly wide-open uncontested looks in that player's most efficient shooting spot on the floor due to having a great playmaker. A player's raw percentage can also be artificially deflated by being placed in a role where they need to manufacture 3 point looks in end-of-clock situations.

There's just way more context than "there's a curse" -- that's frankly a weird thing to think (unless you actually believe in curses, I dunno? I don't believe that's a real, actual thing that affects college basketball players).

Click to expand...

Well, again, you're looking to measure the player's absolute shooting ability (in experimental manners).

But measuring someone's absolute ability has zero to do with helping our team win games; it production and efficiency in that particular setting that matters.

Take Vaaks for example: he will likely shoot a higher percentage 1) if he decreases a bit on the 8.4 3ptA per game and 2) due to our large variety of sufficient playmakers.

Does that skew his outcome away from his experimental, absolute percentage? Yes. But it doesn't matter because he's being more efficient and productive, which is all that matters.

.....

Also, if you'd like to experiment the data, truth is 750 still isn't enough.

Law of Large Numbers is infinite: the more you repeat the experiment, the closer and closer your results will be to the true, expected average.

PleaseMakeABowlGame said:

Every transfer we've had over the last 2 season has experienced a significant 3pt% dip in their first season after arriving.

Ben Humrichous - 41% down to 34%
Kylan Bowell - 38% down to 24%
Andrej Stojakovic - 32% down to 24%
Zvonimir Ivisic - 38% down to 28%
Even Jake Davis was 39% at Mercer and down to 34% his first year at IL.

*The lone exception to this is Tre White

Seeing Vaaks at 35% while shooting 8.4 threes per game, I absolutely thought 37% on 6.5 would be a realistic yardstick. Better percentage = less volume and better shots.

But our recent history makes me temper expectations.

Click to expand...

Sample size can also be a confounding variable. A guy shoots a lot more or less in a different context is going to make it more difficult to compare apples to apples...

Attempts % change
PREV ILL
BH 128 181 141%
KB 174 143 82%
AS 129 82 63%
ZI 125 100 80%
JD 155 64 41%

All of these guys had a significant change in the number of attempts, Only Ben shot them more frequently. Variance in sample size is going to give you variance in results. Doesn't mean there is something in the water making guys shoot worse here. Andre, Z, and especially Davis all accepted a smaller share of the minutes and the offensive load here.

Vaaks will have a similar role and minutes. Usage could be lower. I doubt he takes 8 3's a game again, but he won't have a problem maintaining 30 mpg. I'd bet my beach house he will get more open looks here, as the Providence offense was a 2 man show. He had to take a lot of contested shots.

HouTexIllini said:

You make good points in the time/length of the sample size. However, that still doesn’t clearly define anything more than a smaller sample size.

The findings of a longer sample size could be flawed more than small one. There is no clear or definitive qualifying measure other than being behind the 3 point line. For example both sample sizes of multiple seasons or just a couple seasons include early season or late season shots, wide open shots, hand in the face shots, transition 3s, set 3s, catch-and-shoot 3s, half court buzzer beaters, etc.

That’s a bunch of mumbo jumbo, but I say that to ask does a high sample size tell more than a smaller sample size in terms of recency in this instance?

Click to expand...

A large sample size doesn't tell you more about recency, its tells you more about baseline skill. Small samples have too much variance, especially for events with binary outcomes.

If we set aside all of the other working ancillary reasons like being on a team with a playmaker who gets you ball in your spot when you're wide open, or being 'the guy' on a team where you have to hoist up 28 footers when the shot clock dwindles, we are left with the main reason small samples aren't trusted -- variance.

On 100 attempts, if 3 of them bounce around and go in instead of being a miss, it makes a 35% shooter appear to be a 41% shooter and then you end up with folks wondering why Ben Humrichous didn't come through on his .414 percentage that he posted his single year at Evansville.

Fly Illini Fly! said:

Well, again, you're looking to measure the player's absolute shooting ability (in experimental manners).

But measuring someone's absolute ability has zero to do with helping our team win games; it production and efficiency in that particular setting that matters.

Take Vaaks for example: he will likely shoot a higher percentage 1) if he decreases a bit on the 8.4 3ptA per game and 2) due to our large variety of sufficient playmakers.

Does that skew his outcome away from his experimental absolute percentage? Yes. But it doesn't matter because he's more efficient and productive, which helps us win more games.

Click to expand...

We're not going to agree on this, but that's okay. Because I do actually care more about a player's actual shooting ability than the percentage they posted being a spot up corner 3 point shooter at BFE University on 80-100 attempts.

I am curious about sample sizes. Can you provide a link to where I can read more about the widely accepted sample of 750?

nmillini said:

I am curious about sample sizes. Can you provide a link to where I can read more about the widely accepted sample of 750?

Click to expand...

Sure can. The 750 attempt threshold was discovered by Darryl Blackport of Basketball Prospectus.

Here are some links I found via a simple Google search on the subject (the first references Darryl Blackport's work):




Interestingly, free throw shooting percentages stabilize much faster at 250 attempts.

PleaseMakeABowlGame said:

Every transfer we've had over the last 2 season has experienced a significant 3pt% dip in their first season after arriving.

Ben Humrichous - 41% down to 34%
Kylan Bowell - 38% down to 24%
Andrej Stojakovic - 32% down to 24%
Zvonimir Ivisic - 38% down to 28%
Even Jake Davis was 39% at Mercer and down to 34% his first year at IL.

*The lone exception to this is Tre White

Seeing Vaaks at 35% while shooting 8.4 threes per game, I absolutely thought 37% on 6.5 would be a realistic yardstick. Better percentage = less volume and better shots.

But our recent history makes me temper expectations.

Click to expand...

"Every transfer...the lone exception." So not every transfer then?

Here's the thing, and why sample size matters if you're going to try to use the results of a sample of players as a predictor for a completely different player. A college basketball season is short, and you've got guys on here who are taking in some instances less than 3 attemps a game. Hit one rough patch and it can completely tank your percentage. That doesn't mean that future player is going to have the same thing happen to him.

Big Z for example. Through game 23 of the season, his 3pt% was 36.2%. Over the next 14 games it was 9.7%. No I did not forget a digit, it was under 10%.

So basically for over 60% of the season Big Z was shooting just as well as he did before he got here. Then in the latter half of the season he hit an extended slump that absolutely tanked his 3pt%.

Let's take a look at Stoj now, not here but at Cal. On the regular season his 3pt% was just 29%. How'd he get it up to 32%? In the two conference tourney games Cal played, he shot 7/14 from 3 (58%) which was good enough to raise his entire average for the season to 31.8%.

Stoj also took a lot fewer 3s here than he did at Stanford and Cal. A lot. And a disproportionate amount were in the beginning of the season. I think he got off to a rough start and then stopped taking them. For the last 15 games of the season he was averaging 1.7 attempts a game. Hard to break out of a slump if you don't shoot, though I am personally glad he didn't try to shoot himself out of it. Would have probably been better for his 3pt% but worse for the team. He deferred the possibility that a two game (or more) game heater would lift his average, because that's not what the team needed from him.

The bigger problem even than sample size is correlation/causation. What is your hypothesis as to why players are shooting worse here, only over the last two seasons? What changed at that point? And how does the fact that Big Z shot fine until more than halfway throught the season fit into that hypothesis.

My guess is 2 seasons was not chosen for any reason other than that it fit the narrative you wanted to advance, which is that transfers coming in invariably get worse at shooting 3s (except for the one that didn't) and therefore we should be skeptical that Vaaks will do well. Maybe this is also why we should transfer in a 9th man who shot 55% from 3 last season so that they can come here and shoot a respectable 35% from 3 in at least 28 games, of which at least 18 will be clear the bench type situations.

I'd like to see Vaaks match Keaton's 2.4 made threes per game. I don't think he'll shoot Keaton's percentage, but he'll likely take more attempts.

NarrowJ said:

Here's the full sample:

View attachment 50478

I think with nearly half of them jumping up a level in play, might explain the 1% difference overall

Editing to add that actually none of these are reliable samples. Its widely accepted that 750 attempts is where a 3 point shooter's percentage fully stabilizes statistically.

Click to expand...

I completely agree on your point about sample size, but I disagree that it is applied per-player. On the whole, all our transfers shot 666-1856 (35.9%) in their year prior to Illinois and 534-1630 (32.8%) in their first year at Illinois, so the aggregated sample sizes are sufficient (and the decreased % wasn't obviously due to increased volume).

Also, while that doesn't confirm a change in their "baseline" skill since the results are highly dependent on the context of the two seasons for each player (especially shot difficulty), we don't ultimately care about players' baseline skill. We care about how they shoot in our system, however that my inflate or deflate their %.

But we still can't apply this trend with certainty to player X, since any individual player had a very low sample size in 3pt attempts last year, their context may change in dramatically different ways than the aggregate of our prior transfers, etc.

wc_illini said:

I completely agree on your point about sample size, but I disagree that it is applied per-player. On the whole, all our transfers shot 666-1856 (35.9%) in their year prior to Illinois and 534-1630 (32.8%) in their first year at Illinois, so the aggregated sample sizes are sufficient (and the decreased % wasn't obviously due to increased volume).

Also, while that doesn't confirm a change in their "baseline" skill since the results are highly dependent on the context of the two seasons for each player (especially shot difficulty), we don't ultimately care about players' baseline skill. We care about how they shoot in our system, however that my inflate or deflate their %.

But we still can't apply this trend with certainty to player X, since any individual player had a very low sample size in 3pt attempts last year, their context may change in dramatically different ways than the aggregate of our prior transfers, etc.

Click to expand...

If someone forced me to read into the differences:

Grandison: was focal point of defenses at Holy Cross as their leading scorer, at Illinois a catch-and-shoot 3 & D player
Guerrier: likely just variance, role was similar at Illinois as it was at Syracuse & Oregon
Domask: early season shooting slump + jump up in competition level
Boswell: no clue here, he shot 38% at Arizona... I know they used him differently
Humrichous: big variance (ie, not really a 41% guy just had good variance at Evansville) + jump up in comp
White: increased 4.3%, probably variance
Davis: jump in comp (though he shot significantly higher % this year than he did at Mercer)
Stojakovic: variance + sample size (again, all these players suffer from it to some degree)
Ivisic: late season slump tanked his %

But I can't bring myself to use "Illinois 3 point shooting is cursed" as a feasible explanation for any of it, especially considering the overall difference is negligible and likely explained by nearly half of the players moving up from low/mid-major to B1G.

nmillini said:

I am curious about sample sizes. Can you provide a link to where I can read more about the widely accepted sample of 750?

Click to expand...

The Illini have only had 4 players reach 750 3 point attempts in their career (Frazier, Bradford, Brown, Richardson,) so I am unconvinced that it is a usable bar to clear.

NarrowJ said:

If someone forced me to read into the differences:

Grandison: was focal point of defenses at Holy Cross as their leading scorer, at Illinois a catch-and-shoot 3 & D player
Guerrier: likely just variance, role was similar at Illinois as it was at Syracuse & Oregon
Domask: early season shooting slump + jump up in competition level
Boswell: no clue here, he shot 38% at Arizona... I know they used him differently
Humrichous: big variance (ie, not really a 41% guy just had good variance at Evansville) + jump up in comp
White: increased 4.3%, probably variance
Davis: jump in comp (though he shot significantly higher % this year than he did at Mercer)
Stojakovic: variance + sample size (again, all these players suffer from it to some degree)
Ivisic: late season slump tanked his %

But I can't bring myself to use "Illinois 3 point shooting is cursed" as a feasible explanation for any of it, especially considering the overall difference is negligible and likely explained by nearly half of the players moving up from low/mid-major to B1G.

Click to expand...

If I take out the five who moved up (Humrichous, Davis, Domask, Harmon, and Grandison), the others were 403-1156 (34.9%) the year before, and 370-1134 (32.6%) their first year here. Their roles may still have been different.

The five who did move up were 236-700 (37.6%) the year before, and 164-496 (33.1%) the year after, but those sample sizes are on the small side.

wc_illini said:

I completely agree on your point about sample size, but I disagree that it is applied per-player. On the whole, all our transfers shot 666-1856 (35.9%) in their year prior to Illinois and 534-1630 (32.8%) in their first year at Illinois, so the aggregated sample sizes are sufficient (and the decreased % wasn't obviously due to increased volume).

Also, while that doesn't confirm a change in their "baseline" skill since the results are highly dependent on the context of the two seasons for each player (especially shot difficulty), we don't ultimately care about players' baseline skill. We care about how they shoot in our system, however that my inflate or deflate their %.

But we still can't apply this trend with certainty to player X, since any individual player had a very low sample size in 3pt attempts last year, their context may change in dramatically different ways than the aggregate of our prior transfers, etc.

Click to expand...

I agree with this, but I think the problem is when using it as a predictive measure. If you can't come up with some reasonable (and univerally applicable) explanation for why a decreased 3pt% exists in the aggregate, I do not think it has any predictive value when it comes to future transfers. Otherwise, it's just a case of some players' average goes up, some players' average goes down, and in this small sample size we've just happened to have more go down than up so the numbers are down in aggregate.

the juiceman cometh said:

I agree with this, but I think the problem is when using it as a predictive measure. If you can't come up with some reasonable (and univerally applicable) explanation for why a decreased 3pt% exists in the aggregate, I do not think it has any predictive value when it comes to future transfers. Otherwise, it's just a case of some players' average goes up, some players' average goes down, and in this small sample size we've just happened to have more go down than up so the numbers are down in aggregate.

Click to expand...

My hair brained conclusion is we have some players moving up in competition level + first year in new system

Fly Illini Fly! said:

I'd like to see Vaaks match Keaton's 2.4 made threes per game. I don't think he'll shoot Keaton's percentage, but he'll likely take more attempts.

Click to expand...

Keaton made 2.4 and Shannon made 2.4

Vaaks at Providence made 2.9, but his attempts will certainly come down.

As long as he stays the same shooter, seems attainable.

the juiceman cometh said:

I agree with this, but I think the problem is when using it as a predictive measure. If you can't come up with some reasonable (and univerally applicable) explanation for why a decreased 3pt% exists in the aggregate, I do not think it has any predictive value when it comes to future transfers. Otherwise, it's just a case of some players' average goes up, some players' average goes down, and in this small sample size we've just happened to have more go down than up so the numbers are down in aggregate.

Click to expand...

None at all? I agree that it can't be applied to any individual future transfer, but ~1600 attempts seems meaningful for future transfers in aggregate.

Year-by-year comparisons are obviously much more prone to sample size and context issues, but they are interesting to compare with Illinois' team 3pt attempt rates:
Season | Transfers' aggregate 3pt% change from prior year | Team 3pt attempt rate
2020-21 | +3.6% | 30.0
2022-23 | -1.0% | 41.9
2023-24 | -1.8% | 38.6
2024-25 | -2.4% | 47.2
2025-26 | -3.6% | 49.7

3pt attempt rate does not necessarily correlate with shot quality since a well-designed offense with lots of skilled players can generate lots of good looks. But it could imply a change in style for incoming transfers that requires some adjustment. Just a thought.

Illini 993 said:

The Illini have only had 4 players reach 750 3 point attempts in their career (Frazier, Bradford, Brown, Richardson,) so I am unconvinced that it is a usable bar to clear.

Click to expand...

So, we've never had a bad shooter at Illinois. We've only had good shooters and shooters who did not shoot enough.