Luke Kennard, 3pt Shooting, FT Shooting and some stats.

[dyedwab]

Figured this would be a good place to continue a conversation about Luke Kennard and his 3pt shooting and whether he is a "good" 3pt shooter or not and what his incredible FT shooting percentage means with respect to his likelihood of being an elite 3 point shooter.

Very quick and dirty look at some relevant stats.

Coming into this season, 7 of the Top 10 FT shooters by percentage in Duke history also played during the 3 pt shooting era. Here they are, along with there career 3pt shooting percentage.

Career FT %Rank 3pt Shooting %
1. J.J. Redick .406
2. Trajan Langdon .426
3. Jon Scheyer, .381
4. Quinn Cook .375
6. Seth Curry .420
9. Christian Laettner .485
10. Ryan Kelly .379

So, what we think is true intuitively seems to be provable, which is that good FT shooter tend to be high level 3pt shooters. Also notable, is that it's not necessarily a straight line. Ryan Kelly (.263) and Quinn Cook (.250) were not good 3pt shooter their freshmen years. And Laettner shot a grand total of 13 3pters in his first two seasons.

I think what this says is that Luke Kennard should be a better 3pt shooter than his current percentages show, but that it's not necessarily an instant correlation.

[uh_no]

Figured this would be a good place to continue a conversation about Luke Kennard and his 3pt shooting and whether he is a "good" 3pt shooter or not and what his incredible FT shooting percentage means with respect to his likelihood of being an elite 3 point shooter.

Very quick and dirty look at some relevant stats.

Coming into this season, 7 of the Top 10 FT shooters by percentage in Duke history also played during the 3 pt shooting era. Here they are, along with there career 3pt shooting percentage.

Career FT %Rank 3pt Shooting %
1. J.J. Redick .406
2. Trajan Langdon .426
3. Jon Scheyer, .381
4. Quinn Cook .375
6. Seth Curry .420
9. Christian Laettner .485
10. Ryan Kelly .379

So, what we think is true intuitively seems to be provable, which is that good FT shooter tend to be high level 3pt shooters. Also notable, is that it's not necessarily a straight line. Ryan Kelly (.263) and Quinn Cook (.250) were not good 3pt shooter their freshmen years. And Laettner shot a grand total of 13 3pters in his first two seasons.

I think what this says is that Luke Kennard should be a better 3pt shooter than his current percentages show, but that it's not necessarily an instant correlation.

kp talks about it. among the entire country ft sitting is the best predictor of 3pt shooting aside from last year's 3pt shooting

[CDu]

This came up in another thread, but Kennard shot 2-18 to start his career. Since then, he has shot 24-62 (38.7%). I will not be at all surprised if he ends this season with a 3pt% of 37% or higher.

[kAzE]

I've been on the Kennard train since game 1, I think he's already our best all-around offensive player. Looking forward to seeing what he can do over the next several games until Amile returns.

[Kedsy]

This came up in another thread, but Kennard shot 2-18 to start his career. Since then, he has shot 24-62 (38.7%). I will not be at all surprised if he ends this season with a 3pt% of 37% or higher.

After your lecture in the other thread about small sample sizes, I'm surprised to hear this line from you.

Luke's 3-pt shooting cannot be fairly described as "bad at the beginning, then good." He's had three good distance games (12 for 19 in the three games) and in his other 15 games he shot a collective 23% (14 for 61). The three good games were not particularly close to each other (they came in games 7, 12, and 18), and two of them were against Utah State and Elon.

It's possible the Notre Dame breakout game will put him on a hot streak from downtown -- and I have no doubt that in future seasons his three-point shooting will come around and mirror his excellent FT shooting -- but at the moment I don't see any strong reasons to believe, for this season at least, that he's really a 37+% shooter in a 32.5% package.

[westwall]

After your lecture in the other thread about small sample sizes, I'm surprised to hear this line from you.

Luke's 3-pt shooting cannot be fairly described as "bad at the beginning, then good." He's had three good distance games (12 for 19 in the three games) and in his other 15 games he shot a collective 23% (14 for 61). The three good games were not particularly close to each other (they came in games 7, 12, and 18), and two of them were against Utah State and Elon.

It's possible the Notre Dame breakout game will put him on a hot streak from downtown -- and I have no doubt that in future seasons his three-point shooting will come around and mirror his excellent FT shooting -- but at the moment I don't see any strong reasons to believe, for this season at least, that he's really a 37+% shooter in a 32.5% package.

Sigh! Statistics may be useful, but in this case they deny what we see with our eyes.

[uh_no]

He's had three good distance games

what the heck is your "good" game metric? He's had 6 games of > 40%

If 40+% doesn't fit your definition of good three point shooting, then I understand why you're so hard to please in this case.

Grayson has had about as many poor long range shooting nights as luke recently, yet I don't hear anyone calling for him to stop shooting. he only has 3 games >40% since the start of december to luke's 5...and nearly as many clunkers.

Ultimately I see your point. either side can morph the data to their side...but the ancillary evidence (the FTs, the history, the green light) does not support your case.

Ultimately we have the same problem that any evaluation of a largely random process has. Is the result we have seen due to a statistical aberration, or is our model wrong? I hold the belief, and have since the very beginning that luke was simply in that streak of tails...and has since regressed towards the mean. Before he got "hot" i had a post detailing how every shooter of near 40% is likely to have a streak as bad as luke's start to the season at least once in their 4 year college career.

so neither of us will be able to prove our case by simply citing irrelevent breakdowns of the three point data. We will view it in a way that supports our beliefs. Unfortunately, to successfully argue your position, you must also explain:
1) why is luke one of the best FT shooters in the country, and yet can't hit threes at a good clip, despite the incredibly strong correlation between the two?
2) how did luke go from being a great high school shooter, winning the national three point contest even, to being a mediocre shooter? Speed of the game? could be, IDK...but for your thesis to be plausible, there needs to be an explanation for this evidence
3) why does he have the green light? Is coach K simply wrong about him? does he not have any other options?

You may have some reasons or hypotheses for any or all of those. but in the world of bayesian probabilities, and given my observations, there is a massively higher confidence that luke is an innately good three point shooter that had a poor streak than anything else.

Kenpom lists the following stats in order of predicting three point shooting percentage in the long run, derived from actual data on a large scal (and incidently deals with the exact same situation we're dealing with here...) http://kenpom.com/blog/index.php/web...int_percentage

1. 3-point percentage
1. 2-point percentage
3. FT percentage
4. 3-point attempt rate
5. Turnovers

Luke was known as a shooter coming out of high school and won the HS three point contest.
Luke has the best 2 point % among guards on the team 57% to grayson's 53%
Luke has the best FT% on the team, and 8th best FT % in the country
Luke has the highest 3 point attempt per minute of the team
luke has the best TO % on the team and is ranked 81st in the country

So in all five indicators, which were empirically determined to be the best way to predict three point shooting in the long run, luke is the best on the team. Is Luke the aberration? maybe he is, maybe he isn't.

But right now, I have 6 strong pieces of evidence in my favor (the 5 best indicators, and a green light from the best coach in history), and you have a a slump in the first few games of a college career. I think even you have to agree, my priors are far stronger. Maybe you've hit on that 1% when all of them are true, and the guy ends up not being a good shooter. But I'll take my chances with the 99%.

[Wander]

Stephen Curry, the best 3 point shooter in college basketball history, shot under 40% his last year at Davidson. JJ Redick, the second best shooter in college basketball history, only led Duke in 3 point % one out of four seasons.

Duke tends to have lots of good 3 point shooters, which I think has skewed our fanbase's opinion a little bit about what constitutes a good 3 point shooter. Kennard's number of 3 point shots made would rank second on UNC or Kansas, who will both be ranked in the top 3 tomorrow. He's a good shooter.

[CDu]

After your lecture in the other thread about small sample sizes, I'm surprised to hear this line from you.

Luke's 3-pt shooting cannot be fairly described as "bad at the beginning, then good." He's had three good distance games (12 for 19 in the three games) and in his other 15 games he shot a collective 23% (14 for 61). The three good games were not particularly close to each other (they came in games 7, 12, and 18), and two of them were against Utah State and Elon.

It's possible the Notre Dame breakout game will put him on a hot streak from downtown -- and I have no doubt that in future seasons his three-point shooting will come around and mirror his excellent FT shooting -- but at the moment I don't see any strong reasons to believe, for this season at least, that he's really a 37+% shooter in a 32.5% package.

No, he has had three phenomenal games (50% or more), 2 other good games (40%), and 3 other okay games(33%). He has had 2 low-attempt (<=2) games totalling 1-3. So he has a whopping 8 games of 18 on the not good side. My point was that 4 of them were clustered at the beginning, and that particular cluster was worse than any other bad game cluster since. Since then, he has been a good, though not consistent, shooter.

And again, he is literally 4 makes from having a good %, 6 makes from a very good %. Given the circumstantial evidence that has been provided previously, I think it is much more likely that he is a good shooter who has fallen prey to the sample size demons rather than he has somehow forgotten his shot and yet Coach K has let him keep firing.

Obviously it remains to be seen. He might not continue to rebound from the cold start. Yes, he is a streaky shooter, not a metronomic shooter. But it won't take some phenomenal hot streak for him to get to 37%. He just needs to be 4 shots better than 37% in his next 80 attempts to counteract the 4-shot dropoff. I am not saying he definitely will get there. Just that, given ALL the information (both production and contextual), it would not surprise me at all to see him get to 37% by season's end.

[CDu]

Kenpom lists the following stats in order of predicting three point shooting percentage in the long run, derived from actual data on a large scal (and incidently deals with the exact same situation we're dealing with here...) http://kenpom.com/blog/index.php/web...int_percentage

1. 3-point percentage
1. 2-point percentage
3. FT percentage
4. 3-point attempt rate
5. Turnovers

Luke was known as a shooter coming out of high school and won the HS three point contest.
Luke has the best 2 point % among guards on the team 57% to grayson's 53%
Luke has the best FT% on the team, and 8th best FT % in the country
Luke has the highest 3 point attempt per minute of the team
luke has the best TO % on the team and is ranked 81st in the country

So in all five indicators, which were empirically determined to be the best way to predict three point shooting in the long run, luke is the best on the team. Is Luke the aberration? maybe he is, maybe he isn't.

I obviously totally agree with your overall point. But it is probably fair to say that the 3-point percentage evidence is the best on the team. I'd say, though, that his historical data (when combined with his current-year data) say he should be a pretty good 3pt shooter though. And when you combine that with the other 4 indicators, the evidence strongly suggest that he is indeed a good 3pt shooter who happens to have had a bad stretch so far in his career.

And to Kedsy's comment about my use of small samples in the evidence, my point was not to say that the 2-18 start is definitely distinct from his more recent 24-62 stretch. My point was a part of a larger point that his sample is very small, and that a small stretch of bad shooting can sway overall results over such a short stretch. Had he merely gone 6-18 instead of 2-18 in that stretch, he's a 37.5% shooter on the season and nobody is questioning his shooting prowess. Are we really going to decide to throw out all of the supporting evidence that suggests he is a good 3pt shooter based on a difference of 4 shots?

If we had no prior information regarding Kennard's shooting ability, I'd be inclined to take a hypothesis that maybe he is a mediocre 3pt shooter. But we aren't lacking that information. We know he was a great shooter in high school. We know he is a great 2pt shooter and FT shooter. We know that Coach K has given him the green light, which suggests that he's hitting them in practice (otherwise, there is just no way Coach K is okay with him continuing to put up such a high volume of 3s). All of that information, in my opinion, outweighs the small sample of attempts he's had so far. I think it's FAR more likely that the 2-18 start is the anomaly, and that he really is the shooter that his past and Coach K's opinion suggests he is.

Or, for another perspective, let's say Kennard goes 11-20 in his next 4 games. He'd then be a 37% shooter for the season. Are you going to throw out your view that he's not a good 3pt shooter simply because of that hot stretch? That's the problem with basing everything on such a small sample of attempts at this point in his career.

In the same way that Battier probably wasn't a bad 3pt shooter as a freshman despite making only 4 of 24 attempts that year. He was a good shooter in high school and a good shooter his sophomore through senior years. So it's far more likely that he was a good shooter in his freshman year who just happened to miss 5-6 more attempts than he should have based on his abilities.

[huey]

Oooh- can I add another highly anecdotal, low-sample size variable to the debate?

What about the types of shots Luke's been hitting? Against Notre Dame, I swear he must have pulled up at least 30 feet from the basket. Who does he think he is, a Curry? With his lightning quick release, he also puts up shots with a defender pretty darn close. I'd say <50% of his shots are the Matt Jones, set and open, variety. Wish they had SportsVu (or whatever) camera data online for college like the NBA (and that I, well, knew how to manipulate it).

All I know is that he's put up some shots that would've gotten most Duke players, let alone freshman yanked immediately. I hope it's K's faith in his three point shooting.

But the more important question, is he ever going to be as good a three-point shooter as Lance Thomas? I'm not sure what his sample size said about that at Duke...

[uh_no]

But the more important question, is he ever going to be as good a three-point shooter as Lance Thomas? I'm not sure what his sample size said about that at Duke...

nobody can touch marshall, who I believe shot 100% over the past 2 years

[Indoor66]

nobody can touch marshall, who I believe shot 100% over the past 2 years

I think you sell him short. IRRC, he is 100% over 5 years.

[CDu]

I obviously totally agree with your overall point. But it is probably fair to say that the 3-point percentage evidence is the best on the team. I'd say, though, that his historical data (when combined with his current-year data) say he should be a pretty good 3pt shooter though. And when you combine that with the other 4 indicators, the evidence strongly suggest that he is indeed a good 3pt shooter who happens to have had a bad stretch so far in his career.

Typing fail on my part. I meant to say "not the best on the team."

[Wander]

Typing fail on my part. I meant to say "not the best on the team."

I would like to propose that total number of 3's made (or equivalently, 3's made per game) is the best metric of the "simple stats" to tell how good a 3 point shooter someone is.

[CDu]

I would like to propose that total number of 3's made (or equivalently, 3's made per game) is the best metric of the "simple stats" to tell how good a 3 point shooter someone is.

On a well-coached team, there is probably a good argument to be made for this as best proxy of the "simple stats". The underlying assumption being that a good coach will not let a bad (or even mediocre) shooter take a lot of 3s. And as such, the fact that a shooter takes enough shots to make a bunch of 3s (without getting benched or being told to stop shooting them) is a good indicator that said shooter is a good 3pt shooter.

I like your outside-the-box thinking.

[kAzE]

I just go by the eye test. Does the guy have a good form/arc on his jump shot? Does he get it off quickly? Does his shot look comfortable from 25 feet? Is he a good free throw shooter? In all cases, if it's a yes, he's probably a good shooter. All shooters have hot and cold streaks. Even half a season's worth of games is nowhere near a large enough sample size to pass final judgment on anyone's ability to shoot.

[gumbomoop]

Starting the year out, Luke had to adjust to playing with this team and finding his role, but now he's starting to get comfortable. I think his move into the starting lineup is permanent going forward, and he's probably going to be trusted with more and more minutes given the way he's played.

Kennard is one of the smartest offensive players in the league.

Although this thread focuses on Kennard's 3-pt shooting, I see no reason to post a new thread on a separate Luke-topic. I've imported a couple of previous comments from other threads.

I don't know whether Kennard will continue to start, and sure don't want Thornton to get discouraged, so I'll hope Derryck is getting plenty of encouragement. It's not good if several [all but Derryck?] of our perimeter players are consistently going 37-40 minutes. Although in a sense we're "deeper" on the perimeter than inside, it's not at all deep. But if Derryck can give us a steady 20 mpg, then the other 4 perimeter guys can stick around 35 each, giving all of them a breather in each half. And obviously we need Chase and Sean to provide a few breather minutes for Marshall, too.

On Kennard specifically, I do assume his 3-pt % will gradually rise into the respectable+ range, but from the beginning -- for me, starting with his announcement for Duke -- I was actually more impressed with his many other talents: handle, passing, vision, smarts, court sense, amphibious drive-finishing. I just didn't pay enough attention to his D in high school, so what I judge to be his very solid D adds meaningfully to his already-significant value. Ditto his willingness to be a vocal leader, even as a new guy. It looks as if, for example, he understands where guys are supposed to be on D, and speaks up.

So, never shy to speak up about Luke's talents, I'm inclined to claim, as a friendly amendment to fdd's comment, that Luke is at least among the smartest defensive players, too, or is likely to develop in that direction. He fights through screens some, moves his feet pretty well, gets in good rebounding position. He'll take a few charges by beating his man to the spot. He communicates.

I welcome the change in the way Luke is being used on O, and do wonder whether his being used earlier in the season mostly as a 3-pt "specialist" partially explains his getting off to a woeful start in that role. I think all posters will agree that he is way more than a 3-bomber, and that he makes plays for himself and for others. I rather expect the ball will be in his hands a little more as the season continues. Not that he will morph into our PG, for he isn't a classic PG, in the way that Derryck can become.

But, extending kAzE's comment, I think Luke is more comfortable because his role has changed noticeably since those early games. His minutes have gone up (1) because Amile went down, but also (2) because his overall play blossomed -- strikingly, at times -- the more he used all of his multiple skills.

[loran16]

People in this thread are making a whole lot out of very little.

First, yes, free throw shooting is a better predictor of future 3 point shooting than current 3 point shooting. This is also true on the next level - it's why many thought (correctly) that Justise Winslow's shooting last year wouldn't translate to the NBA (and it hasn't). And it's why many feel the same about Ingram. But it works in Kennard's favor.

Second - remember, it's not like we didn't know anything about Kennard coming in - we have PRIOR information - that he's a known HS Sharpshooter. This suggests even more that he's likely to improve his shooting #s from range since it's unlikely the talent scouts from before were very wrong.

Third - regression (which is what we are expecting from Kennard) being likely doesn't mean we'll see it in any particular game going forward. Odds are we do see him hit in the high 30s over the rest of the season. That said, he'll have good games and bad, and may have another cold streak. There's no reason to think it will come in the form of multiple games in a row, as Kedsy tried to argue it should.

Fourth- this hasn't mattered THAT much so far. Kennard currently 3rd on the team in possessions used takes the third highest percentage of shots while on the floor. He's being used a lot. And for guys being used around as often as he is or more (being used on 20% of a team's possessions), Kennard is TWELTH in offensive efficiency (Grayson is 7th actually). Kennard has surprisingly shot a very good % from 2 this year - his 57% rate is actually the highest on the team aside from Marshall and Amile! He doesn't turn the ball over despite using the ball so often. And he makes those free throws.

Kennard might be an elite player at Duke if the 3 comes - which it likely will. At the moment, he's REALLY damn good on offense already, and this whole debate is somewhat academic.

[Indoor66]

People in this thread are making a whole lot out of very little.

First, yes, free throw shooting is a better predictor of future 3 point shooting than current 3 point shooting. This is also true on the next level - it's why many thought (correctly) that Justise Winslow's shooting last year wouldn't translate to the NBA (and it hasn't). And it's why many feel the same about Ingram. But it works in Kennard's favor.

Second - remember, it's not like we didn't know anything about Kennard coming in - we have PRIOR information - that he's a known HS Sharpshooter. This suggests even more that he's likely to improve his shooting #s from range since it's unlikely the talent scouts from before were very wrong.

Third - regression (which is what we are expecting from Kennard) being likely doesn't mean we'll see it in any particular game going forward. Odds are we do see him hit in the high 30s over the rest of the season. That said, he'll have good games and bad, and may have another cold streak. There's no reason to think it will come in the form of multiple games in a row, as Kedsy tried to argue it should.

Fourth- this hasn't mattered THAT much so far. Kennard currently 3rd on the team in possessions used takes the third highest percentage of shots while on the floor. He's being used a lot. And for guys being used around as often as he is or more (being used on 20% of a team's possessions), Kennard is TWELTH in offensive efficiency (Grayson is 7th actually). Kennard has surprisingly shot a very good % from 2 this year - his 57% rate is actually the highest on the team aside from Marshall and Amile! He doesn't turn the ball over despite using the ball so often. And he makes those free throws.

Kennard might be an elite player at Duke if the 3 comes - which it likely will. At the moment, he's REALLY damn good on offense already, and this whole debate is somewhat academic.

...And on this board, that is shocking HOW, prey tell?