2014 Draft Projection Models

[Statman]

Washington is fighting hard for the playoffs and better seeding, and given Ariza's good shooting this season Porter is just pushed completely out of the rotation. He's averaging nine minutes a game, and it's hard for players to play well in such opportunities. He's also coming off an injury, and he's only 20 years old.

Any more thoughts on Doug McDermott? What's his best comparison? How much will he be able to score in the pros?

No NBA comps if you include steals/block rates - since his are literally lower than anyone that's ever made the NBA, by a long shot.

I think some team MIGHT try to get him pt for a while despite some serious defensive limitations (and O limitations in terms of not getting all those post points he did in college and being merely a short/poor rebounding stretch 4 or a slow/defensive liablity/medioce handles SF who only shoots 3's). He might stick for years as just a 3pt shooter who comes in to just do that when there's a O scrub on the other team in he can "cover".

But, who knows, maybe he'll be that ONE outlier who can have very below average st/bk rates (& mediocre A/TO rate to boot) in college and be a good NBA player. But, those are some real bad odds to buch for an NBA team to spend a 1st round pick on.

[Barncore]

Hey VJL, i noticed you have some guys in your spreadsheet that have decided to go back to college, such as Delon Wright and Brice Johnson. Just a heads up.

[JimiHendrix]

No NBA comps if you include steals/block rates - since his are literally lower than anyone that's ever made the NBA, by a long shot.

I think some team MIGHT try to get him pt for a while despite some serious defensive limitations (and O limitations in terms of not getting all those post points he did in college and being merely a short/poor rebounding stretch 4 or a slow/defensive liablity/medioce handles SF who only shoots 3's). He might stick for years as just a 3pt shooter who comes in to just do that when there's a O scrub on the other team in he can "cover".

But, who knows, maybe he'll be that ONE outlier who can have very below average st/bk rates (& mediocre A/TO rate to boot) in college and be a good NBA player. But, those are some real bad odds to buch for an NBA team to spend a 1st round pick on.

I know most draft models don't look good for McDermott, but sometimes you've got to use your eyes too. If McDermott is on the floor with athletes (which has never happened at Creighton), then his issues on defense will be overcome. Yes, individually he'll still be a problem, yes, he won't get steals or blocks, but he'll never miss a rotation, and he'll never be out of position. McDermott is a quality "team defender" which is something no box score stat can account for.

Offensively, well, of course he won't score 26 PPG in the NBA. He just doesn't have that kind of ability. But consider this: McDermott was the focus of Creighton's of the offense, which "padded" his stats, but it also worked to his disadvantage in that he was always the focus of the defense. They sought out to stop him, not his teammates, and he still scored at a high level. In the pros, he'll be expected to be a second or third option on the offensive end. This will open the floor for him, therefore giving him quality looks he otherwise never would've had in college. He'll be able to conserve energy considering his lower usage, which could lead to higher efficiency.

I think for McDermott, it's pretty realistic to expect 10-13 points per game out of him as a high end role player, but it's not out of the question that he could become an 18 PPG scorer at some point. He has too much talent and he's able to create space with the dribble. He's a knockdown three-point shooter, and he's skilled enough where he could create his own offense against the average NBA defender. Will he finish at the rim like he did in college? No, but I don't think it hurts his game enough to say he's a total bust.

I think what we're seeing now is a lot of people experiencing confirmation bias when it comes to Doug McDermott. "He's not athletic."..."He was featured in his offense."...The draft projection models hate him."....So he must be overrated...he must be a bust.

That isn't necessarily the case because sometime you must look beyond the box score, and with McDermott, I believe that to be the case.

[VJL]

"Hey VJL, i noticed you have some guys in your spreadsheet that have decided to go back to college, such as Delon Wright and Brice Johnson. Just a heads up. "

I'm going to post final numbers and start knocking guys off the list after the tourney. That should be the final update until the combine when I will add the new star probability outputs.

[nbo2]

I know most draft models don't look good for McDermott, but sometimes you've got to use your eyes too. If McDermott is on the floor with athletes (which has never happened at Creighton), then his issues on defense will be overcome. Yes, individually he'll still be a problem, yes, he won't get steals or blocks, but he'll never miss a rotation, and he'll never be out of position. McDermott is a quality "team defender" which is something no box score stat can account for.

His role on defense playing for his dad for four years was to do nothing except rest for offense...

More importantly, he's nowhere near athletic enough to guard NBA wings and nowhere near strong enough to guard NBA bigs and play inside offensively

[JimiHendrix]

I know most draft models don't look good for McDermott, but sometimes you've got to use your eyes too. If McDermott is on the floor with athletes (which has never happened at Creighton), then his issues on defense will be overcome. Yes, individually he'll still be a problem, yes, he won't get steals or blocks, but he'll never miss a rotation, and he'll never be out of position. McDermott is a quality "team defender" which is something no box score stat can account for.

His role on defense playing for his dad for four years was to do nothing except rest for offense...

More importantly, he's nowhere near athletic enough to guard NBA wings and nowhere near strong enough to guard NBA bigs and play inside offensively

I totally agree with that statement, but what I'm saying is this: McDermott's terrible defense has never been overshadowed by the rest of team. Everyone has been equally athletic or only slightly more. For the most part, there haven't been many great man-to-man defenders at Creighton and there certainly hasn't been an elite rim protector. You're right, McDermott is going to struggle defending many wings in the NBA. You're damn right.

But what I'm saying is he's a competitive and reliable off-ball defender. You can't say this about a lot of players. McDermott understands rotations, he knows the right spot to be at all times, and from what I can tell, he does a good job of calling out back screens and such. What I'm saying is statistics can't qualify that (yet), which is one of the limitations of using only box scores for assessing defense.

At Sloan, I think in 2013, Stan Van Gundy went on a rant about how J.J. Redick couldn't defend at all when he first came into the league, but he learned how to become a great team defender, which is why he won playing time near 30 minutes per game. Could he defend some of the great wings in the NBA man-to-man? No. But he learned how to play terrific off-ball defense by always beating his man to a spot or by helping properly on pick-and-rolls, and then closing out on his man if the pass went that way. If I remember correctly, SVG said something along the lines of, "Redick would never be on an All-NBA defensive first team, but he was a guy I could rely on to never make mistakes when he was out there."

I think Doug McDermott can be that exact type of player in time: a complimentary offensive threat and a reliable team defender.

[VJL]

JimiHendrix:

All of the things you note are perfectly reasonable, but the point remains that there really is not a precedent for a player with McDermott's statistical profile succeeding in the NBA. Steals and blocks are the main culprit, but his poor distribution statistics add to the concern as well. We are talking about a guy who collected 34 steals and 8 blocks across a 4-year and 145 game NCAA career.

You note how J.J. Reddick was able to overcome certain defensive shortcomings and contribute as a team defender, but Redick collected steals at more than 5X McDermott's rate in college and had just as many blocks. Even Ryan Anderson, who is probably the least prolific player in terms of NCAA defensive stats to turn into a great pro, doubled McDermott's steal rate and collected 7X as many blocks. Steve Novak, arguably the most extreme offensive specialist in the NBA more than doubled McDermott's steal rate and doubled his block rate.

I'm not even saying McDermott has no chance of succeeding, but the challenge he faces in rising above his box-score defensive statistics is on another level compared to any player coming before him.

[Statman]

JimiHendrix:

All of the things you note are perfectly reasonable, but the point remains that there really is not a precedent for a player with McDermott's statistical profile succeeding in the NBA. Steals and blocks are the main culprit, but his poor distribution statistics add to the concern as well. We are talking about a guy who collected 34 steals and 8 blocks across a 4-year and 145 game NCAA career.

You note how J.J. Reddick was able to overcome certain defensive shortcomings and contribute as a team defender, but Redick collected steals at more than 5X McDermott's rate in college and had just as many blocks. Even Ryan Anderson, who is probably the least prolific player in terms of NCAA defensive stats to turn into a great pro, doubled McDermott's steal rate and collected 7X as many blocks. Steve Novak, arguably the most extreme offensive specialist in the NBA more than doubled McDermott's steal rate and doubled his block rate.

I'm not even saying McDermott has no chance of succeeding, but the challenge he faces in rising above his box-score defensive statistics is on another level compared to any player coming before him.

Agreed.

BTW VJL - you are the one I gave the heads up on w/ Rondae Hollis-Jefferson a couple months ago - right? I notice he's top 30 in EPW for you at the moment. Not bad for a 6th man frosh that was on no one's radar until very recently.

I'm very curious where my model will have small school guys like Payton & Hunter compared to yours. Cleaning 18 seasons of errors in college player stats is a time consuming pain in the ass - it'll be worth it when done.

Do you eventually use three years of data if available (weighted I assume?) when done w/ your projections before the draft? I'm thinking three years with mine if the data is available (juniors & seniors). I'll probably do the same with the NBA career curves - so injury guy projections like Rose don't suck after 2 injury years. Three bad injury years, well, their career curve projection should probably suck at that point.

[Statman]

All of the things you note are perfectly reasonable, but the point remains that there really is not a precedent for a player with McDermott's statistical profile succeeding in the NBA. Steals and blocks are the main culprit, but his poor distribution statistics add to the concern as well. We are talking about a guy who collected 34 steals and 8 blocks across a 4-year and 145 game NCAA career.

Also, I just went on a little rant on another board about Frank Kaminsky, and how he might end up my #1 rated college player by the end of this season (he's 5 at the moment I think). NOT draft model - just ranking players by how well they performed in college all things considered. McDermott was 2nd to KJ McDaniels (179 to 180, Kaminsky 9th at 169 before the two HUGE games) my last update on 3/25. http://hoopsnerd.com/?p=322

Anyway - some just can't wrap their brain around a 13/6 guy being ranked #1 NATIONALLY in performance. I kinda love that Kaminsky has been top 10 in my rankings all year (trust the metrics) - because he has passed the eye test with flying colors every time I've watched him. Years ago my ratings had 13 & 6 Roy Hibbert top 10 nationally, I felt so much better about that when he became a good NBA player.

But, can you imagine if you traded Kaminsky for McDermott - and had their respective new coaches game plan them in? I'm guessing Kaminsky would easily be a 20/10 guy w/ a bunch of blocks and better a/to rate than McDermott if he played at Creighton (he'd play less mpg than Doug - fouls, cardio). McDermott would probably be a 17 & 5 guy at Wisconsin with crazy nice shooting efficiency, and an uptick in assist/steal rate (steal rate no longer historically low, just merely bad). I kinda feel Creighton would maybe be just as good as they were in real life. I'm guessing Wisconsin would surely be worse.

If we tried to imagine Kaminsky being a Euro with his ball skills - wouldn't he be an easy top 10 pick? He's a young junior, turning 21 in a few days.

As I'm typing this - had the urge to pull up my "old" rankings for last season, curious about Kaminsky.

http://classic.hoopsnerd.com/uploads/20 ... sFINAL.pdf

Playing less minutes hurts the rating btw, how much depending on quality of team/teammates and how low the minutes are (low minutes on a BAD team hurt the rating alot - on a good team not nearly as much because sometimes it's hard to get playing time w/ a bunch of good teammates). Using HnI rank on the right of the PDF - ignores missed games (better predictor):

Highest ranked player '12-13 under 20 mpg, Mitch McGary (94th w/ a 146 at 19.7 mpg).
Highest ranked player '12-13 under 19 mpg, Casey Prather (98th w/ a 145 at 17.1 mpg).
Highest ranked player '12-13 under 17 mpg, Montrezl Harrell (136th w/ a 141 at 16.2 mpg).
Highest ranked player '12-13 under 16 mpg, Frank Kaminsky (160th w/ a 138 at 10.3 mpg).

So, my rankings had Frank Kaminsky as the 160th best player in the nation in '12-13, despite only playing 10.3 mpg. He was a stud just waiting for the playing time to prove it.

Man, I guess I can't complain about those results, in hindsight.

[VJL]

statman:

regarding the # of years, are you talking about the IVs or the DV? My DV is just the single best season a player ever has. I found that with college data the most recent and the previous year carry almost equal amounts of information, but all years before that only offer a very marginal improvement in prediction.

[miabbett]

First of all, I really like the work, the system you've built up is very impressive. Second, I was wondering if anyone here had any thoughts about players coming from different levels of competition and how they might project differently. For example, is there anything in your system that could identify and adjust accordingly for a guy like Elfrid Payton who put up great numbers in a weak conference?

[VJL]

miabett:

Team strength of schedule and margin of victory are both included in the model and have a powerful effect on predictions. That said, there are nuances to the relationship between level of competition and expected NBA production that the model glosses over. For example, some skills like scoring efficiency and rebounding are really strongly impacted by SOS, while other like blocks and steals seem completely unaffected by competition. In theory I could account for these issues by including interaction terms, but I prefer to keep the model simple.

[J.E.]

Anybody willing to share their college data?

[VJL]

J.E.

Sure. Sent PM.

[Statman]

J.E.

Sure. Sent PM.

I really want to "fix" the massive amount of data errors that spring up with the 90,000 or so player seasons I'm working with. When my data is more accurate (checked against multiple sources), I'll present all the results - & we can talk about what you might need at that point (older seasons?).