2017 draft analytics

Home for all your discussion of basketball statistical analysis.
pdevos
Posts: 34
Joined: Fri Jun 26, 2015 6:18 pm
Contact:

Re: 2017 draft analytics

Post by pdevos » Tue Apr 04, 2017 5:57 am

As far as Markelle Fultz goes...here's some fun for you.

One player is Markelle Fultz, the other is D'Angelo Russell (Lakers, of Ohio State, 2015 Draft #2 overall).

Who is who?

Advanced (per sports-ref):
https://ibb.co/ija9iv

Per 100 Possessions (also per sports-ref):
https://ibb.co/eaCN3v

Shot selection per Hoop-Math:
https://ibb.co/ctyZGF

Maybe you need some help....bring in some physical attributes...

Here's some measureables on the two:
https://ibb.co/i0s0qa

Looks:
https://ibb.co/hRX0qa

Some might say -- no way, they're totally different? Are they or are they more similar than we'd care to admit for the 'vigor' of this 2017 NBA Draft class.

I'll see if I can put together that shot chart data on them later.

But who is Player A and who is Player B? :)

Statman
Posts: 548
Joined: Fri Apr 15, 2011 5:29 pm
Location: Arlington, Texas
Contact:

Re: 2017 draft analytics

Post by Statman » Tue Apr 04, 2017 6:19 am

bchaikin wrote:I like ALL the draft modeler guys because they are willing to step out & try to produce results that may fly in the face of convention & open them to ridicule.

exactly...

and those that openly criticize/ridicule are rarely those that also do their own modeling - because anyone who has poured through the college player stats for 340+ college teams (and for multiple years) knows what an exhaustive chore it is...

the online experts/prognosticators that bemoan the rankings of modelers will not rank players themselves, only criticize those that do. they are after-the-fact experts...
They also won't produce retrodictions of a model showing all the hits & misses like I have & I believe a few others do. Mine goes back 20 years now.

Statman
Posts: 548
Joined: Fri Apr 15, 2011 5:29 pm
Location: Arlington, Texas
Contact:

Re: 2017 draft analytics

Post by Statman » Tue Apr 04, 2017 6:22 am

tarrazu wrote:I have not done any draft modeling outside of analyzing college stats but for those who have, is any weight put on high school data? For the most part, I'm talking about the 1-and-done players.

I imagine you run into the issues of adjusting for competition, actual data (outside of the top Nike/Adidas tournaments?), and eventually enough meaningful college minutes to wash out most high school data. Might something like FT% and 3PT% benefit from some additional high school data when the sample size is lacking? Is there play by play data with any level of detail at the high school level?
If a team locks me up & gets me the HS data data (at least all the major tourneys), it will be included in my model. I'm too cheap to pay for it myself, & I can't find it anywhere for free. I think it could help the modeling a little.

Crow
Posts: 6208
Joined: Thu Apr 14, 2011 11:10 pm

Re: 2017 draft analytics

Post by Crow » Fri Apr 07, 2017 7:52 pm

Based on 3pt shot rate I'll guess Russell was player B.

shadow
Posts: 218
Joined: Fri May 29, 2015 3:38 am

Re: 2017 draft analytics

Post by shadow » Fri Apr 07, 2017 8:01 pm

DRtg gives it away. Fultz is obviously player A.

pdevos
Posts: 34
Joined: Fri Jun 26, 2015 6:18 pm
Contact:

Re: 2017 draft analytics

Post by pdevos » Fri Apr 07, 2017 11:23 pm

Crow, shadow.

You both are correct! Not that you guys had conflicting answers.

They're very similar in size, reach (identical actually), hairstyle, and most statistics -- except Russell liked the 3-ball a bit more and Fultz was a bit more porous on defense.

Admittedly, I had Russell as my #3 prospect in 2015. So if he flops, I wouldn't have missed it. I still expect him to get better. But might be to cover my own butt. I thought he had much better poise at Ohio State and thought it would show more. But it might be something at LA. I have no ideas honestly.

Fultz, to me, lackadaisical on defense. On offense, wonderful ability to drive downhill and in transition. But I wasn't impressed in half court sets if he couldn't get into the lane. I couldn't tell if it was awful coaching, awful teammates, or his seemingly indifference on defense also showing it's face on defense. I found it a hard time to see he really cared. And "the game" wasn't too easy for him. Washington was bad, awful bad. He had all the opportunities in the world to try to take over games -- as nothing to lose except losing, which is a win.

As hyped as this class is -- I still feel he's the 'default #1' pick vs, OMG, he's the #1 pick! For me, such draft examples would be: Shaq, Webber, Iverson, Duncan, Brand, Yao, LeBron, Rose, Wall, Kyrie, AD, Towns. I only mention that as so many "draft folks" say Ball is surefire HOFer as in the best spawn of Steve Nash, John Stockton, and Jason Kidd -- no kidding ;) Crow has pointed out on a few occasions how few NBA top 10 picks even become average NBA players. Then Fultz is supposed to go before him. Markkanen is Dirk. I just don't see it. If Steve Nash, John Stockton, Jason Kidd, or Dirk Nowitzki were in this draft I've had no problem taking any of them over these guys without blinking an eye. Yet supposedly on equal footing.

If combining the past 3 years. I have Fultz just ahead of Russell and Russell is ahead of Ben Simmons (Kat, Zinger, Fultz, Russell, Simmons). Ball isn't in my top 10. I have a lot of thoughts that feel make me think he's going to bust altogether. Just trying to find a few more trends and video -- wasn't really watching him until the 1st Kentucky game.

My "best case" models [and eye test] say Ryan Anderson for Lauri Markkanen. Though, for me he's more likely to regress closer to Steve Novak or 32 year old Channing Frye. Fultz makes me feel CJ McCollum. But he's not as good of a shooter. Maybe a better passer. But not the (OhMyGosh) comps to 'a smaller James Harden' comps I see.

shock3
Posts: 22
Joined: Thu Mar 21, 2013 7:13 pm

Re: 2017 draft analytics

Post by shock3 » Mon Apr 10, 2017 4:06 am

Statman wrote:
tarrazu wrote:I have not done any draft modeling outside of analyzing college stats but for those who have, is any weight put on high school data? For the most part, I'm talking about the 1-and-done players.

I imagine you run into the issues of adjusting for competition, actual data (outside of the top Nike/Adidas tournaments?), and eventually enough meaningful college minutes to wash out most high school data. Might something like FT% and 3PT% benefit from some additional high school data when the sample size is lacking? Is there play by play data with any level of detail at the high school level?
If a team locks me up & gets me the HS data data (at least all the major tourneys), it will be included in my model. I'm too cheap to pay for it myself, & I can't find it anywhere for free. I think it could help the modeling a little.
I'm guessing Krossovr and Synergy have play by play for Nike, adidas and Under Armour summer games. I could possibly get access to them although I don't know how easy it is to download. I compiled the advanced stats for adidas and Nike last summer, and I'm planning on doing the same this year. I haven't done anything in terms of predictive models, though.

EvanZ
Posts: 912
Joined: Thu Apr 14, 2011 10:41 pm
Location: The City
Contact:

Re: 2017 draft analytics

Post by EvanZ » Tue Apr 11, 2017 10:27 pm

pdevos wrote:
As hyped as this class is -- I still feel he's the 'default #1' pick vs, OMG, he's the #1 pick! For me, such draft examples would be: Shaq, Webber, Iverson, Duncan, Brand, Yao, LeBron, Rose, Wall, Kyrie, AD, Towns. I only mention that as so many "draft folks" say Ball is surefire HOFer as in the best spawn of Steve Nash, John Stockton, and Jason Kidd -- no kidding ;) Crow has pointed out on a few occasions how few NBA top 10 picks even become average NBA players. Then Fultz is supposed to go before him. Markkanen is Dirk. I just don't see it. If Steve Nash, John Stockton, Jason Kidd, or Dirk Nowitzki were in this draft I've had no problem taking any of them over these guys without blinking an eye. Yet supposedly on equal footing.

If combining the past 3 years. I have Fultz just ahead of Russell and Russell is ahead of Ben Simmons (Kat, Zinger, Fultz, Russell, Simmons). Ball isn't in my top 10. I have a lot of thoughts that feel make me think he's going to bust altogether. Just trying to find a few more trends and video -- wasn't really watching him until the 1st Kentucky game.

My "best case" models [and eye test] say Ryan Anderson for Lauri Markkanen. Though, for me he's more likely to regress closer to Steve Novak or 32 year old Channing Frye. Fultz makes me feel CJ McCollum. But he's not as good of a shooter. Maybe a better passer. But not the (OhMyGosh) comps to 'a smaller James Harden' comps I see.
With you man. I am not impressed by this class. 2018, however...

CrackersPhinn
Posts: 14
Joined: Sat Sep 08, 2012 7:56 am

Re: 2017 draft analytics

Post by CrackersPhinn » Wed Apr 12, 2017 10:00 pm

Biggest difference between Fultz and Russell is ability to get to the rim by himself and also draw contact. Fultz was a far more dynamic own-shot creator in college than Russell. Partly that is superior skill. Partly that is superior athleticism.

pdevos
Posts: 34
Joined: Fri Jun 26, 2015 6:18 pm
Contact:

Re: 2017 draft analytics

Post by pdevos » Thu Apr 13, 2017 4:34 pm

CrackersPhinn wrote:Biggest difference between Fultz and Russell is ability to get to the rim by himself and also draw contact. Fultz was a far more dynamic own-shot creator in college than Russell. Partly that is superior skill. Partly that is superior athleticism.
I think you're 100% spot on. Being able to get to the rim (1) and (2), finish. I don't think those skills can be understated. 3pt shooting can be up and down for most players. Even Steph Curry, who arguably, I don't think it's arguably, but it is arguable for many, is the greatest 3 pt shooter ever can string together a streak of 6-7 bad games where his 3PT% is under 30%. Not too many guys will differ by 25-40% off their FG% around the rim for more than a game or two.

In assessing Fultz vs Russell -- stats wise, physical profile wise they're as close a pair of players I've seen in the past 15 years other than the Harrison twins from Kentucky. Thinking out loud -- Is this then where 'scouting' can only tell the difference between two seemingly similar substances?

Back to the ability to get to the rim. It's kinda something I think I missed with Kris Dunn. I never doubted he could get to the rim. [He does, even now]. He's more athletic than 70% of guards who will guard him. But he can't seem to finish [hopefully, that's a "yet" - for the sake of this Wolves fan]. Russell, I loved his pocket passing. He really impressed me with some passes he made, both the vision, and the execution -- although, the latter definitely was questionable at times. He was quick, not fast. Fultz is a whole other level of quick to explosive AND (the 2nd part), he can finish...with either hand, scoops, hooks, finger rolls, floater, etc. Next to Jayson Tatum of Duke, he seems to have as many individual 'moves' developed as you will see in awhile. Mile Bridges, in my view, would be third. Left hand, right hand, post spin left, post spin right, side step, etc.

I watched at least 8+ games of both Fultz and Smith, Jr as they presented the possibility or potential of two the most dynamic explosive scoring point guards since...Kyrie maybe(?). So I perhaps had very high expectations. I'd watch a few DraftExpress videos, a few highlights and plan out my Thursday nigh, Saturday afternoon, or Saturday night to watch them. And in probably half of those games for Fultz and 3/4 for Smith, Jr. They'd just disappeared by simple schemes or "the ball isn't in my hands I'm going to watch this game cause I got a good point of view" type of feel.

My main concerns around Fultz I'd like clarity on, are ironically, applied also to DSJ. I never sensed a 'sense of urgency' about Futlz when a game was within reach. Two things I get from that. 1) my gosh, they were 'this good' without even trying?! and 2) hmmm, don't seem to love competition -- the NBA is a whole different league. Their once outlier athleticism [or all physical attributes] is going to be 'average'...one of many. In fact, on many nights it might even be less than their opponent (e.g. Westbrook, Wall, Kemba Walker, Harden, Lillard, Bledsoe).

But in those games I watched intently, hoping to see this ultra-dominant talent take over...whether down by 3, up by 5 there seemed no sense of ownership or urgency or fight. And Fultz's defense, in many games, it seemed like he wasn't really interested. Most 19 year kids aren't, but also most aren't that obvious about it! Note: he did seem to relish transition defense, that is the fast-break blocks and steals. Jumping, straight line speed (Free safety interceptions) vs lateral get your butt down plays.

Maybe he doesn't like Mozart when it comes to musical taste.... (lame joke attempt at pace/style). On offense, neither seemed to have a game plan or desire or forward thinking on how to beat a team vs a 1-on-1 instance here and there and even more preference for a fast-break. Could this be a preference thing that evolves from playing AAU ball and then playing in the Pac-12? Note: The Pac-12, this year, more than usual, there was a lot of fast-breaks (most teams had 25-33% of possessions in transition per hoop-math.com). FYI Ohio State when Russell was there was 33.9% fastbreak during 2014-15. So noting environment. I'm not sure what the NBA is. Note: http://stats.nba.com/teams/scoring/#!?s ... _FB&dir=-1 shows % of points due to fast-break, but not possessions. I could probably reserve engineer it, asking for a free pass here or easy reference :) But the point in contention would be the nature of the game changing quite bit. Duke, by contrast has 20% transition. I am a big believer of pushing transition at every chance (anything that has greater chance of mismatches or 'forced' decision making for the other team), but pulling back if no obvious advantage [something you coach].

I have less worry about Fultz due to that 'downhill' ability and profile [getting into the lane] than Smith Jr.

But these two share that 'unaware' feeling, which is in contrast to Lonzo Ball, who seemed rather aware -- even in half-court sets. Although, he seems to have trouble getting by (on offense) or staying in front of any guard with decent explosion (e.g. Melton USC, Simmons ARZ). Narrow profile with high hips makes it hard to have balance for lateral explosion. De'Aaron Fox obviously shows what a great athlete can do to Ball. P.S. I hope this isn't seen as 'negatives' for negatives' sake. In my un-expert opinion, there's a strong correlation as to why players don't do well in the NBA is not because of their lack of strengths, but rather it is having weaknesses that are easily exposed. It also seems that Offensive weaknesses are more easily tolerated. DeAndre Jordan can't shoot an 8 footer, make a free throw, pass, etc, etc. But he's those "Bob Knight" qualities. Knight said [paraphrasing], "...tell me who my best rebounder is and who my best defender is and I'll pick the other 3 starters out from the rest." There's many bigs who fit this. If you want a wing, Tony Allen, Thebo Sofa sofa (or however his name is spelled), MKG, Rondo, Pat Beverley, etc.

In trying to make this analytical -- statistically analytical. Would a good project idea be to show which 'weaknesses' are typical 'zero sum game' scenarios for players? And can this be done?

I was toying with this. e.g. Robert Williams for Texas A&M. He's 6'8. Long wingspan. 7'3". We saw Jordan Mickey and many others have this same profile of height, shot blocking ability, and wingspan. Not uncommon to find actually. I did some research of players who were 'rim protectors' in the NBA. That is blocked at least 2 per 36 mins. Of players under 6'9, only Bo Outlaw had more than 2 seasons that met that criteria. If Williams is 6'9, you can add Ben Wallace to the mix. But that's 1 data point, doesn't help. Once you get to 6'10 or taller and were a similar shot blocker in college (e.g. blk per 40) the 'rim protection' seemed to hold in the NBA. Though often, these players also showed very good rebounding ability. So maybe the rebounding ability [likely I believe] got them on the court more than the rim protection, but enabled enough minutes to get those blocks[?].

Another area. "individual defense' in bigs, which I focused on 3 stats -- DReb, Blk, Stl per 40mins. Mostly with further analyzing Lauri Markkanen. His per 40 numbers for those cats respectively, 6.3 dreb, 0.5 spg, 0.7 bpg. It's hard to quantify how bad that is. His most hopeful comp is Ryan Anderson -- don't have dreb stats available, I looked at Cal Bears official site too....nothing. So total rebounding is still a big difference 12.1 to 9.3. Almost 33% less. Soooo....how high would you draft a highly likely outcome of Steve Novak? Or a "taller" Doug McDermott?

EvanZ
Posts: 912
Joined: Thu Apr 14, 2011 10:41 pm
Location: The City
Contact:

Re: 2017 draft analytics

Post by EvanZ » Thu Apr 13, 2017 9:27 pm

Fultz shot 65% on free throws. It's a big red flag, if people are expecting him to be some kind of elite (or even very good) 3pt shooter. I think people should consider what Fultz will be if he's, say, a 34% 3pt shooter. Is he still a #1 pick in that scenario (which I think is closer to an eventuality)? Maybe. I'm not sure I'm sold, but I also think it's a much weaker draft than most people, so he might still be. I'm not sure who I would take ahead of him, unless I reach.

Zach Collins has some impressive stats. I would make him the #1 center off the board. Maybe a top 5 pick?

Statman
Posts: 548
Joined: Fri Apr 15, 2011 5:29 pm
Location: Arlington, Texas
Contact:

Re: 2017 draft analytics

Post by Statman » Fri Apr 14, 2017 3:26 pm

pdevos wrote:
Another area. "individual defense' in bigs, which I focused on 3 stats -- DReb, Blk, Stl per 40mins. Mostly with further analyzing Lauri Markkanen. His per 40 numbers for those cats respectively, 6.3 dreb, 0.5 spg, 0.7 bpg. It's hard to quantify how bad that is. His most hopeful comp is Ryan Anderson -- don't have dreb stats available, I looked at Cal Bears official site too....nothing. So total rebounding is still a big difference 12.1 to 9.3. Almost 33% less. Soooo....how high would you draft a highly likely outcome of Steve Novak? Or a "taller" Doug McDermott?
This site should help you find some stats to work with in terms of your analysis.....

http://basketball.realgm.com/ncaa/confe ... ts/desc/1/

colts18
Posts: 304
Joined: Fri Aug 31, 2012 1:52 am

Re: 2017 draft analytics

Post by colts18 » Fri Apr 14, 2017 4:51 pm

You don't need high school stats. Have you thought about putting high school recruiting rankings in your model? Might help with predictive ability.

Statman
Posts: 548
Joined: Fri Apr 15, 2011 5:29 pm
Location: Arlington, Texas
Contact:

Re: 2017 draft analytics

Post by Statman » Sat Apr 15, 2017 12:15 am

colts18 wrote:You don't need high school stats. Have you thought about putting high school recruiting rankings in your model? Might help with predictive ability.
I could do that - but I have been on a mission to prove proper statistical modeling alone is better than current methodology. Combine results, recruit rankings, mocks, etc. can all smooth out outliers, but I won't use any of that data (yet) because I don't want any doubt from anyone that looks at my work how much weight I put into actual player production (100%). If I were being paid for the work I do, I'd obviously incorporate non statistical data if it helped predictive correlations - because the scope would then be extended from the best model based of player production to just the best model, period. From the work I've done, I'd be surprised if the large majority of the final "as close to perfect as we can get" model wasn't still heavily player production based.

pdevos
Posts: 34
Joined: Fri Jun 26, 2015 6:18 pm
Contact:

Re: 2017 draft analytics

Post by pdevos » Sat Apr 15, 2017 1:37 am

Statman wrote:
pdevos wrote:
This site should help you find some stats to work with in terms of your analysis.....

http://basketball.realgm.com/ncaa/confe ... ts/desc/1/
It does! Much appreciate the re-direct.

Post Reply