Kevin Durant has become a popular topic in this forum. So far the following columns have featured his exploits.
July 12, 2007: Looking Back at the NBA Draft, Part Two
July 7, 2007: Disappointing Durant
July 21, 2007: Durant Disappoints Again
October 31, 2007: Will Kevin Durant Be the Best Rookie?
November 16, 2007: Choosing the Best Rookie in November
November 27, 2007: Evaluating Future Stars in Baseball and Basketball
November 28, 2007: The Top Rookies, Again
These posts tell the following story:
Durant was very productive in college last year. On draft night – based on the college numbers — it was perfectly reasonable to wonder if the Blazers should take Greg Oden or Durant.
And then we saw Durant in summer league play. Although I recall one sportswriter saying that “Durant did not disappoint” this summer, the data told a very different story. Against many players who are not employed by the Association this year, Durant was very, very bad. And, as Table One illustrates, this poor performance continued in pre-season basketball and across the first few weeks of the regular season.
Table One: Kevin Durant in Pre-Season and After 15 Games
My sense is that most everyone understands that Durant hasn’t played very well this year. He has problems hitting his shots and he’s prone to turnovers.
There is an expectation, though, that with time this will change.
Certainly I think that’s possible. But I want to emphasize that there are no guarantees. And to illustrate this point, I want to re-hash the story of Jerry Stackhouse and Carmelo Anthony.
The Jerry Stackhouse story has been told in the following posts:
June 15, 2006:Myth and Measurement after Game Three of the NBA Finals
July 20, 2007: Stackhouse vs. Jordan
October 14, 2007: Looking Back at the 1995 Draft or An Antidote for the Potential Drug
While the Melo story was told in the following columns.
July 8, 2006: One Prediction is (Almost) Right: Carmelo Anthony Gets Max Money
September 12, 2007: Did Melo Resurrect the Nuggets?
October 24, 2007: Melo, King James, and the Human Highlight Film
For those who don’t want to read through these posts, let me review the early career performance of each player.
Let’s start with Stackhouse. Here is what Stack did in each of his first five seasons [WP48 – Wins Produced per 48 minutes].
1995-96: -0.4 Wins Produced, -0.007 WP48
1996-97: -2.8 Wins Produced, -0.043 WP48
1997-98: 4.2 Wins Produced, 0.080 WP48
1998-99: 0.4 Wins Produced, 0.016 WP48
1999-00: 5.1 Wins Produced, 0.078 WP58
In terms of Wins Production, 1999-00 was the peak for Stackhouse. Now in his 13th NBA season, he has yet to post a WP48 that exceeded the average mark of 0.100. And as noted previously in this forum, this is because Stackhouse can’t shoot efficiently and he is prone to turnovers (sound familiar?).
Now let’s look at Melo. Here are his first four seasons.
2003-04: 2.0 Wins Produced, 0.032 WP48
2004-05: 0.2 Wins Produced, 0.004 WP48
2005-06: 4.8 Wins Produced, 0.079 WP48
2006-07: 5.1 Wins Produced, 0.098 WP48
After fifteen games this season, Anthony has posted a 0.065 WP48 and is on pace to produce 4.0 wins. Of course, Melo is still young. And unlike Stackhouse, he can hit his shots at an average level of efficiency. But Anthony is simply not outstanding at any aspect of the game. And to be outstanding, you have to do something outstanding. Yes, it’s that simple.
Let me note that both Anthony and Stackhouse did improve after their rookie season. But let me also note that both performed at a higher level than Durant is performing at right now. So Durant is going to have to improve more just to get to the level of these players.
And I also want to note how far Stackhouse and Anthony improved. Neither player has reached the level of productivity one would associate with “a franchise savior.” Certainly that’s what Seattle’s fans were hoping for when this team took Durant last summer. But the early returns tell us that the savior hasn’t arrived yet.
Let me emphasize that it’s still very early. Durant could certainly improve. But right now, he’s not a very productive basketball player.
And that is my point. Players should be evaluated in terms of what they actually have done. Not in terms of what we imagine they might do at some point in the future. When we look at what Durant has done, he is not the best rookie in 2007-08. He isn’t even in the top five or the top ten. Yes, and I repeat myself, he could become the best rookie. But based on what we have seen so far, there is no guarantee that’s going to happen.
– DJ
Our research on the NBA was summarized HERE.
The Technical Notes at wagesofwins.com provides substantially more information on the published research behind Wins Produced and Win Score
Wins Produced, Win Score, and PAWSmin are also discussed in the following posts:
Simple Models of Player Performance
What Wins Produced Says and What It Does Not Say
Introducing PAWSmin — and a Defense of Box Score Statistics
Finally, A Guide to Evaluating Models contains useful hints on how to interpret and evaluate statistical models.
Wages of Wins Journal 
Baba O'Riley
November 29, 2007
This is a good post. Durant missed 15 of 21 shots tonight. That just isn’t going to help a team win.
dustin
November 29, 2007
It helps the other team win :)
I’m sorry I had to say that.
Tim
November 29, 2007
So tomorrow we hear about the Bulls?
Rasta
November 29, 2007
Good summary of the Durant debate.
You probably set an all-time record for most links in a single column.
Tommy_Grand
November 29, 2007
Thanks for an interesting column. It got me to wondering about the performance 2nd year players. For Portland, Roy and Aldridge look decent (based on my Mark-1 eyeball analysis). I’d guess both are productive, above average players. Does that bear out statistically, or does rigorous, non-subjective ananysis put the lie to my intuition?
Tommy_Grand
November 29, 2007
I also like the performance of Ronnie Brewer, Daniel Gibson, and Rudy Gay.
Owen
November 29, 2007
Brewer has an amazing skill set. He is a below average rebounder, but he has more than twice as many steals as turnovers so far this year. (42-19), Must be a record of some sort….
Happy to have him on my fantasy squad…
mrparker
November 29, 2007
There was an interesting study using similarity scores on 82games.com to predict NBA success. In his draft class ronnie brewer was predicted to be the most successful.
It was written last summer if you are interested in finding the study.
Jon
November 29, 2007
I will say that Durant will never be a great Wins Produced player. So Berri will certainly never believe he is a great player But I also don’t think it is fair to compare Durant to Melo or Stackhouse or really anyone, because what he did in college has never been done before. Also, while I don’t know about Stackhouse, Melo was physically developed by the time he came into the league. Durant clearly is not. In my opinion, that is one of the leading causes for his low shooting percentage, high turnover rate, and a rebound rate that is lower than expected.
dbg
November 29, 2007
It’s interesting that you compared Durant to Carmelo and Stackhouse. Both of those, according to Wins Produced, are not nearly as productive as they are perceived to be. But with Durant perception was not the problem. He really was productive in college. I don’t think Seattle can be faulted for picking him.
So my question is this. By comparing Durant to Carmelo are you saying that you think Durant will improve enough to become slightly below average but will be hailed as a star?
dberri
November 29, 2007
dbg,
I think I am saying two things. First, right now Durant is not playing well. Second, as the case of Melo and Stack demonstrate, playing time does not necessarily turn you into a productive player.
That being said, I do think it is possible Durant is going to develop into a productive player. The sample we have seen so far is still too small.
But what we have seen is not good.
I would also add that even if Durant doesn’t improve dramatically, if he continues to score he will be thought of as a star.
Baba O'Riley
November 29, 2007
Bill Simmons spoke about Durant today, “I still love Durant, but it looks like he’s stuck at the wrong college with the wrong major and the wrong friends right now. Play him at the “4,” keep him near the basket on defense and let him do his thing, for God’s sake. “
mrparker
November 29, 2007
Guys,
Melo averaged 22 and 11 in college on a team that didn’t necessarily run the way texas did last year.
So how is what Durant did last year never been done before?
Josh
November 29, 2007
I won’t argue that Durant is playing well right now.
But it’s just silly to pick two other players and argue that, because they didn’t improve markedly, the odds are that he won’t improve much either.
To make the comparison, you’d have to adjust for the quality of the coaching each player gets, the system he’s placed in, whether he’s playing the proper position for his skills (and how you define that is a big question in itself), the quality of the players surrounding him, and probably a dozen other factors.
Re the Simmons comment: Carlesimo has said one reason he’s not playing Kevin at the 3 or 4 is because he doesn’t want him to get physically pounded by bigger, stronger men on defense. That may or may not be a good decision, but it’s at least reasonable.
dberri
November 29, 2007
Josh,
You mis-read my argument. I wasn’t taking a sample of two and drawing a conclusion about Durant. I was merely pointing out that it is not a given that Durant will improve. And I used Melo and Stack as two players who did get to play quite a bit but never improved.
As I have said, it is still quite possible that Durant will improve. It just isn’t a given that it will happen.
Jason
November 29, 2007
I’ve always found the perception of players to be something rather strange.
I’ve said this before but I think it bears repeating: a player who came out of college with *remarkably* similar numbers to Durant was Glenn Robinson. Granted Robinson’s numbers came after two years at Purdue (ineligible as a frosh and leaving after his Jr. year) three removed from HS rather than one removed from HS, but nonetheless, the level of collegiate dominance and statistical line was strangely similar, even if you compare only their first years in college. Robinson scored well for his career and wasn’t *terrible* but was somewhat below average in overall efficiency. He was actually pretty close to the median in efficiency. Not what you’re looking for in the #1 overall.
What’s interesting is that while Robinson did make it the all-star game a couple of times (indicating that his scoring got rewarded), for most of his career I think people didn’t regard him as a star, but as a ball-hog who scored at his teammates’ expense and didn’t really improve his club despite his 20ppg.
I’m not certain that Durant will be regarded as a star just for scoring. I suspect that once the honeymoon of the rookie season wears away, his team will have to win some games, else he too will get the ball-hog rep.
dberri
November 29, 2007
Good comment, Jason,
I would add that I think Robinson played almost his entire career under the contract he signed after being drafted. So we do not know how Robinson was valued by the market. PERs says he was above average almost his entire career, so that suggests he would have been paid pretty well on the market (although, as you note, he was not very productive NBA player).
Jon
November 29, 2007
While the stats between Durant and Glenn Robinson may look similiar, how they got them is compeltely different. Their skill set are completely different. Durants skill set (ability to get to basket, break people down on drrible) more closely resembles a Kobe Bryant/Tracy McGrady type with a little less quickness and a little more height than a Glenn Robinson/Carmelo Anthony type. And besides the points and rebounds Durant put up, he also had a very good amount of steals and blocks. That leads me to believe that Durant will have a different career than Robinson.
Don’t know if anybody has read Hollinger’s article, but he came up with some formula that predicts how well a college player will do in the NBA. While it is by no means perfect, it has shown more accuracy than a player’s draft position. (Hollinger drafts better than teams do) And Durant BLEW his formula out of hte water. I believe Hollinger said Durant was the greatest prospect (according to his formula) of the past 10 years by a wide margin.
Jon
November 29, 2007
** Maybe I was wrong to categorize Glenn Robinson and Carmelo Anthony into the same skill set, I don’t know what you think **
Owen
November 29, 2007
Econometrics is a very useful discipline, but I am afraid no academic approach can begin to quantify how comically horrible it is to be a Knicks fan right now…
Totally off topic I know…
JW
November 29, 2007
Maybe you do this in other posts, though I check here rather frequently and haven’t seen it, but it seems that you give short shrift to aging curves and projection. Of course, when we ask ourselves “how good is Durant right now” we can give an objective answer and analyze accordingly.
However, if we want to look at how a player projects, we have a lot more to look at. You try to look at comparable players and say “he might not improve” but it’s much more important to look at the interactivity of a player’s skill set and see how that improves with age.
Nate Silver does this with PECOTA quite rigorously. Delmon Young, for example, didn’t exactly produce this year. He was probably a below average outfielder overall. However, we take his age, his isolated power, build, etc. into account and can project him into the future as a very useful player.
I’m not saying it can be as precise in basketball, but I think it’s an important point. How do skinny 19 year-old, 6-9 rookie shooting guards with large wingspans develop given the performance they put up? Do we interact the shooting percentage with the age, and does that predict their future better?
Jason
November 29, 2007
When I said remarkably similar, I wasn’t simply referring to points and rebounds. Across the board they were remarkably similar.
I’m not going to try to judge this as a beauty contest and explain *how* they got their numbers (e.g. what ‘skillset’ produced them). But whatever skillsets they had led to remarkably similar results across the board, with Durant doing a little better at almost everthing, but nowhere (save protecting the ball, and I suspect this may have something to do with the rest of the team around them–Durant having superior teammates) does anything scream that Durant was in another class from Robinson.
Robinson went #1 overall and I don’t really remember it being considered a questionable pick at the time. He went #1 in a draft where 2 and 3 were also considered ‘can’t miss’ players. So I think that the comparison between results and perception of what those results would translate to on the NBA court holds.
Jon
November 30, 2007
All I mean with the skill set is that player develops his certain skills. For example, *maybe* a player like Durant who has a better ability to get to the basket, use his dribble, is a skill that takes more time to develop in the NBA. It is not just about numbers, but how those numbers got there.
I am not saying this is anything like a beauty contest. But just because two players produce the same output at a certain time (although different age) does not mean they will continue to produce the same output. If Durant has a different skill set, he will develop differently than Glenn Robinson did as he refines his skills. Those skills will take him somewhere different than where Glenn Robinson’s skills took him. While a different skill set may lead to the same college numbers, that is no indication that it will lead to the same NBA numbers. And just because after 15 games Durant has been an inefficient shooter as Robinson had, that is no reason to conclude they are the same type of player and Robinson’s career path is a likely destination for Durant
Jon
November 30, 2007
Maybe this will clear up what I am saying. Antonio Gates as a senior at Kent State average 20.6 points per game, 4.1 assists, and 7.7 rebounds per game. I am sure you can find other players who have produced that or something similar to that and been a productive NBA player, or even an NBA player. Yet Antonio Gates did not make the NBA. His skill set (which includes size) did not fit the NBA. The same applies to Durant and Robinson. While both put up great numbers in college, one of their skill sets may be more suitable for the NBA game. In this case, many people believe that Durant’s is, and this will in turn make him a more productive player.
mrparker
November 30, 2007
Jon,
Antonio Gates is/was a 6’5″ power forward. Gates does not have NBA size. Carlos Boozer was a 2nd round pick because of his lack of size and he’s quite a bit larger than Antonio Gates.
The entire point of this site is to dispell the notion that scoring = winning. While Durant possesses the ability to inspire his coaches to let him shoot the ball 20 plus times a game, he is below average in every other measureable aspect of the game except stealing and blocking. Berri’s research contends that Durant(for right now at least) is Adam Morrison in 06 with a better ft/fg rate.
Were the NBA scored by judges giving marks for technical merit, presentation, and short program then Durant would consistently be helping his team win. However, since this is the NBA and not figure skating Durant will not be helping his team win any time soon. In fact by my calculation he would have to shoot the ball at 70 plus percent just to get to average.
In short unless paired with an all time good player Durant will never play for a team that competes for a championship.
mrparker
November 30, 2007
Not to put words into Berri’s mouth. My saying Berri’s research was a reference to similar wp48 for two rookie seasons. I was in no way shape or form trying to speak for dberri.
Jon
November 30, 2007
I know. I am also glad you can reach a conclusion on Durant’s career after 15 games. Very impressive, that you can already conclude Durant will never be able to play for a championship. Not once did I make the argument that Durant is playing well right now. My whole argument is Durant will grow more as a player than Melo or Stackhouse or Robinson.
My whole point with Antonio Gates was it is not just about the numbers you put up in college. It is also about your skill set. Which means height, athleticism, what your actual basketball skills are, are all taken into account. And considering Durant’s heigh, wingspan athelticism, and skills he brings regarding ball handling and ability to get to the rim, once his body fills out and he gets more experience he will improve vastly as a player. Also, his rebounding numbers are not terrible for a shooting guard.
Jon
November 30, 2007
And unless Jason or dberri back you up with your conclusions, I think it is absolutley ridiculous that you claim he would have to shoot 70% just to be an average player. Jason or berri can you verify that? What would his wp48 be if he shot 70%?
mrparker
November 30, 2007
firstly,
70% would put him at average for my own calculations which have predicted successful transitions to the nba from college. You can go back to the last post on rookies(top rookies, again) to find what I wrote in June is close to true now.
Im not sure what wp48 would be if he shot 70%. I don’t pretend to know how to calculate the exact formula for wp48 but my numbers are pretty close, but not always in agreement. I came up with my own numbers because I needed something that gave the same results for college and pro basketball which winscore does not do. So my numbers are different, but are usually in agreement for the nba but not in agreement for college. For instance, my numbers say Durant wasn’t even the best player on his team last year. But we will have to wait for Augustine to turn pro before this can be proven true or false.
Secondly,
Durant’s been praised for averaging a double, double in college. My numbers said that wouldn’t translate to the NBA which has proven true so far. Emphasis on so far. My prediction, not conclusion, is that Durant will not be productive enough to lead any team to a championship. However, I actually do think that he will end up being around a wp48 of .15-2. His shooting percentage will improve,he’ll get to the line more and his turnovers will go down. That being said I dont think he is in the top 5 of his own draft class.
mrparker
November 30, 2007
Third,
Melo is an interesting player. My contention is that if Denver chose to use him as a power forward that his production would sky rocket. They don’t have the roster to make such a move but Melo is being underutilized as strictly a perimeter player. If you watch any all star/olympic game where he can play the 4 and therefore choose to play anywhere on the court without causing spacial issues he is a force to be reckoned with. However he lacks the all around game to be of much value as stricly a perimeter player, but would be a good enough rebounder(think better version of Shawn Marion) to play the 4 for a team that plays uptempo.
I guess my point is that I don’t agree with any contention that Carmelo will always be below average.
Four,
Sounds like you love analyzing basketball just like I do. I didn’t mean to come off as preachy or as someone who was intentionally trying to piss you off. This Durant topic gets my blood boiling and I tend to hit the submit button without rereading what I wrote. Im pretty sure I sound like a prick to someone else reading what I write.
Thats not my intent. I have strong feelings about the NBA especially since our Olympic teams haven’t been doing well. I tend to really scoff at any player being talked of as elite who isnt what wp48 scores as above .25 because they might end up on th olympic team and we might end up losing a game when in reality we are so far ahead of the world that we should never lose a game by more than 1 point.
This is pretty far off topic by now but with our young guys being scorers as well as productive(chris paul, lebron james, Dwight Howard, Melo when he’s playing the 4) I don’t see this being a problem anytime soon. But I will always contend that Durant belongs nowhere near the olympic team. Thats the first thing that I think of when I hear Skip Bayless and others slobbering over Durant. Im tempted to pray for this not to ever happen.
good night
Jon
November 30, 2007
How are you able to predict what college players will do in the NBA? Because I know Hollinger just ranks the college players and predicts how good he will be, but does not even attempt to predict what their numbers will be. He just gives a list of how good of a prospect they will be. I find it extremely interesting that you are able to predict Durant won’t be able to average a double-double
Baba O'Riley
November 30, 2007
FWIW, I predict Durant will improve once he plays with better people and doesn’t have to spend all his time on the perimeter.
Jason
November 30, 2007
I understand that traditional scouting takes into account “skill set” when evaluating players. It’s not that I don’t believe people do this or that I don’t understand *how* they do this, it’s a question of whether or not it matters, and that’s something that I’d like to see as a more objective question.
I’m also not sure that at draft time the opinion of Durant was much higher than the opinion of Robinson before he was drafted. I don’t remember anyone thinking Robinson had anything but superstardom written all over him. He went #1 in what, at the time was considered a very deep draft. Don’t evaluate what people thought of him with the advantage of hindsight. In comparing the *expectations* and the *opinion* of his skills (not numbers, but “skills”), think to what people said at the time.
What I wonder though is if evaluating based on “skill set” has a significantly positive predictive measure over evaluating by more objective measures. I’m separating out physical stature from ‘skill set’ as that’s an ojbective measure. You’re either the height and weight you are or you aren’t. “Athleticism” gets fuzzy. We can measure how high a guy can jump, how fast he can run on a given day, but these are somewhat elastic as well and don’t necessarily indicate what a player does on the court. But saying you’re a slasher/driver scorer vs. a jump-shooter perimeter type is more of a value judgement.
My point isn’t that there can be a difference. My point is that these differences are subjective and I do not know that the subjective opinions are going to result in more accurate predictions. Do we know that it is?
I’ve seen a ton of “can’t miss” players miss. Some of them it was clear weren’t able to handle the game physically, some appeared to be head cases. Some though are mysteries. Jimmy Jackson and Reggie Williams were both regarded as ‘can’t miss’ do-everything players (drafted by teams that needed everything and eventually turning into something amounting to adequate subs. I *don’t* recall anyone saying that Robinson’s “skill set” was going to turn him into a pedestrian NBA player. I recall him being the “can’t miss immediate impact” player. I recall people marveling at how strong he was, how high he could jump, how he could burn you from anywhere on the court (unless, like Dook did, you spent all your efforts to make him try to drive left on every single possession). The objective measures said he was going to be great. The subjective ones said this too. He wasn’t.
The PECOTA system intrigues me. If we handle everyone like a complete individual, an isolated case, the hype machine that seems to surround basketball like no other sport, can run away with a player quickly, pronouncing stardom (sometimes long after a player is demonstrably not a star) based on a court of opinion, arguing how unique a player is, how this combination of skill and size blah blah blah. This can and does snowball rapidly. Before the draft you hear about a player’s “stock rising” but that’s just a public opinion poll and they tend to be positive feedback loops more than honest evaluations. You see it here when yahoos jump out and scream (in their own borderline literate way) that there’s no way that Moon can be better than Durant and anyone who says otherwise is an idiot. You see it when people fume at the notion that rebounds awarded to particular players really could be that important to wins. The arguements to support such fuming are either assertion (which is an argument of opinion) or anecdotal (which is opinion combined with a single datum).
What I believe Dave and colleagues are trying to do is look beyond the opinion and hype, question conventional wisdom, and see if the opines have value or if they’re off target. They are trying to separate out those things that are opinions, and thus subject to feedback loops associated with opinions that can render them meaningless, but, because of the deep held belief in them, unassailable in the minds of those who hold them, from those things that we can hold as truly tangible and see what of those tangible measures are actually indicating success and/or predicting future success.
There are many variables beyond the stat sheet. These *may* (or may not) be important. If there was a way to classify a player and compare them based on objective similarities other players to see what their development curve looked like, we’d have a tool to see if patterns were there or, if not, if it was either pure dice roll or perhaps the individual scouting really was the only way to go.
mrparker
November 30, 2007
Jon,
I wrote extensively on the comments section during the draft on my favorite prospects. My system is somethign that I’m unwilling to share, but I’m not unwilling to share my predictions, because they tend to be true. I hope to one day take my research to a team and be a memeber of the front office. If this is at all possible it will most likely take 10 years or more but its not a possibility that I’m putting out of the question. If my work is that good then it should one day prove profitable and I’d like to make sure I’m the one to profit from that research.
I’ve been at this for only 3 years now so I actually only have 2 classes of players under my belt from which to point.
My favorite players to come out have been Chris Paul, Sean May, and David Lee.
This year’s draft had 4 of my favorite players in Conley Jr, Noah, Horoford, and Oden.
Basically, its the way the player plays the game and not necessarily the numbers which predict NBA success.
Im not here to steal Berri’s thunder, this is his site and I’m blown away by the way he put together actually predicting wins. My system still can’t do that..I guess its not a goal of mine, but if it were I probably wouldn’t succeed.
spike
November 30, 2007
I’m new to this stuff. Is 35 points on 60% FG (from 2- and 3-point range) to go with 5 boards, 2 assists, 2 steals, 3 blocks and 7 turnovers good? Yes, Durant is turnover prone, but he also won the game for Seattle tonight with two huge threes and enough clutch free throw shooting to keep it out of reach.
Your “4 favorites” sure are lighting it up so far this year. As is Sean May.
Jon
December 1, 2007
mr parker, it still really doesnt make any sense to me that you can claim durant will not average a double-double based on his college performance… until you explain more it really doesnt make any sense
mrparker
December 1, 2007
spike,
Sean May is injured but Charlotte would have been a force if he had been able to play this year.
One man’s opinion of course. Appreciate the sarcasm though.
Jon,
My research is about the way that people play basketball. The way that Durant plays basketball he would never sacrifice more shots for more rebounds. I guess my number represent adaptability. Durant is a pure scorer, and plays far away from the basket. If he does average a double-double then I will have to scrap my research because he will have proven me wrong.
spike
December 1, 2007
The main point “pro-Durant” people have been trying to make is that Durant is not ready for the role he’s been given. Conley and Noah are playing pretty well in very limited limits, while Horford is doing well as his team’s 4th or 5th option. Based on their performance so far, are we sure they’re going to be “team saviors” either? Can we even say that Conley would be doing as well if he weren’t being played in such limited circumstances?
It seems like Durant just brings out the stat geeks because they’re desperate to prove that scoring isn’t as important as everything else, so they put their crosshairs on him. What these statistical measures really do is tell you how well a player is doing at filling the role he’s been given. Conley is doing very well as the team’s backup point guard who only plays 11 minutes a game. Horford is doing very well as the team’s “guy who needs to rebound and take fewer than 7 shots a game”. Durant is performing slightly below average in the “team savior” role. But take Horford two years ago and see how well he’d do in the “team savior” role as a 19-year old.
When assessing who the best rookie is, you can either look at the guys who won all of these statistical measures, without regard to who was asked to do what, or you can incorporate the role of each player as well and see who had the overall best season. I still think Durant is having the best season so far because he’s playing as a 1st option on a team with no other established pros.
mrparker
December 1, 2007
I can’t say that I’m in the anti-durant camp. Like I said, I have him rated as around what wp48 would rate as .15-.2 at his peak. That being said I am also in the “Durant will never do for his team what the guys you mentioned in your previous comments will do for theirs” camp.
I can see where your coming from as far as roles in the team go. However, its my contention that if Durant were going to be all-world he would be able to create for his teammates, which he hasn’t done at either the college or pro level so far. A productive basketball mind says..I shoot 40% and all the defense is focused on me so let me get someone else a better shot.(Conley Jr. mentality) Or all the defense is surrounding me so I will give and go with my point guard to get closer to the rim for o. rebs or easy buckets(brandon roy mentatlity).
Durant has the doin it myself mentality which hasn’t produced a championship for any 1st option without the help of another player who was an absolute force and didn’t need the basketball to be productive.
I guess I will accept the tag as stat geek, which is ironic in some ways.
Jon
December 1, 2007
Not to be offensive, but trust-me, you don’t get the label as “stat-geek”. Until you have proven yourself much more as a statistician, you don’t deserve it. People like dberri and Jason who make claims and back it up with pure statistical reasoning deserve it. In my mind, you have yet to do that.
Jon
December 1, 2007
And to Dberri and Jason, I was not trying to make fun of you or anything, I was just saying that I respect what both of you say because on all of your posts you have shown great statistical knowledge
mrparker
December 2, 2007
Aw damn, I was really hoping to be labeled as stat geek. I guess I’ll just have to be happy with hitting 65+% of my bets on basketball because of everything that I understand that most people do not.
My new goal in life is to prove myself to Jon.
Jon
December 2, 2007
All I am saying is you just keep making claims. I predicted Durant would not average a double-double. I project Durant to eventually have a WP of .15-.2. I hit 65%+ of my bets on basketball. You make claims, but back none of it up with the proof or your work. Compare that to what Dberri or Jason posts and you will see where I find the difference.
mrparker
December 2, 2007
That is why I don’t have a blog.
I made predictions before the season on this season’s crop of rookies. So far everything I said has proven true.
Im not sharing my work, but the predictions were made before the seaon.
If you were around before the season you would have seen what I wrote.
I reposted it in the comments section of the post titled “Top, Rookies Again” if you want to see what I wrote 5 months ago. The comments section is pretty long but I did state my opinions before anything happened. If those predictions end up being solid then theresearch must be solid as well.
https://dberri.wordpress.com/2007/06/26/win-score-and-the-nba-draft/
Here is the link to my tracked wagering numbers if you want to see where I am with my wagering
http://www.wagerline.com/sportscontests/profile.aspx?user=-51884&type=0&sportID=7
I have tried to make piece with you but seem hell bent on being snarky. Cool with me. I’ll be around we’ll probably always disagre but I’ll be right more than I’ll be wrong.
Jon
December 2, 2007
First of all you claim everything you said about this year’s rookie is true. For you to say you are right or me to say you are wrong would just be plain stupid. 15 games is nowhere near enough to judge a rookie class. And anyways, even through 15 games, some of your predictions are questionable. Acie Law? Aaron Gray?
And than regarding some other statments you made. Chicago Bulls are the best run franchise? Hawks should just give up on Marvin Williams. While you have made some correct statments, others are wrong. As expected.
And regarding your waging, what happend in ’06 and ’05. It seems to me like it is more random whether or not you do well. For some who follows basketball extensively, that is not surprising. It is not like you don’t know what you are talking about. But what you say still remains a guess, which explains why this year you are doing really well why in other years past you have not done so well
Also, I just don’t understand how a system can take college numbers and predict the actual numbers they will get in the NBA. John Hollinger, a respected statistician, just ranks them as prospects. He does not attempt to predict rookie numbers. Until you show how you are able to get your numbers, of course I am going to be skeptical.
Look, I am not saying you are stupid. My point is, anyone who follows basketball as much as you do, regardless of the “system” you use, will be able to make pretty good educated guesses on basketball. But this system you use, until it can be looked at by other statisticians, has no credibility. I am an avid NBA fan, and many predictions I have claimed have been true. Many others have been wrong.
If you like your system so much, I am just confused as to why you would not publish it and reap the benefits (cash). I mean you already use it to make money in the first place (wagering), so why not just try to make more?
mrparker
December 2, 2007
jon,
The bulls have one of the best rosters in basketball. That is if you wan to measure by the % of above average basketball players. IMO, that is the best way to evaluate a front office. Because that means that when players get injured a capable replacemnt steps in. If I’m running a franchise I would want to minimize risk. But thats just me.
Aaron Gray is at 16 and 10 per 40 with requisite high rookie turnover rate. I think that is great for a bench player. That is alot of production out of a guy drafted that low.
Acie Law is shooting 50% getting 10 pts 6assists and 2steals per 40 minutes which is great for the first 15 games.
Mike Conley is at 15pts and 11assists and shooting 50% per 40 minutes
I feel as though young guys progress, so seeing guys that I liked doing well early is a good sign and they should all get better not worse.
As far as wagering goes, in those two years I made over 1500 bets. In the years where I bet less my percentage is much higher. Not that it matters but I couldn’t get my win% predictions(for seasons) to go along with game by game wagers. Now with the help of sites like this where alot of brainstorming goes on I learned how to do that. Now I can pick and choose my bets better and my basketball win% are pretty good.
Im unwilling to share because I have had other work that I’ve done before stolen with no credit given to me. So, what I think is understandably, I’m unwilling to share something that alot of other people might be able to use until its copyrighted or trademarked or whatever I would need to get done to get credit.
I don’t want any Gil Brandts out there making me their David Lewin. That was one f-d up situation and something that I hated to see happen. Especially since my projection were partly inspired by his.
I guess I’m not looking for any validation for now I’m just a dumb kid who brags ever time he’s right about something which cheeses alot of people off and is something I need to knock off.
If my system is unique and successful then I can rent myself out as a consultant in the future. Possibly at least.
Kent
December 2, 2007
Mr. Parker, good luck with your system.
mrparker
December 2, 2007
oh yeah and Marvin Williams,
Atlanta has Josh Childress, Marvin Williams is ok but Josh Childress is a very underrated asset and doesn’t have to shoot. But shoots the ball very well.
I shouldn’t have said they should give up on Marvin Williams. Its possible that I was drunk when I wrote that and if so I still should be slapped for writing that.
mrparker
December 2, 2007
thanks for the well wishes kent
Jon
December 2, 2007
I am not gonna argue with you over all of these points, as it is pointless and won’t get anywhere. On many of the things above, we just look at it much differently and nothing either of us say will change that.
But all I will say is for me personally, while you may do great work with your system and projections or whatever it is that you exactly do, I for one will be skeptical until I see the work behind it.
Maybe you consider me a jerk, but that is just how it is for me. I don’t like looking at Dberri because I necessarily agree with all his remarks, but because I am interested in his work and how he got there. I am not impressed with predictions and claims, as I have seen that many of the best basketball minds still are not correct with their projections. I am more impressed with how someone comes up with their projections and claims. Because anybody can make a prediction, but only very few can back it up with great statistical analysis. So until you get your work published, I will always be skeptical, even if I agree with some of your points. (Which I do, we in fact like many of the same rookies. I just have no statistical reasoning behind it, I just like them. You have statistical reasoning behind it, I believe, but choose not to share it.)
Jon
December 2, 2007
And just so you know, when I say I have no statistical reasoning behind it, I mean I have no formula or metric that I have created like you or Dberri. Obviously I consider stats when assessing a player or team
Kent
December 3, 2007
“I am not impressed with predictions and claims, as I have seen that many of the best basketball minds still are not correct with their projections. I am more impressed with how someone comes up with their projections and claims.”
Bah! If someone makes a sequence of predictions that are all correct than what does it matter what the methodology is?
Jon
December 3, 2007
Well that hasn’t happened. Nobody will continue to be right all the time without some methodology. I mean making a few correct guesses or one. Now if someone were to predict the records of all the teams and get them right than of course regardless of the methodology I would be impressed. I am talking about more reasonable predictions. And I am not saying Mr Parker is not an idiot. He clearly has a great knowledge of basketball. Let me clarify what I am saying.
If Mr Parker were to make an overwhelming amount of predictions that were extremely accurate, than yes, of course I would be amazed. But what it seems like to me is that he makes predictions, and for the most part he is fairly accurate, but of course there are times where he was wrong. And this is to happen with almost anybody who follows basketball as closely as Mr. Parker does and makes predictions. So what impresses me is when I see statistical proof of their or reasoning for their prediction. Or in other words, what Dberri does, Mike G of APBR metrics, John Hollinger, etc. If you can see the difference between Mr Parker and those three, it is that they have published methods while Mr. Parker has kept his a secret.
mrparker
December 3, 2007
I guess trying to back up my claims about Durant with “proof” of other claims I’ve made is a bit silly.
All I’m really trying to say is I see Durant as a guy who’s going to get about 5 reb and 3 assists a game eventually. I feel this way because rebounding from the wing positiion in college has never translated to the nba for any other player. Glenn Robinson and Carmelo Anthony are two recent members of the college sf double double club who ended up averaging around 6 rebs a game. Also Durant was poor at creating opportunities for his teammates in college averaging under 2 assists per game.
These claims could be a mit more substantiated if Texas goes on to be better offensively without him this year than they were with him last year. So far it appears that this might be the case(133 offensive efficiency vs. 114), but it is very early in the season. This at least has happened against similar competition with last year’s and this year’s strength of opponent’s defense averaging out to approximately 97.4.*data provided by kenpom.com
Kent
December 3, 2007
“And I am not saying Mr Parker is not an idiot.”
Sneaky double negative! So you’re saying Mr. Parker is an idiot?! ( :-) I realize it’s a typo)
Jon raises an interesting question. If you have to choose between 2 predictors which do you choose if one has slightly better accuracy but a black box methodology and the other has a fully transparent methodology but slightly worse accuracy? Interesting question.
mrparker
December 3, 2007
transparency…that way you can check the output
spike
December 3, 2007
“In the years where I bet less my percentage is much higher. Now I can pick and choose my bets better and my basketball win% are pretty good.”
This backs up my point pretty well about Conley, Law, and Noah. When your sample size is around 10 minutes a game, does that really predict what their performance would be if they were to play 40 minutes a game?
Also, I’d say 6 boards a game from a small forward is pretty good, considering no wing player has more.
mrparker
December 4, 2007
I think that Berri has done research that supports a notion that production doesn’t change with an increase in minutes. I could be wrong.
6 boards a game from a wing is not bad but it still doesn’t get these guys above average on wp48.