PF Draft Prospect Stats
It's beginning to seem likely that if the Blazers hang on to their #24 pick they will draft a PF to back up LMA. Now, this first tells us a few things about the type of player the Blazers are likely to draft. First, the player has to be available at #24 or worth trading up for. Second, it seems unlikely that the Blazers will draft someone based on potential, as they are not looking for a starter. They'll be looking for a productive, efficient player who can back up LMA. This means we can narrow our search to productive, efficient power forwards that will be available at #24 or worth trading up for.
The big question then becomes how to measure "productive and efficient." It seems obvious that the best way to measure a player's productivity is to measure how many wins they create for their team (that's the point of the game, to win). A more productive player will produce more wins. It's as simple as that. Of course, that still leaves us with the question of how to measure the wins a player contributes. Fortunately, we actually have a stat for that. It's the Win Score, which was developed to come the closest to predicting a player's contributions to a team's win total.
This means we can take a look at a player's Win Score in college to give us an idea of how productive they will be at the NBA level. Conveniently, DraftExpress provides a Win Score per 40 on almost all potential draftees (they obviously don't have them for the Euros). Of course, these stats have to be taken with a grain of salt. Not all productive college players make productive NBA players. Many of them do, but some just have a style of play that suits college much better than the NBA. So these stats have to be combined with a look at how well a player's game could translate to the NBA. But for the most part, productive college players make productive NBA players.
So first, we make a list of PFs who are likely to be drafted:
Blake Griffin
Jordan Hill
DeJuan Blair
Tyler Hansbrough
Taj Gibson
Jeff Pendergraph
Josh Heytvelt
Jeff Adrien
Now, let's narrow down the list. First, we can throw out Blake Griffin and Jordan Hill as early lottery picks that the Blazers almost certainly won't get a chance at. DeJuan Blair and Tyler Hanbrough are projected to go ahead of #24, but we'll continue to consider them because they're not out of trading range. Now, let's find the most productive out of these in terms of win score per 40:
1. DeJuan Blair 21.8
2. Tyler Hansbrough 14.1
3. Jeff Pendergraph 13.4
4. Taj Gibson 13.3
First thing you'll notice out of our top four is that DeJuan Blair was an amazing contributor to Pitt's win total. He even had a better WS/40 than Blake Griffin. Of course, I'm not saying he'll be better than Blake Griffin. There are several issues with his game (6'7" guy with bad knees) that suggest that his game might not make a smooth transition to the NBA level. Still, he could lose more than seven points off of his win score and still be better than Tyler Hanbrough. For a reference point, this blog (which is the king of all stats win related and worth a read if you find this post interesting) gives us an average WS/40 for PFs drafted in the league to be 12.48. In other words, all of our top four are above average, but DeJuan Blair significantly so.
Let's assume however, that the Blazers don't trade into the late lottery to get Blair. We now are faced with a choice between Hansbrough, Pendergraph, and Gibson. As you can see, the difference between them is negligible, especially considering room for continuing change and development in the NBA. Hansbrough is slightly better than the other two, but probably not worth trading up for. Between Gibson and Pendergraph, the difference in productivity is nonexistent, and it simply comes down to whomever the Blazers like more. Considering how Pendergraph seemed to impress in his workout for the Blazers on Wednesday, I'd assume it would be Pendergraph.
So what conclusions can we draw from all of this? First, we can see that there are four better than average PFs at efficiency that the Blazers should consider. Second, out of these four, Blair was the most efficient in college. If we trade up, it should be to get Blair, not Hanbrough. Third, because Hansbrough, Pendergraph, and Gibson all are close to statistically even, it makes most sense for the Blazers to stay at #24 and draft Pendergraph. This makes even more sense considering how Pendergraph seemed better than Hansbrough at his workout with the Blazers. He seems to be what you'd want in a backup PF. He's a smart, low-mistake player who most importantly contributes above average win totals to the team he plays for. Since the point of the game is to win, it makes most sense to draft the players that are best at winning.
(Note: I just want to emphasize that WS/40 does not tell us everything about a player. I'm not saying it's the only thing teams should draft on. I don't want a thousand comments telling me that stats aren't the only thing you should consider. I'm just saying that productivity tends to translate from college to the pros and should definitely be considered in our debate over the best PF for the Blazers to draft. I'd also like to once again highlight the significant contribution of this blog to this post.)
1 recs |
22 comments
Comments
James Johnson 13.3
Draft Cole Aldrich 2010
by jlarose78 on Jun 19, 2009 9:50 PM PDT reply actions 0 recs
Good eyes
Good point, I didn’t have him up there because he’s a bit of a tweener. While DaJuan Blair hauls down a monster amount of rebounds to justify putting him at the PF, Johnson is more of a F. And we’ve already got one of those in Travis Outlaw. I guess you could include him if you want. I just don’t think we need another player potentially gumming up the works in our large number of guys capable of playing the 3.
I must not fear. Fear is the mind-killer.—Dune
by Muad'Dib on Jun 19, 2009 10:00 PM PDT up reply actions 0 recs
The only reason I really like him is
Because he could probally beat up Shaq
Draft Cole Aldrich 2010
by jlarose78 on Jun 19, 2009 10:05 PM PDT up reply actions 0 recs
Fear is the mind-killer
You don’t have to be bene gesserit trained to know that DeJuan Blair is the Ma’adi
by two4larue on Jun 19, 2009 10:26 PM PDT reply actions 0 recs
Strongly agree
First I must state I am also a big reader of the Wages of Wins blog. Anyone who likes stats or being right should go read it. But anyway since that angle has been well covered, I’m going to let loose with my non-statistical analysis:
I watched a number of Pitt games, and Blair is just a beast. I agree that if we’re going to trade up, he’d give us more of what we want than Psycho T.
More importantly though, we don’t need a lottery pick. We’ve got enough of those the last few years – we just want a solid bench contributor. I like Pendergraph because he’s not much of a risk. He’s 22, 230 lbs, and 6’9" – NBA sized, with plenty of college experience. Another thing the WoW blog points out is that basketball players are fairly consistent over time, and while college to the NBA is a leap, Pendergraph will be doing the same kinds of things. He won’t be a black hole of FGA like Outlaw – the shots he takes will go in almost always.
His DraftExpress profile gives a nice positive writeup. However, he’s projected to go early second round, and the Blazers have picks 33 and 38 (if what I’m reading is right). NBADraft.com’s mock has him going to us at 38. So now that I’ve looked into this more, I’d love to get Pendergraph at 38 – maybe even 33 depending on who else is available. But it doesn’t look like we should take him at 24, because we don’t need to.
Sorry for the rambling nature of this post.
by austinpwnz on Jun 19, 2009 10:56 PM PDT reply actions 1 recs
More importantly though, we don’t need a lottery pick. We’ve got enough of those the last few years – we just want a solid bench contributor.
If you can go get that lottery pick at the cost of players that won’t be on the court (Sergio, #24, #33, #39) anyway, you go do that. 15 minutes of an B- player is better than 15 minutes of a C- player always. Blair is one of those hyper efficient type players that is going to make the most of limited minutes on the court. He’s the exact player we want our vacant backup 4 spot.
by as11osu on Jun 19, 2009 11:10 PM PDT up reply actions 0 recs
Sorry, let me clarify.
Of course getting something for nothing like that would be good. I guess what I mean is you don’t go for a “potential” player, someone coming out of college early whose strengths are up in the air. We don’t need to take a risk on a lottery player is perhaps a better way to say it. Blair is probably one of the less risky guys to trade up for (Stephen Curry is the first example that comes to mind of a risky and high-cost trade up).
by austinpwnz on Jun 19, 2009 11:58 PM PDT up reply actions 0 recs
Anyone who likes stats or being right should go read it.
You will have to change that to
Anyone who likes statsuntil Nick Fazekas pans out.
Life is exhausting when you are this stupid.
by jonestr on Jun 20, 2009 12:09 AM PDT up reply actions 0 recs
He stunk in Dallas. But his stats in the French league and Euroleague (on a team co-owned by Tony Parker) are pretty good
Maybe he could be an alternative to Joel Freeland. Don’t know if the Clippers still hold his free agent rights.
by Norsktroll on Jun 20, 2009 9:05 AM PDT up reply actions 0 recs
He stunk for Dallas?
Huh? He played a grand total of 9 minutes for Dallas that season. Look at his numbers with the Clippers. 260 minutes and a PER of 19.6. TS% = .594, TRB% = 19.1.
That’s pretty much the extent of his NBA work (not just solid, but legitimately good).
http://www.82games.com/0708/07LAC15A.HTM
http://www.82games.com/0708/07LAC15C.HTM
http://www.82games.com/0708/07LAC15D.HTM
by as11osu on Jun 20, 2009 9:16 AM PDT up reply actions 0 recs
They’re stats, not a crystal ball. He didn’t say “anyone who likes stats or being omniscient.” You can’t just find one example of a statistical analysis not working and call it wrong. You have to look at whether it gets it right on a consistent level. Sometimes it’ll be wrong because there are way too many variables to account for. But most of the time, win based analysis is correct.
I must not fear. Fear is the mind-killer.—Dune
by Muad'Dib on Jun 20, 2009 1:28 AM PDT up reply actions 0 recs
I was not attacking the work
only the idea that by reading the WoW journal you will be “correct”, if by using WoW that made you “correct” then the work would not be debatable, which it is, to debunk the idea that this work makes you “correct” you only need one counter example.
BTW I was just perusing his draft rankings from years past, and it has more than a few notable flubs in it. As someone who frequently reads WoW, I am always surprised by how he is not all that critical of his analysis even when there is gaping chasm between his analysis and reality. I much prefer Hollinger’s analysis as he will tell you what the possible warts are in his modeling.
Life is exhausting when you are this stupid.
by jonestr on Jun 20, 2009 1:49 AM PDT up reply actions 0 recs
Sorry, excuse my over-enthusiasm
I think it’s safe to say that WoW is a lot more accurate when applied purely to the NBA rather than from the drafting standpoint, and I was referring to the model and blog as a whole. The model is far more accurate in predicting team wins based on an number of individual performances than it is at predicting future individual performance. Not to mention any draft rankings or predictions have so many more factors than stats, I think his predictions can still be way off without invalidating the vast majority of his statistical analysis. The draft rankings are only a small part of WoW and really not the focus of most of the research.
Also I was kind of joking/purposely being over-the-top, so sorry if I seemed arrogant. Didn’t mean anything by it.
by austinpwnz on Jun 20, 2009 2:22 AM PDT up reply actions 0 recs
here is the problem
he doesn’t really use WoW to run his predictions for team success, in the end, he uses point differential like everybody else. A study (which I am having difficulty finding) showed that P.E.R. was the best metric for predicting wins. WoW did not do very good at all and the author of the study railed at Berri for being arrogant and not willing to fix his broken metric.
WoW is somewhat useful for identifying all-around players and good roleplayers. However, WoW makes a lot of assumptions that I don’t necessarily agree with. One, and I believe it was Timbo that asked this originally, is it better to have players who are a 5 at everything or players who are a 10 at some things? Also, Berri over values defensive boards a bit.
Anyhow, I’ve lots of criticisms of WoW, it’s useful, but not great. and Berri has his little blog where he trumpets his successes and almost never admits the shortcomings. He sees basketball through the lens of WoW instead of seeing WoW through the lens of basketball IMO.
Come on you gotta listen unto me,
lay off that whiskey and let that cocaine be. ~Johnny Cash
by HurraKane212 on Jun 20, 2009 10:05 AM PDT up reply actions 0 recs
very nice critique of WoW
what a nice little conversation that started from a remark that I meant to be silly
I love BE
Life is exhausting when you are this stupid.
by jonestr on Jun 20, 2009 10:40 AM PDT up reply actions 0 recs
Here is the problem with your problem
First, the efficiency of WoW’s metrics actually has been tested. They ran the numbers for the 2003-2004 NBA season and came up with an average error of 1.67 wins from the actual number of wins. That’s pretty darn accurate. You claim that PER is a more accurate model. The simplified version of PER is Game Score. Game score has a .99 correlation with PER. In other words, there’s no significant difference. Game score itself has a .99 correlation with NBA Efficiency. That means that PER is essentially no different from NBA Efficiency at telling us what a player does. NBA Efficiency only predicts 23% of a team’s wins. In other words, PER explains at most a quarter of a team’s win total. What PER and NBA Efficiency do explain well is player salary. But I’d rather win than pay people lots of money. Of course, if you ever actually find this study you cite, I’d be pleased to read it. It’s just that right now I’ve seen a lot of numbers that refute what you’re suggesting and you didn’t provide any numbers to back up your assertions.
I must not fear. Fear is the mind-killer.—Dune
by Muad'Dib on Jun 20, 2009 3:45 PM PDT up reply actions 0 recs
PER is a better mthod of evaluating players than the WoW metrics
The problem with WoW is that it ignores diminishing returns. For instance, it assumes that if you put a bunch of good defensive rebounders on the court together that you can just add up the number of rebounds that they normally get and that will be the number of rebounds that the team will get. The problem is that those players are taking rebounds away from each other. If you put 5 Joel Przybillas on the court together they will be a good rebounding team, but total defensive rebounding rate will still be well below 100% instead of the 150%+ that Berri’s metrics would predict.
There are also diminishing returns on scoring. Berri assumes that if you put a bunch of low-volume but high efficiency scorers on the court together that they will all be able to maintain their efficiency while increasing scoring volume. The reality is that a team like that would struggle to score and the efficiency of the players would drop dramatically.
That is why studies like this one have found that the WoW metrics do a poor job of evaluating how much a player helps his team win, and are inferior to PER. However, PER isn’t really that great either, I think statistical +/- is significantly better than either PER or the WoW metrics.
by trk on Jun 22, 2009 1:24 PM PDT up reply actions 0 recs
I'd like to see that study
Because I haven’t seen anything like that or any reason to believe that PER is better for predicting wins.
PER is messed up anyway.
http://dberri.wordpress.com/2006/11/17/a-comment-on-the-player-efficiency-rating/
The relevant sentence being:
“Given these values, with a bit of math we can show that a player will break even on his two point field goal attempts if he hits on 30.4% of these shots. On three pointers the break-even point is 21.4%.”
I was going to post a bunch more but Muad’Dib covered it all, more or less. PER is a formula that a guy created to fit his assumptions, where WS is based on actual data rather than the impressions of a single player. Plus, in terms of data, I trust someone who does econometrics for a living rather more than someone who just writes articles.
To address your specific criticisms, the win score metric compares players to the average output at that position. So if, in your case, average is a 4 at everything, is a player who’s 5 at everything better than someone who is a 10 in one thing and 1 in everything else? Oftentimes yes, since being a 1 at turnovers or a 10 in shot attempts (but not FG%) can hurt a great deal.
Also, it seems to me (and I haven’t always thought this) that d.rebs are just as important as o.rebs. Because if you don’t get a defensive rebound from somebody, that means you’re giving up an offensive rebound, which can potentially harm you just as much as getting an offensive rebound can potentially help you. That prevention is worth a lot.
Anyway, arguing on the internet and all that, but I feel you’re dismissing WoW because it doesn’t fit what you already believe. And it is interesting to talk about.
by austinpwnz on Jun 20, 2009 9:46 PM PDT up reply actions 0 recs
Again, WoW is useful
but I think there are problems with it based upon all the other metrics as well. I too, was surprised by that study (wherever it is atm) that said PER was as accurate as it is. It may be that they ran a PER differential… can’t remember.
I read a lot of APBRmetrics sites on a regular basis and I have taken statistics in my undergrad, so I’m not just spouting off based upon what I think is important in basketball.
I like d-rebounds, inasmuch as they prevent opponent O rebounds, but D rebounds can’t be weighed as heavily as O rebounds because D rebounds are easier to get. Also, O rebounds are more likely to lead to fouls drawn.
If I were to create a metric, I would heavily value ball control. O rebounds, D rebounds, Steals, Charges Drawn, etc. I would look at bad pass turnovers as a major negative, as well as “pocket picked per 40 minutes.” A player like Roy might get the ball stolen more than LMA since he has it in his hands more, but over 40 min, it would be informative to see who’s coughing the ball up most.
Blocks are hard to value because there is a psychological component to having a shotblocker inside and I don’t have the data to see how much that is worth.
I can’t wait for a healthy Greg Oden, because hopefully it will allow our guards to gamble a bit more for steals. A couple more steals per game will push us into 60 wins territory.
Come on you gotta listen unto me,
lay off that whiskey and let that cocaine be. ~Johnny Cash
by HurraKane212 on Jun 21, 2009 10:36 AM PDT up reply actions 0 recs
Well it makes sense there are problems with it
“based on all the other metrics”…since they disagree.
Your description of a metric pretty much sums up what I like about WoW and what other metrics are undervaluing, as far as I can tell. And it is pace-adjusted and per-minute…
I still say defensive rebounds are equal to offensive rebounds simply out of prevention. The marginal benefit of a defensive rebound (preventing the other team’s o.reb) is the same as getting an offensive rebound when you look at the net benefit versus the other team.
I think the psychological component of a shotblocker would manifest itself in the other team taking worse shots, and thus a lower shooting percentage (or getting less free throws).
Charges drawn would be a nice thing to throw in, but it comes up so seldom it would be like comparing the blocked shots of PGs. And Roy gets more assists (and presumably better shots) than LaMarcus due to increased possessions.
Also, GO’s stats per 36 minutes are pretty sweet (aside from fouls), check basketball-reference.com. When he’s healthy and can stay on the court, his draft position will be vindicated.
But the difference between WP48 and other metrics is that WP48 was created to explain experimental data, whereas PER etc were created based on the preconceived notions someone has about the value of certain statistics. WP48 doesn’t explain WHY, but it is mathematically based off of years of NBA stats.
by austinpwnz on Jun 21, 2009 1:48 PM PDT up reply actions 0 recs
Taj Gibson
I like him to eventually play a Tyrus Thomas-type role & backup both Forward positions.
Offseason:
PG Options: Terrence Williams(D), Rodrigue Beaubois(D)
SG Options: Mickaël Piétrus(T)
PF Options: Ronnie Turiaf(T) - Taj Gibson(D), Kevin Seraphin(D)
C Options: Alexis Ajinca(T)
The French Invasion !
by TheGreatDane17 on Jun 20, 2009 12:22 AM PDT reply actions 0 recs

by 




















