PDA

View Full Version : ESPN's take on value of a draft pick...



docpaul
06-03-2009, 04:21 PM
I really enjoyed this article, and found it a different take on Count55's work, as it focused on the growing popularity of "win averages".


Many of you might not have access to insider, so here goes:


http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222914



When the NBA draft started more than 50 years ago, it was like a bare-knuckles fight -- there was no limit on the number of rounds, allowing teams to draft players for as long as they could stand, including a record 21 rounds in 1960.
Times change. Since then, the draft has been shrunk (to 10 rounds by the mid-1970s) and downsized (to seven in 1985) and desiccated (to the current two rounds in 1989). Obviously, drafts can change the course of league history (see: Magic Johnson (http://sports.espn.go.com/nba/players/profile?playerId=2334), Michael Jordan (http://sports.espn.go.com/nba/players/profile?playerId=1035)), and every year on draft night, we like to point and stare at the big suits (http://4.bp.blogspot.com/_8tfveb4LEvY/SGWhbtISpYI/AAAAAAAAAKQ/a7yKX-llcHg/s400/gordon.jpg) and the Big Suit (http://www.chinadaily.com.cn/english/doc/2004-11/23/xin_31110123084947115761.jpg).
But Insider's D.R.A.F.T. Initiative (Data-Related Analysis For Truth) crew has been poring over draft data for months, analyzing draft picks (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4206291) and their subsequent careers, and we've arrived at a surprising conclusion: The NBA draft isn't that big a deal. That's because, in any given year, there isn't enough talent to give many teams any hope of landing a star, let alone a reliable backup.
To get deeper into the ideas of the D.R.A.F.T Initiative (educating yourself on value and methodology in the process), please sign up for ESPN Insider. http://assets.espn.go.com/i/in.gif There isn't a Mike Piazza or a Tom Brady lurking in the late rounds, because there are no late rounds. Crunch some numbers, as ESPN researcher Tom Haberstroh did, and the draft could easily do with more downsizing. Beyond the first five picks, the quality falls off rapidly. Beyond the first 10, the selection process is a proverbial crapshoot. Actually, it's not that proverbial; teams drafting after the fifth pick are quite likely to pick a crap player and look back on it while using closely related linguistic variants of the word "shoot."
Check out this graph from our pick-based analysis page (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4206291):
http://assets.espn.go.com/i/mag/blog/2009/0602_graph.jpgThe squiggly blue line plots EWA (http://insider.espn.go.com/nba/insider/news/story?id=4222771), John Hollinger's estimated wins added (http://sports.espn.go.com/nba/columns/story?columnist=hollinger_john&page=PERDiem-090325) stat, for each draft slot, averaged over the past 20 years of draft picks. EWA is a VORP (http://sports.espn.go.com/mlb/hotstove06/columns/story?columnist=neyer_rob&id=2751842)-like metric that measures a player's contributions to his team compared to those of a baseline replacement player. Built off Hollinger's player efficiency rating (http://sports.espn.go.com/nba/columns/story?columnist=hollinger_john&id=2850240), EWA considers not only the player's efficiency but also how many minutes that player spends on the court. The red line is the best-fit curve that predicts a given slot's value based on the downward-trending distribution of the data.
So what does the picture show? Draft early, and don't draft often. Maybe, if you're the Clippers, don't draft at all.
And yet ...
There are exceptions, anomalies, funky trends. The draft is cruel to most GMs, but it's been kind to a few smart clubs (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4221335). The Spurs have mined diamonds both early (Tim Duncan (http://sports.espn.go.com/nba/players/profile?playerId=215)) and late (Tony Parker (http://sports.espn.go.com/nba/players/profile?playerId=1015)). In general, picks six through eight have been awful, yet No. 9 has been revolutionary. While there's a strong case to eliminate the second round altogether -- hey, it would be in keeping with the NBA's history -- there's also the lure of a Rashard Lewis (http://sports.espn.go.com/nba/players/profile?playerId=469) or a Gilbert Arenas (http://sports.espn.go.com/nba/players/profile?playerId=974) still on the board while the janitors are vacuuming Cheez Doodles off the green room floor.
In other words, since the draft as a whole isn't the talent show it's made out to be, it's even more important for teams to match up their roster needs with the players who are available -- and to realize when those needs simply can't be met. Or, they need to recalibrate their expectations. Instead of hoping the supposed "best player available" will max out all his skills, when a team is drafting at a point likely to generate a role player, it would be wise to target a prospect with one clear, NBA-ready ability.
http://a.espncdn.com/photo/2009/0601/nba_g_cabarkapa01_200.jpg (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222914#)<cite>Rocky Widner/NBAE/Getty Images</cite>It's more likely that your team will wind up drafting someone like Cabarkapa than you'd realize.


How to do that? Glad you asked. Over the next month, we'll explain our method in more detail and analyze expected value in a potential draftee. We'll comb through our research to identify and explain trends in the draft, show the difference between teams that succeeded and those that failed, and the success rates of players taken in different draft slots.
We'll also analyze players who worked out well for their teams, well beyond their predraft workouts, compare them with players who went bust, and compare both groups with the current crop of draft candidates. We'll talk about high-school players versus internationals versus collegians, and show that while a little knowledge (a year of college) may be a dangerous thing, it's a lot more dangerous to draft a guy with more knowledge (three or more years in school).
We'll explore the past two decades' sleepers and busts, and why some clubs wind up with Darkos (Milicic (http://sports.espn.go.com/nba/players/profile?playerId=2171)), Zarkos (Cabarkapa (http://sports.espn.go.com/nba/players/stats?playerId=1997)) and Markos (Jaric (http://sports.espn.go.com/nba/players/profile?playerId=387)), while others manufacture Manus (Ginobili (http://sports.espn.go.com/nba/players/profile?playerId=0272)).
We fervently hope, too, to explain the statistical insignificance of first-round draft choices whose last names end in "I."
Mostly, we'll give you an idea of what to expect from the draft, given your team's slot. More important, we'll let you know when, and why, you shouldn't expect too much.
Luke Cyphers is a senior writer for ESPN The Magazine and ESPN Insider.

docpaul
06-03-2009, 04:23 PM
Focusing more specifically towards the Pacers:

http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4221335

Which teams have drafted best over the past 20 years? And which are the worst draft-day franchises? To figure that out, you can't just compile a list of names. Even the Warriors can trot out some gems, which is what happens when you end up in the lottery almost every season. But just because they've nabbed Antawn Jamison (http://sports.espn.go.com/nba/players/profile?playerId=385), Chris Webber (http://sports.espn.go.com/nba/players/profile?playerId=897), Jason Richardson (http://sports.espn.go.com/nba/players/profile?playerId=1018) and Andris Biedrins (http://sports.espn.go.com/nba/players/profile?playerId=2427), it doesn't mean they're excellent at evaluating talent.
Instead, to measure a team's success in the draft, you have to rate its picks against the expected value of a given slot. For instance, Andrew Bogut (http://sports.espn.go.com/nba/players/profile?playerId=2747) is a fine player with a career EWA of 5.4. That's a lottery-worthy number. But it's not good enough to be worthy of the No. 1 pick (expected EWA of 7.8), which means the Bucks cost themselves some wins.
With that in mind, we studied every pick over the past 20 years for each franchise and came up with a total number of wins above or below expectations. We've also presented that information on a per-pick basis, since the Clippers obviously have had more opportunities to draft than the Bobcats.
How does that relate to the ultimate letter grades? On average, teams actually draft exceptionally close to expected levels. The average team over the past 20 years has picked a player with an EWA that is just 0.1 below the expected rating at a given spot. The average grade, then, becomes a "C" which is where the Rockets landed in the chart. Every tenth of a point (plus or minus) away from 0.1 resulted in a grade change (from C to C+, B+ to A-, C- to D+, etc.). The Spurs, whose rating is off the charts, get their A+ with a special gold star. And the Clippers? Even a perfect draft this year wouldn't give them a passing grade.
For each team's detailed draft history, just click on the linked name.
[NOTE: For a draft-day swap (such as Dirk Nowitzki (http://sports.espn.go.com/nba/players/profile?playerId=609) for Robert Traylor (http://sports.espn.go.com/nba/players/profile?playerId=857)), the team ending up with the player is credited with the selection. So, in that case, Dallas gets Nowitzki and Milwaukee gets Traylor.]
Team-by-team analysis

<table><thead> <tr><th>Rank</th> <th>Team</th> <th>Grade</th> <th>Wins above expected</th> <th>Wins per pick</th> <th>2009 Draft Pick</th> </tr> </thead> <tbody> <tr class="last"> <td>1</td> <td>Spurs (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223674)</td> <td>A+</td> <td>22.1</td> <td>0.85</td> <td>25</td> </tr> <tr class="last"> <td>2</td> <td>Lakers (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222499)</td> <td>A+</td> <td>27.2</td> <td>0.76</td> <td>29</td> </tr> <tr class="last"> <td>3</td> <td>Suns (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223677)</td> <td>A+</td> <td>28.8</td> <td>0.67</td> <td>14</td> </tr> <tr class="last"> <td>4</td> <td>Raptors (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223671)</td> <td>A</td> <td>15.3</td> <td>0.44</td> <td>9</td> </tr> <tr class="last"> <td>5</td> <td>Cavaliers (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222402)</td> <td>A-</td> <td>15.3</td> <td>0.44</td> <td>30</td> </tr> <tr class="last"> <td>6</td> <td>Celtics (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222406)</td> <td>B+</td> <td>13.5</td> <td>0.33</td> <td>28</td> </tr> <tr class="last"> <td>7</td> <td>Thunder (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223679)</td> <td>B-</td> <td>11.1</td> <td>0.24</td> <td>3</td> </tr> <tr class="last"> <td>8</td> <td>Hornets (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222446)</td> <td>B-</td> <td>5.7</td> <td>0.18</td> <td>21</td> </tr> <tr class="last"> <td>9</td> <td>76ers (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222225)</td> <td>B-</td> <td>6.3</td> <td>0.14</td> <td>17</td> </tr> <tr class="last"> <td>10</td> <td>Jazz (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222450)</td> <td>C+</td> <td>2.5</td> <td>0.07</td> <td>20</td> </tr> <tr class="last"> <td>11</td> <td>Warriors (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223683)</td> <td>C+</td> <td>2.9</td> <td>0.06</td> <td>7</td> </tr> <tr class="last"> <td>12</td> <td>Heat (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222440)</td> <td>C+</td> <td>0.2</td> <td>0.01</td> <td>18</td> </tr> <tr class="last"> <td>13</td> <td>Pistons (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223659)</td> <td>C</td> <td>-0.6</td> <td>-0.2</td> <td>15</td> </tr> <tr class="last"> <td>14</td> <td>Blazers (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223682)</td> <td>C</td> <td>-3.2</td> <td>-0.07</td> <td>24</td> </tr> <tr class="last"> <td>15</td> <td>Grizzlies (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222434)</td> <td>C</td> <td>-2.5</td> <td>-0.07</td> <td>2</td> </tr> <tr class="last"> <td>16</td> <td>Rockets (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223676)</td> <td>C</td> <td>-3.8</td> <td>-0.08</td> <td>23</td> </tr> <tr class="last"> <td>17</td> <td>Bucks (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222230)</td> <td>C-</td> <td>-4.3</td> <td>-0.11</td> <td>10</td> </tr> <tr class="last"> <td>18</td> <td>Magic (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222502)</td> <td>C-</td> <td>-4.7</td> <td>-0.12</td> <td>27</td> </tr> <tr class="last"> <td>19</td> <td>Timberwolves (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223681)</td> <td>C-</td> <td>-5.3</td> <td>-0.13</td> <td>6</td> </tr> <tr class="last"> <td>20</td> <td>Mavericks (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222520)</td> <td>D+</td> <td>-9.5</td> <td>-0.22</td> <td>22</td> </tr> <tr class="last"> <td>21</td> <td>Kings (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222486)</td> <td>D</td> <td>-13.0</td> <td>-0.32</td> <td>4</td> </tr> <tr class="last"> <td>22</td> <td>Pacers (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223657)</td> <td>D</td> <td>-11.9</td> <td>-0.35</td> <td>13</td> </tr> <tr class="last"> <td>23</td> <td>Bulls (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222297)</td> <td>D-</td> <td>-24.4</td> <td>-0.49</td> <td>16</td> </tr> <tr class="last"> <td>24</td> <td>Knicks (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222493)</td> <td>F</td> <td>-16.8</td> <td>-0.53</td> <td>8</td> </tr> <tr class="last"> <td>25</td> <td>Nuggets (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223641)</td> <td>F</td> <td>-22.6</td> <td>-0.54</td> <td>26</td> </tr> <tr class="last"> <td>26</td> <td>Nets (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222523)</td> <td>F</td> <td>-21.6</td> <td>-0.62</td> <td>11</td> </tr> <tr class="last"> <td>27</td> <td>Wizards (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223686)</td> <td>F</td> <td>-22.8</td> <td>-0.63</td> <td>5</td> </tr> <tr class="last"> <td>28</td> <td>Hawks (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4221754)</td> <td>F</td> <td>-30.2</td> <td>-0.66</td> <td>19</td> </tr> <tr class="last"> <td>29</td> <td>Bobcats (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222226)</td> <td>F</td> <td>-11.8</td> <td>-1.18</td> <td>12</td> </tr> <tr class="last"> <td>30</td> <td>Clippers (http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4222409)</td> <td>F</td> <td>-52.0</td> <td>-1.18</td> <td>1</td></tr></tbody></table>

docpaul
06-03-2009, 04:27 PM
...and most specifically... an analysis of all Pacer draft picks. I think the flaws of the EWA model of analysis begin to make themselves apparent.

For example, it makes poor assumptions based on limited data (ie, Brandon Rush).

http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223657

See the chart below for a detailed, pick-by-pick analysis of the Pacers' drafts over the last 20 years. The EWA column shows the player's actual production, while the eEWA column shows the expected value of that draft slot. The "net" column is the difference between the two. For more analysis of their picks, click here (http://insider.espn.go.com/nbadraft/draft/tracker/team?draftyear=2008&team=ind).
<table><thead> <tr class="last"><th>Draft Year</th><th>Pick</th><th>Player</th><th>EWA</th><th>eEWA</th><th>NET</th></tr> </thead> <tbody><tr class="last"><td>2008</td><td>17</td><td>Roy Hibbert (http://sports.espn.go.com/nba/players/profile?playerId=3436)</td><td>2.56</td><td>2.00</td><td>0.56</td></tr> <tr class="last"><td>2008</td><td>13</td><td>Brandon Rush (http://sports.espn.go.com/nba/players/profile?playerId=3457)</td><td>-1.79</td><td>2.60</td><td>-4.39</td></tr> <tr class="last"><td>2007</td><td>39</td><td>Stanko Barac</td><td>0.00</td><td>0.40</td><td>-0.40</td></tr> <tr class="last"><td>2006</td><td>31</td><td>James White (http://sports.espn.go.com/nba/players/profile?playerId=3037)</td><td>0.03</td><td>0.80</td><td>-0.77</td></tr> <tr class="last"><td>2006</td><td>17</td><td>Shawne Williams (http://sports.espn.go.com/nba/players/profile?playerId=3038)</td><td>0.07</td><td>2.00</td><td>-1.93</td></tr> <tr class="last"><td>2005</td><td>46</td><td>Erazem Lorbek</td><td>0.00</td><td>0.00</td><td>0.00</td></tr> <tr class="last"><td>2005</td><td>17</td><td>Danny Granger (http://sports.espn.go.com/nba/players/profile?playerId=2760)</td><td>7.11</td><td>2.00</td><td>5.11</td></tr> <tr class="last"><td>2004</td><td>59</td><td>Rashad Wright</td><td>0.00</td><td>-0.50</td><td>0.50</td></tr> <tr class="last"><td>2004</td><td>29</td><td>David Harrison (http://sports.espn.go.com/nba/players/profile?playerId=2383)</td><td>-0.02</td><td>1.00</td><td>-1.02</td></tr> <tr class="last"><td>2003</td><td>49</td><td>James Jones (http://sports.espn.go.com/nba/players/profile?playerId=2009)</td><td>0.37</td><td>-0.10</td><td>0.47</td></tr> <tr class="last"><td>2002</td><td>14</td><td>Fred Jones (http://sports.espn.go.com/nba/players/profile?playerId=1718)</td><td>0.24</td><td>2.40</td><td>-2.16</td></tr> <tr class="last"><td>2001</td><td>40</td><td>Jamison Brewer (http://sports.espn.go.com/nba/players/profile?playerId=980)</td><td>-0.12</td><td>0.30</td><td>-0.42</td></tr> <tr class="last"><td>2000</td><td>27</td><td>Primoz Brezec (http://sports.espn.go.com/nba/players/profile?playerId=94)</td><td>0.93</td><td>1.10</td><td>-0.17</td></tr> <tr class="last"><td>2000</td><td>56</td><td>Jaquay Walls</td><td>0.00</td><td>-0.40</td><td>0.40</td></tr> <tr class="last"><td>1999</td><td>26</td><td>Vonteego Cummings (http://sports.espn.go.com/nba/players/profile?playerId=172)</td><td>0.14</td><td>1.20</td><td>-1.06</td></tr> <tr class="last"><td>1999</td><td>5</td><td>Jonathan Bender (http://sports.espn.go.com/nba/players/profile?playerId=50)</td><td>-0.01</td><td>4.50</td><td>-4.51</td></tr> <tr class="last"><td>1998</td><td>25</td><td>Al Harrington (http://sports.espn.go.com/nba/players/profile?playerId=308)</td><td>3.15</td><td>1.30</td><td>1.85</td></tr> <tr class="last"><td>1997</td><td>12</td><td>Austin Croshere (http://sports.espn.go.com/nba/players/profile?playerId=167)</td><td>1.58</td><td>2.80</td><td>-1.22</td></tr> <tr class="last"><td>1996</td><td>52</td><td>Mark Pope (http://sports.espn.go.com/nba/players/profile?playerId=667)</td><td>-0.29</td><td>-0.20</td><td>-0.09</td></tr> <tr class="last"><td>1996</td><td>10</td><td>Erick Dampier (http://sports.espn.go.com/nba/players/profile?playerId=181)</td><td>2.91</td><td>3.10</td><td>-0.19</td></tr> <tr class="last"><td>1995</td><td>52</td><td>Fred Hoiberg (http://sports.espn.go.com/nba/players/profile?playerId=339)</td><td>0.96</td><td>-0.20</td><td>1.16</td></tr> <tr class="last"><td>1995</td><td>23</td><td>Travis Best (http://sports.espn.go.com/nba/players/profile?playerId=58)</td><td>1.50</td><td>1.40</td><td>0.10</td></tr> <tr class="last"><td>1994</td><td>41</td><td>William Njoku</td><td>0.00</td><td>0.30</td><td>-0.30</td></tr> <tr class="last"><td>1994</td><td>44</td><td>Damon Bailey</td><td>0.00</td><td>0.10</td><td>-0.10</td></tr> <tr class="last"><td>1994</td><td>15</td><td>Eric Piatkowski (http://sports.espn.go.com/nba/players/profile?playerId=660)</td><td>1.45</td><td>2.30</td><td>-0.85</td></tr> <tr class="last"><td>1993</td><td>39</td><td>Thomas Hill</td><td>0.00</td><td>0.40</td><td>-0.40</td></tr> <tr class="last"><td>1993</td><td>51</td><td>Spencer Dunkley</td><td>0.00</td><td>-0.20</td><td>0.20</td></tr> <tr class="last"><td>1993</td><td>14</td><td>Scott Haskin (http://sports.espn.go.com/nba/players/profile?playerId=1681)</td><td>0.00</td><td>2.40</td><td>-2.40</td></tr> <tr class="last"><td>1992</td><td>14</td><td>Malik Sealy (http://sports.espn.go.com/nba/players/profile?playerId=763)</td><td>1.00</td><td>2.40</td><td>-1.40</td></tr> <tr class="last"><td>1991</td><td>13</td><td>Dale Davis (http://sports.espn.go.com/nba/players/profile?playerId=187)</td><td>4.17</td><td>2.60</td><td>1.57</td></tr> <tr class="last"><td>1991</td><td>41</td><td>Sean Green (http://sports.espn.go.com/nba/players/profile?playerId=3722)</td><td>-0.03</td><td>0.30</td><td>-0.33</td></tr> <tr class="last"><td>1990</td><td>45</td><td>Antonio Davis (http://sports.espn.go.com/nba/players/profile?playerId=189)</td><td>2.66</td><td>0.10</td><td>2.56</td></tr> <tr class="last"><td>1990</td><td>46</td><td>Kenny Williams (http://sports.espn.go.com/nba/players/profile?playerId=3538)</td><td>0.50</td><td>0.00</td><td>0.50</td></tr> <tr class="last"><td>1989</td><td>7</td><td>George McCloud (http://sports.espn.go.com/nba/players/profile?playerId=524)</td><td>1.01</td><td>3.80</td><td>-2.79</td></tr></tbody></table>

owl
06-03-2009, 04:39 PM
The only thing I really got out of that is the following quote from the article above.

"Instead of hoping the supposed "best player available" will max out all his skills, when a team is drafting at a point likely to generate a role player, it would be wise to target a prospect with one clear, NBA-ready ability."

I believe that should really be an important component of the Pacers at 13.

count55
06-03-2009, 04:46 PM
Yeah, it's a good read.

On the Pacers: This Page has the breakdown:

http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223657

We're being dragged down by Bender, Fred Jones, Scott Haskin, and George McCloud...all of which make sense. However, our second worse pick...according to this...is Brandon Rush at -4.39. I'd be curious to see what his numbers looked like, Pre and Post All-Star break.

Our best pick (since 1989) was...duh...Danny Granger at +5.11. Rounding out the top 5 were Antonio, Baby Al, Dale, and Fred Hoiberg. (It's important to keep in mind that they're comparing against the expectation for the pick.)

Some notes:

Roy Hibbert was our 6th best pick, at +0.56.

21 of the 34 picks were below average for the slot, 12 were above average, and one (Lorbek, believe it or not) was exactly on average.

I'd love to have access to the type and volume of data they have.

ChicagoJ
06-03-2009, 05:08 PM
The thing I do like about this is that it shows you how little to expect from a pick anywhere below about #15 in a bad year to may #18 in a good draft. As in: nearly nothing.

rexnom
06-03-2009, 05:10 PM
I don't know - I don't think this is a particular useful metric. A lot of things seem off - like Croshere being a worse pick than Brezec - and adjustments need to be made.

For example, shouldn't the eEWA change every year based on the average EWA of the class? Or the average EWA of the picks made afterwards or the players eligible. Since the 2005 draft was stronger than the 2006 draft (even after 17), I think Danny's eEWA should be higher than Williams's. Also, player pools have changed (now we have more international players but no high school players) which changes the eEWA considerably based on each year.

Los Angeles
06-03-2009, 05:15 PM
...and most specifically... an analysis of all Pacer draft picks. I think the flaws of the EWA model of analysis begin to make themselves apparent.

For example, it makes poor assumptions based on limited data (ie, Brandon Rush).

http://insider.espn.go.com/nba/draft2009/insider/news/story?id=4223657

See the chart below for a detailed, pick-by-pick analysis of the Pacers' drafts over the last 20 years. The EWA column shows the player's actual production, while the eEWA column shows the expected value of that draft slot. The "net" column is the difference between the two. For more analysis of their picks, click here (http://insider.espn.go.com/nbadraft/draft/tracker/team?draftyear=2008&team=ind).
<table><thead> <tr class="last"><th>Draft Year</th><th>Pick</th><th>Player</th><th>EWA</th><th>eEWA</th><th>NET</th></tr> </thead> <tbody><tr class="last"><td>2008</td><td>17</td><td>Roy Hibbert (http://sports.espn.go.com/nba/players/profile?playerId=3436)</td><td>2.56</td><td>2.00</td><td>0.56</td></tr> <tr class="last"><td>2008</td><td>13</td><td>Brandon Rush (http://sports.espn.go.com/nba/players/profile?playerId=3457)</td><td>-1.79</td><td>2.60</td><td>-4.39</td></tr> <tr class="last"><td>2007</td><td>39</td><td>Stanko Barac</td><td>0.00</td><td>0.40</td><td>-0.40</td></tr> <tr class="last"><td>2006</td><td>31</td><td>James White (http://sports.espn.go.com/nba/players/profile?playerId=3037)</td><td>0.03</td><td>0.80</td><td>-0.77</td></tr> <tr class="last"><td>2006</td><td>17</td><td>Shawne Williams (http://sports.espn.go.com/nba/players/profile?playerId=3038)</td><td>0.07</td><td>2.00</td><td>-1.93</td></tr> <tr class="last"><td>2005</td><td>46</td><td>Erazem Lorbek</td><td>0.00</td><td>0.00</td><td>0.00</td></tr> <tr class="last"><td>2005</td><td>17</td><td>Danny Granger (http://sports.espn.go.com/nba/players/profile?playerId=2760)</td><td>7.11</td><td>2.00</td><td>5.11</td></tr> <tr class="last"><td>2004</td><td>59</td><td>Rashad Wright</td><td>0.00</td><td>-0.50</td><td>0.50</td></tr> <tr class="last"><td>2004</td><td>29</td><td>David Harrison (http://sports.espn.go.com/nba/players/profile?playerId=2383)</td><td>-0.02</td><td>1.00</td><td>-1.02</td></tr> <tr class="last"><td>2003</td><td>49</td><td>James Jones (http://sports.espn.go.com/nba/players/profile?playerId=2009)</td><td>0.37</td><td>-0.10</td><td>0.47</td></tr> <tr class="last"><td>2002</td><td>14</td><td>Fred Jones (http://sports.espn.go.com/nba/players/profile?playerId=1718)</td><td>0.24</td><td>2.40</td><td>-2.16</td></tr> <tr class="last"><td>2001</td><td>40</td><td>Jamison Brewer (http://sports.espn.go.com/nba/players/profile?playerId=980)</td><td>-0.12</td><td>0.30</td><td>-0.42</td></tr> <tr class="last"><td>2000</td><td>27</td><td>Primoz Brezec (http://sports.espn.go.com/nba/players/profile?playerId=94)</td><td>0.93</td><td>1.10</td><td>-0.17</td></tr> <tr class="last"><td>2000</td><td>56</td><td>Jaquay Walls</td><td>0.00</td><td>-0.40</td><td>0.40</td></tr> <tr class="last"><td>1999</td><td>26</td><td>Vonteego Cummings (http://sports.espn.go.com/nba/players/profile?playerId=172)</td><td>0.14</td><td>1.20</td><td>-1.06</td></tr> <tr class="last"><td>1999</td><td>5</td><td>Jonathan Bender (http://sports.espn.go.com/nba/players/profile?playerId=50)</td><td>-0.01</td><td>4.50</td><td>-4.51</td></tr> <tr class="last"><td>1998</td><td>25</td><td>Al Harrington (http://sports.espn.go.com/nba/players/profile?playerId=308)</td><td>3.15</td><td>1.30</td><td>1.85</td></tr> <tr class="last"><td>1997</td><td>12</td><td>Austin Croshere (http://sports.espn.go.com/nba/players/profile?playerId=167)</td><td>1.58</td><td>2.80</td><td>-1.22</td></tr> <tr class="last"><td>1996</td><td>52</td><td>Mark Pope (http://sports.espn.go.com/nba/players/profile?playerId=667)</td><td>-0.29</td><td>-0.20</td><td>-0.09</td></tr> <tr class="last"><td>1996</td><td>10</td><td>Erick Dampier (http://sports.espn.go.com/nba/players/profile?playerId=181)</td><td>2.91</td><td>3.10</td><td>-0.19</td></tr> <tr class="last"><td>1995</td><td>52</td><td>Fred Hoiberg (http://sports.espn.go.com/nba/players/profile?playerId=339)</td><td>0.96</td><td>-0.20</td><td>1.16</td></tr> <tr class="last"><td>1995</td><td>23</td><td>Travis Best (http://sports.espn.go.com/nba/players/profile?playerId=58)</td><td>1.50</td><td>1.40</td><td>0.10</td></tr> <tr class="last"><td>1994</td><td>41</td><td>William Njoku</td><td>0.00</td><td>0.30</td><td>-0.30</td></tr> <tr class="last"><td>1994</td><td>44</td><td>Damon Bailey</td><td>0.00</td><td>0.10</td><td>-0.10</td></tr> <tr class="last"><td>1994</td><td>15</td><td>Eric Piatkowski (http://sports.espn.go.com/nba/players/profile?playerId=660)</td><td>1.45</td><td>2.30</td><td>-0.85</td></tr> <tr class="last"><td>1993</td><td>39</td><td>Thomas Hill</td><td>0.00</td><td>0.40</td><td>-0.40</td></tr> <tr class="last"><td>1993</td><td>51</td><td>Spencer Dunkley</td><td>0.00</td><td>-0.20</td><td>0.20</td></tr> <tr class="last"><td>1993</td><td>14</td><td>Scott Haskin (http://sports.espn.go.com/nba/players/profile?playerId=1681)</td><td>0.00</td><td>2.40</td><td>-2.40</td></tr> <tr class="last"><td>1992</td><td>14</td><td>Malik Sealy (http://sports.espn.go.com/nba/players/profile?playerId=763)</td><td>1.00</td><td>2.40</td><td>-1.40</td></tr> <tr class="last"><td>1991</td><td>13</td><td>Dale Davis (http://sports.espn.go.com/nba/players/profile?playerId=187)</td><td>4.17</td><td>2.60</td><td>1.57</td></tr> <tr class="last"><td>1991</td><td>41</td><td>Sean Green (http://sports.espn.go.com/nba/players/profile?playerId=3722)</td><td>-0.03</td><td>0.30</td><td>-0.33</td></tr> <tr class="last"><td>1990</td><td>45</td><td>Antonio Davis (http://sports.espn.go.com/nba/players/profile?playerId=189)</td><td>2.66</td><td>0.10</td><td>2.56</td></tr> <tr class="last"><td>1990</td><td>46</td><td>Kenny Williams (http://sports.espn.go.com/nba/players/profile?playerId=3538)</td><td>0.50</td><td>0.00</td><td>0.50</td></tr> <tr class="last"><td>1989</td><td>7</td><td>George McCloud (http://sports.espn.go.com/nba/players/profile?playerId=524)</td><td>1.01</td><td>3.80</td><td>-2.79</td></tr></tbody></table>
So Rush is almost as bad of a pick as Bender?

Something doesn't quite smell right here.

count55
06-03-2009, 05:29 PM
I don't know - I don't think this is a particular useful metric. A lot of things seem off - like Croshere being a worse pick than Brezec, for example. Shouldn't the eEWA change every year based on the average EWA of the class?

Maybe, maybe not.

The purpose is first to identify what the value of each individual draft position is, then measure the player/team against the others taken at that spot.

The thing to look at is whether it's directionally correct or not. As with the analysis I did, there will always be outliers. It's also important to remember what they're saying.

In my analysis, Austin Croshere ended up ranking 14th out of 27 #12 picks, while Brezec ended up, coincidentally, 14th out of 27 #27 picks. Both were slightly below average produced for the slot they were picked. Therefore, I could argue that, as a pick, Brezec was just as good as Austin.

However, that's different than saying that Brezec was as good or better than Croshere, which I don't even consider arguable. For all of the disappointment surrounding Austin, there's no question in my mind that he was a much better player than Primos.

That's demonstrated in this analysis as well, with Austin showing a 1.58 vs. Primos' 0.93.

(BTW...one statistical oddity is that they show guys like Piatkowski and Vonteego Cummings as draftees, despite the fact that we drafted them for other teams. I also wonder how Bayless got a better rating than Rush, despite having a lower PER...I've got to read up on EWA.)

d_c
06-03-2009, 05:33 PM
I also wonder how Bayless got a better rating than Rush, despite having a lower PER...I've got to read up on EWA.)

I don't know the full of it, but EWA has something to do with how many wins a particular player "added" to a team.

Though Bayless didn't do anything to help win any games for Portland, the fact is that Portland improved greatly in the win column last season (and Bayless was on the roster), while the Pacers didn't improve at all. It's probably reflected in that stat.

count55
06-03-2009, 05:34 PM
So Rush is almost as bad of a pick as Bender?

Something doesn't quite smell right here.

Limited data on Rush.

Also...look at the graph...Kobe was taken at 13...so were productive players like Dale Davis, Richard Jefferson, Jalen Rose, and Corey Maggette.

Bizarrely, the #13 pick peaks up and above the #5 pick on the EWA...I really have to understand the EWA...does it factor in the team wins? If so, was Bayless (my earlier question) helped by the fact that Portland went up by 13 wins (and won 54 games) while the Pacers stayed flat (at only 36 games)?

CableKC
06-03-2009, 05:39 PM
I think that the ESPN Insiders stole the idea of this article from count55.

rexnom
06-03-2009, 05:40 PM
Maybe, maybe not.

The purpose is first to identify what the value of each individual draft position is, then measure the player/team against the others taken at that spot.

The thing to look at is whether it's directionally correct or not. As with the analysis I did, there will always be outliers. It's also important to remember what they're saying.

In my analysis, Austin Croshere ended up ranking 14th out of 27 #12 picks, while Brezec ended up, coincidentally, 14th out of 27 #27 picks. Both were slightly below average produced for the slot they were picked. Therefore, I could argue that, as a pick, Brezec was just as good as Austin.

However, that's different than saying that Brezec was as good or better than Croshere, which I don't even consider arguable. For all of the disappointment surrounding Austin, there's no question in my mind that he was a much better player than Primos.

That's demonstrated in this analysis as well, with Austin showing a 1.58 vs. Primos' 0.93.

(BTW...one statistical oddity is that they show guys like Piatkowski and Vonteego Cummings as draftees, despite the fact that we drafted them for other teams. I also wonder how Bayless got a better rating than Rush, despite having a lower PER...I've got to read up on EWA.)
Well, forgetting the Austin thing for a second, my main problem with it is that you cannot say that pick x was a bad pick based on other years is fallacious. How can they say that Andrew Bogut was a poor choice in 2005 based on other years? What if there had been some sort of strike with college players and no one else declared and only international players were available (including Bogut). Would he still be a below average pick? They need to have some sort of variable to account for draft strength. Judging by their explanation, it seems that eEWA are simply points on a best fit line, which precludes the variable I mentioned.

ChicagoJ
06-03-2009, 05:42 PM
If the idea is that the only difference in a team's record is the attributed to the rookie, that is a major flaw.

Corey Benjamin must be one of the worst draft picks ever. After all, he single-handedly dropped the Bulls from being a 62-20 team all the way to a 13-37 (lockout year) team. It was clearly all his fault. It couldn't have a thing to do with Jordan retiring and Pippen getting traded, could it?

rexnom
06-03-2009, 05:43 PM
I don't know the full of it, but EWA has something to do with how many wins a particular player "added" to a team.

Though Bayless didn't do anything to help win any games for Portland, the fact is that Portland improved greatly in the win column last season (and Bayless was on the roster), while the Pacers didn't improve at all. It's probably reflected in that stat.
Well, that's another wrinkle that's awful. What if the Pacers traded their best player for draft picks and expiring contracts? How do they control for other variables?

EDIT: Jay said it better.

count55
06-03-2009, 05:48 PM
Well, forgetting the Austin thing for a second, my main problem with it is that you cannot say that pick x was a bad pick based on other years is fallacious. How can they say that Andrew Bogut was a poor choice in 2005 based on other years? What if there had been some sort of strike with college players and no one else declared and only international players were available (including Bogut). Would he still be a below average pick? They need to have some sort of variable to account for draft strength. Judging by their explanation, it seems that eEWA are simply points on a best fit line, which precludes the variable I mentioned.

Well, where they get into trouble is the grades. Once they do that, they are implying the quality of the job being done by the front office.

What this information really shows you is what benefit the Pacers have gotten from the draft, which is to say, not a lot.

However, you need the other information that you're asking for to determine whether they should have gotten more.

Regarding your Bogut example, Bogut is, in fact, a weak #1 pick, but that doesn't necessarily mean that he was the wrong guy to take in that draft. However, even looking at 2005 by itself, hindsight says that Bogut should have gone no better than third, possibly as low as fifth or sixth (CP3, Deron, and arguably Danny, Bynum, and Monta Ellis). However, that's hindsight.

count55
06-03-2009, 05:49 PM
If the idea is that the only difference in a team's record is the attributed to the rookie, that is a major flaw.

Corey Benjamin must be one of the worst draft picks ever. After all, he single-handedly dropped the Bulls from being a 62-20 team all the way to a 13-37 (lockout year) team. It was clearly all his fault. It couldn't have a thing to do with Jordan retiring and Pippen getting traded, could it?

If I get a chance, I'll look at the EWA calc and try to see what impact that has...if I can find enough detail in the calc itself. I have to believe Hollinger has at least tried to do something to smooth this.

rexnom
06-03-2009, 05:53 PM
Well, where they get into trouble is the grades. Once they do that, they are implying the quality of the job being done by the front office.

What this information really shows you is what benefit the Pacers have gotten from the draft, which is to say, not a lot.

However, you need the other information that you're asking for to determine whether they should have gotten more.

Regarding your Bogut example, Bogut is, in fact, a weak #1 pick, but that doesn't necessarily mean that he was the wrong guy to take in that draft. However, even looking at 2005 by itself, hindsight says that Bogut should have gone no better than third, possibly as low as fifth or sixth (CP3, Deron, and arguably Danny, Bynum, and Monta Ellis). However, that's hindsight.
The problem is more than the grades, I think. They are basing whether Bogut was a bad pick based upon other number ones and not other picks in that draft, like you are.

Another example; Kenyon Martin. That's a terrible number one pick compared to other number one picks but a fairly decent pick based on that draft (only Michael Redd and Jamaal Magloire even made all-star teams from that draft). Though their EWA is based on individual performance (sort of, we've alluded to why that's not a good measure), the eEWA is based on other years' picks, which is unfair to the teams and the players.

ChicagoJ
06-03-2009, 06:28 PM
What this information really shows you is what benefit the Pacers have gotten from the draft, which is to say, not a lot.

Over the past twenty seasons, that is true. Go back just a few more years and you get Smits, Reggie and Chuck (and even Wayman).

By the time this study starts, the Pacers of the 1990s were generally constructed although there were some important trades to tweak the team. We were then drafting pretty late in the first round.

After that time was blown up, it was back to the trade market.

We haven't relied heavily on the draft in a long, long time. So much so that everyone seems to have forgotten how its done and how long it takes.

docpaul
06-03-2009, 08:37 PM
I don't know the full of it, but EWA has something to do with how many wins a particular player "added" to a team.

Though Bayless didn't do anything to help win any games for Portland, the fact is that Portland improved greatly in the win column last season (and Bayless was on the roster), while the Pacers didn't improve at all. It's probably reflected in that stat.

Hi, I think I might have incorrectly assumed that you all follow the basketball statistics community.. so here's perhaps some background if you're interested:

I first learned about this statistical methodology from a book called the Wage of Wins:

http://www.amazon.com/Wages-Wins-Taking-Measure-Modern/dp/0804758441/ref=ed_oe_p/104-9789661-4983110?ie=UTF8&qid=1136403009&sr=1-1

It's really an excellent book which lays important groundwork to statistically model the true "value" of a player (both their offensive and defensive contributions) in terms of the wins they add to a team.

The authors of this book have a pretty excellent blog that I follow regularly:

http://dberri.wordpress.com/

Not too long ago, this author team laid out criticisms of Hollinger's PER metric... perhaps in response to this online debate (which you can find by searching the Wages of Wins blog), Hollinger has evolved his metric... the VA and EWA (one which is a quick division of the other):

http://sports.espn.go.com/nba/columns/story?columnist=hollinger_john&page=PERDiem-090325

It borrows heavily from some of the statistics methodologies in baseball... such as the "value over replacement player" (VORP) model.

There's actually another fairly good derivative called the WARP:

http://sonicscentral.com/warp.html

Hope this is interesting/helpful. :)

MyFavMartin
06-03-2009, 08:51 PM
So if one of your star players goes down the following year, it leads to a drop in team wins, thus making your draft picks look bad?

MyFavMartin
06-03-2009, 08:54 PM
Wasn't Vonteego Cummings traded for Jeff Foster?

docpaul
06-03-2009, 09:03 PM
So if one of your star players goes down the following year, it leads to a drop in team wins, thus making your draft picks look bad?

Actually, "wins added" is more a product of how much better a player is in comparison to a replacement player for his position.

MyFavMartin
06-03-2009, 09:07 PM
Actually, "wins added" is more a product of how much better a player is in comparison to a replacement player for his position.

Is it just for the next season?

MyFavMartin
06-03-2009, 09:15 PM
Actually, "wins added" is more a product of how much better a player is in comparison to a replacement player for his position.

Also, isn't the addage to take the BPA, not draft for need?

MyFavMartin
06-03-2009, 09:19 PM
Also, weren't the Pacers doing something right by making the playoffs 16 of those 20 years?

Anthem
06-03-2009, 10:22 PM
Wasn't Vonteego Cummings traded for Jeff Foster?
Yeah, this is what I'm wondering as well.

Vonteego Cummings doesn't do well, so we got knocked for drafting him. Of course, he never played for the Pacers... we did a draft-day trade for Jeff Foster.

But then, we also didn't draft Brandon Rush. We drafted Bayless, and traded him immediately. Exact same situation.

So either Cummings should be off the list or Rush should. That's a pretty sloppy methodological error for a major magazine.

count55
06-04-2009, 12:19 AM
I don't know the full of it, but EWA has something to do with how many wins a particular player "added" to a team.

Though Bayless didn't do anything to help win any games for Portland, the fact is that Portland improved greatly in the win column last season (and Bayless was on the roster), while the Pacers didn't improve at all. It's probably reflected in that stat.


If the idea is that the only difference in a team's record is the attributed to the rookie, that is a major flaw.

Corey Benjamin must be one of the worst draft picks ever. After all, he single-handedly dropped the Bulls from being a 62-20 team all the way to a 13-37 (lockout year) team. It was clearly all his fault. It couldn't have a thing to do with Jordan retiring and Pippen getting traded, could it?

Upon reading the explanation of the methodology, this is not the case. In effect, they compare the new player's PER to the value of a replacement player:


There's one more step. We want VA to mean something, and in this case we want it to be the approximate number of additional points the player has been worth to his team, over the course of the season, relative to a replacement level player. To get to Point A from Point B requires us to divide the result by 67. Yes, 67. Sorry, that's what works. (If you're curious, a point of PER over the course of 2,000 minutes is worth about 30 points to a team, meaning that one point of PER over one minute is worth 1/67th of a point.)

So what we end up with is the formula:

VA = ((Minutes * (PER -- Position Replacement Level)) / 67)

Where Position Replacement Level = 11.5 for power forwards, 11.0 for point guards, 10.6 for centers, and 10.5 for shooting guards and small forwards.

This generates the "Value Added" (VA). This "replacement player" is kind of a generic player or normal, run of the mill, average player. Then to get the EWA, they:


n fact, you'll notice another column running alongside VA called EWA: Estimated Wins Added. EWA is the same idea as VA, except that the result is expressed in terms of wins instead of points. This is helpful if you're trying to figure out the impact of, say, removing LeBron James from the Cavs or Kobe Bryant from the Lakers. Note that the leaderboard for the two categories will always be identical; EWA is just VA divided by 30, since it takes about 30 points over the course of an 82-game season to add another win.

So, in fact, the team the player plays for, the difference in their wins, and the player he actually replaced don't particularly matter in the calculation.


Yeah, this is what I'm wondering as well.

Vonteego Cummings doesn't do well, so we got knocked for drafting him. Of course, he never played for the Pacers... we did a draft-day trade for Jeff Foster.

But then, we also didn't draft Brandon Rush. We drafted Bayless, and traded him immediately. Exact same situation.

So either Cummings should be off the list or Rush should. That's a pretty sloppy methodological error for a major magazine.

This goes back to what I was trying to say to MajorCold (oops, rex). The rating scale, per se, isn't bad. It's just that they're reaching too far and implying too much when they give grades to teams. Implicit in the grade is the idea that such and such team is good at drafting, while so and so team is bad at drafting. That over-reaches the analysis and creates "noise" or a "distraction" from what the data is telling you.

Take Vonteego Cummings as an example. The simple answer says take Jeff Foster's +0.73 and replace Cummings' -1.06. The result is the Pacers average goes from -0.35 to -0.30, which is immaterial. Also, you could argue that to really understand what the Foster pick "cost" you, Jeff would not be a +0.73.

To get Jeff Foster (drafted 21st), we used both a 26th pick and a 14th pick (the trade was Cummings and a future first, which became Troy Murphy). So, Jeff's 2.33 should not be compared to the normal 21 of 1.60, but rather to a combined 26 (1.20) and 14 (2.40), so Jeff Foster has really generated -1.27, not +0.73.

However, that's not really what the goal of the analysis was. The goal was to figure out what each draft slot gave you, much like the work that I did (and I have an update coming). The problem is, either the editors or the stat guys just flat over-reached. MajorCold is right when he says that you have to take the draft class into account when judging a team's drafting acumen. In other words, how often did they make the best possible pick in that draft?

However, I have absolutely no problem comparing number one picks across the years, as an example.


The problem is more than the grades, I think. They are basing whether Bogut was a bad pick based upon other number ones and not other picks in that draft, like you are.

Another example; Kenyon Martin. That's a terrible number one pick compared to other number one picks but a fairly decent pick based on that draft (only Michael Redd and Jamaal Magloire even made all-star teams from that draft). Though their EWA is based on individual performance (sort of, we've alluded to why that's not a good measure), the eEWA is based on other years' picks, which is unfair to the teams and the players.

It is certainly reasonable to say that both Kenyon Martin and Andrew Bogut were sensible #1 picks in the drafts when they were taken. However, it is also accurate to say that, as #1 picks, they were relatively weak.

Again, this is about production. There are good years to have #1 picks (1997, 2003, 2004), and there are bad years to have #1 picks (1998, 2000, 2001) just to name a few.

Keep this in mind...if you are looking for a perfect analysis, then all of these will fail. There is no such thing. You have to separate the wheat from the chaff. The info here is good and sound, but the prose and editorial narrative is a little spotty...but that's an easy trap in which to fall. You get used to talking to people who pretty much understand the ins and outs of what you're doing, so you resort to short hand. You apply the information to things that aren't entirely appropriate. You take for granted that everybody understands the strengths and limitations of the math, and a productive conversation amongst the well-versed becomes misleading and controversial to the newcomers in the audience.

On the whole, there's a lot to be taken from this. However, it, just like the posts in my "So, What's a draft pick get ya?" series, needs to be kept in perspective.

(BTW...if you're wondering...the reason that Brandon had a bigger negative than Bayless is because...he played more minutes. In effect, his PER of 9.0 was only slightly better than Bayless's 8.2, but he played almost 1200 more minutes. It also appears that they compared both to the 11.0 of a replacement PG, which is arguable...for both.)

Major Cold
06-04-2009, 07:55 AM
Yeah bro I appreciate the quote, but that wasn't me.

MyFavMartin
06-04-2009, 08:08 AM
Yeah, this is what I'm wondering as well.

Vonteego Cummings doesn't do well, so we got knocked for drafting him. Of course, he never played for the Pacers... we did a draft-day trade for Jeff Foster.

But then, we also didn't draft Brandon Rush. We drafted Bayless, and traded him immediately. Exact same situation.

So either Cummings should be off the list or Rush should. That's a pretty sloppy methodological error for a major magazine.

Also, Dampier was traded too. Seems too flawed on multiple fronts. If they did something like an RPI on individual's +/-, I'd think that would normalize for a lot of variables (players PT, level of competition, injuries and team turnover).

Major Cold
06-04-2009, 08:18 AM
Also, Dampier was traded too. Seems too flawed on multiple fronts. If they did something like an RPI on individual's +/-, I'd think that would normalize for a lot of variables (players PT, level of competition, injuries and team turnover).


Dampier was not a draft day trade.

Each draft has many variables that EWA does not take into account. Each team has many variables that EWA does not take into account. There is as much certainty in the EWA than in the draft itself.

Is Rush entire year up against Duns last year? Then I would understand that awful number. Would it be correct to say that the level play of the Pacers was not due to the play of Rush, but to the increase of play from Granger?

count55
06-04-2009, 08:38 AM
Yeah bro I appreciate the quote, but that wasn't me.

:blush:

It was late, and I'd had a few.

rexnom
06-04-2009, 09:28 AM
I'm still not fully convinced. Why can't I have an analysis that tells me if player X was a good pick at spot Y in year Z? Is that too many variables? My problem with this analysis is that they are acting as if this is the info they're telling us, when in reality, they are using pretty numbers and equation to tell us something different. Namely, that player X was a good pick at spot Y. The Z variable is totally discounted.

Major Cold
06-04-2009, 09:33 AM
:blush:

It was late, and I'd had a few.


Excuses...Excuses. You are the MVP. I expect more. It is time for a Milly Vanilli parody:

Blame it on the beer
Yeah, Yeah
Blame it on the time

Major Cold
06-04-2009, 09:35 AM
rexom,

I would think that the Z is the average, and so is X and Y. Any given year it would be a change. So it seems what you are looking for is a ratio of averages and any given year?

Occam's Razor.

rexnom
06-04-2009, 11:14 AM
rexom,

I would think that the Z is the average, and so is X and Y. Any given year it would be a change. So it seems what you are looking for is a ratio of averages and any given year?

Occam's Razor.
I'm too lazy to do the math or come up with a nicer formulation but here's what I think would work (assuming EWA is a good metric, which is a hefty assumption):

so Z=rexnom's eEWA

let Zi=sum(EWA for every player eligible for draft in year Y not including i or players drafted before i)/number of players eligible for draft in year not including i or players drafted before i

Their problem is that their eEWA is just a point from a best fit line they drew from a sum EWA of ALL years, which is silly. Z would calculate the value of each pick based on the players available, not other picks in the same position.

If someone could get me the EWA of all players eligible in the 2000 draft, for example, I bet I could show that while Kenyon Martin was not a great number one pick in general (i.e. he's not Yao, Olaujuwon, Shaq, etc. quality), he was an excellent number one pick in 2000. It is just wholly unfair to judge a team's drafting by this metric.

Naptown_Seth
06-04-2009, 02:45 PM
Well, forgetting the Austin thing for a second, my main problem with it is that you cannot say that pick x was a bad pick based on other years is fallacious. How can they say that Andrew Bogut was a poor choice in 2005 based on other years? What if there had been some sort of strike with college players and no one else declared and only international players were available (including Bogut). Would he still be a below average pick? They need to have some sort of variable to account for draft strength. Judging by their explanation, it seems that eEWA are simply points on a best fit line, which precludes the variable I mentioned.
Exactly, you have the Y=F(x) part but you still need the +C part, the offset.

Say it's a fat year for picks and the Pacers pulled crafty trades for years in order to stockpile picks 5 through 15. Even after players 1-4 are gone you have sitting there 10 guys all headed for PERs better than a typical #5 even gives you, so as you pick those #5+ at 7, 9, 13, you are hitting homeruns over and over in this no-brainer draft.

Then the next year it's me and guys of my skill set. I'm the top rated player in the draft even. No one will have a better career than me. The Bulls draft me at #1 (the correct choice) and years ago managed to get the next 4 picks too. Now they do take the top 5 guys correctly when we look back on things, but each one has a much worse PER than expected for the pick.

So now without weighting the Pacers are brilliant and the Bulls are dumb when it comes to drafting when the truth is one team lucked into a lot of picks in a good draft and the other compiled many high picks in a weak draft.

But weight the 2 drafts against each other and suddenly both teams picked perfectly for their years and score the same. They took the right guy for their spot.

So instead you use EWA or PER to slot the correct draft order and score teams for missing based on that instead. Do a differential that shows how much more EWA/PER they could have had with that same pick, rather than compare to estimated amount for that pick generic.

Not only does that solve the math, it also keeps teams from benefitting when they draft Duncan, the so-called "no s***" pick, as in "wow, you realized that Tim was the best guy, no s***". ;)

So the Spurs get a 0.00 for that one. There was nothing better they could have taken at that spot. And this helps when a team misses the best pick but still gets a pretty good one. Say you are comparing picking Melo, Bosh, Wade, etc. So you didn't take the best one but you did take a pretty good one.

This also slams teams that take a real dud when a stud was still on the board. EVERY team gets punished for passing on Arenas, for example. But teams that still took a pretty good player get punished less than ones that chased crap with an elite on the board.

You might think this hurts the Warriors who then score a 0.00 for taking the best guy left in the 2nd round since a 0.00 for a sleeper like that is wrong....but it's a much better score than the huge -15.00 or whatever the teams before them were racking up passing on him.

Since in the end you have a sum total to compare that means that the relative scoring is key and 0.00 is great when faced with -15 instead.

Naptown_Seth
06-04-2009, 02:57 PM
(BTW...if you're wondering...the reason that Brandon had a bigger negative than Bayless is because...he played more minutes. In effect, his PER of 9.0 was only slightly better than Bayless's 8.2, but he played almost 1200 more minutes. It also appears that they compared both to the 11.0 of a replacement PG, which is arguable...for both.)
Doesn't PER take into account per minutes aspects, such as FGMade - FGMissed (ie, FG%). I thought PER was meant to normalize over minutes to a degree. You play more and you do more bad things too, not just more good things. Subtract them against each other and playing more time doesn't get you anywhere.

I realize that it's not a dead-on perMin average, but mathematically it has aspects of that kind of normalization in it. Therefore I think they are wrong to do a second adjustment for playing time.

rexnom
06-04-2009, 03:19 PM
Exactly, you have the Y=F(x) part but you still need the +C part, the offset.

Say it's a fat year for picks and the Pacers pulled crafty trades for years in order to stockpile picks 5 through 15. Even after players 1-4 are gone you have sitting there 10 guys all headed for PERs better than a typical #5 even gives you, so as you pick those #5+ at 7, 9, 13, you are hitting homeruns over and over in this no-brainer draft.

Then the next year it's me and guys of my skill set. I'm the top rated player in the draft even. No one will have a better career than me. The Bulls draft me at #1 (the correct choice) and years ago managed to get the next 4 picks too. Now they do take the top 5 guys correctly when we look back on things, but each one has a much worse PER than expected for the pick.

So now without weighting the Pacers are brilliant and the Bulls are dumb when it comes to drafting when the truth is one team lucked into a lot of picks in a good draft and the other compiled many high picks in a weak draft.

But weight the 2 drafts against each other and suddenly both teams picked perfectly for their years and score the same. They took the right guy for their spot.

So instead you use EWA or PER to slot the correct draft order and score teams for missing based on that instead. Do a differential that shows how much more EWA/PER they could have had with that same pick, rather than compare to estimated amount for that pick generic.

Not only does that solve the math, it also keeps teams from benefitting when they draft Duncan, the so-called "no s***" pick, as in "wow, you realized that Tim was the best guy, no s***". ;)

So the Spurs get a 0.00 for that one. There was nothing better they could have taken at that spot. And this helps when a team misses the best pick but still gets a pretty good one. Say you are comparing picking Melo, Bosh, Wade, etc. So you didn't take the best one but you did take a pretty good one.

This also slams teams that take a real dud when a stud was still on the board. EVERY team gets punished for passing on Arenas, for example. But teams that still took a pretty good player get punished less than ones that chased crap with an elite on the board.

You might think this hurts the Warriors who then score a 0.00 for taking the best guy left in the 2nd round since a 0.00 for a sleeper like that is wrong....but it's a much better score than the huge -15.00 or whatever the teams before them were racking up passing on him.

Since in the end you have a sum total to compare that means that the relative scoring is key and 0.00 is great when faced with -15 instead.
Yup. Thank you. I thought I was going insane there for a second. And, to me, it's just absurd and weird to base the value of a pick based on other years and not that same year.

count55
06-04-2009, 03:36 PM
Doesn't PER take into account per minutes aspects, such as FGMade - FGMissed (ie, FG%). I thought PER was meant to normalize over minutes to a degree. You play more and you do more bad things too, not just more good things. Subtract them against each other and playing more time doesn't get you anywhere.

I realize that it's not a dead-on perMin average, but mathematically it has aspects of that kind of normalization in it. Therefore I think they are wrong to do a second adjustment for playing time.

It does, but when you do the VA, you bring the minutes back into the equation.


VA = ((Minutes * (PER -- Position Replacement Level)) / 67)

This makes sense to some degree. Think of it this way: If a player's not playing, then he can't really help the team win (unless he's Jamaal).

Roy Hibbert had a PER of 16.1, while Rashard Lewis had a PER of 16.8. However, Rashard Lewis was on the floor almost 3x as much as Roy, 2859 vs. 1009, so it would make sense for him to have a much larger VA/EWA despite the relatively similar PER's. (8.25 vs. 2.56)

Of course, the screwy thing with Bayless and Rush was that their PER's were both below water when compared to the "generic replacement player," who has a PER of 11.0. Theoretically, with a guy who's better than the replacement (like Roy or Rashard), the more they play, the more they help. In Brandon and Jerryd's cases, the implied notion was the more they played, the more they hurt.

The comparative Formulas:


Bayless = ((655 * (8.2 -- 11.0)) / 67)= -27.37/30 = -0.91


Brandon = ((1803 * (9.0 -- 11.0))/67)= -53.82/30 = -1.79

If you normalize (or equalized) the minutes either way, Bayless becomes worse than Brandon.

count55
06-04-2009, 03:45 PM
Yup. Thank you. I thought I was going insane there for a second. And, to me, it's just absurd and weird to base the value of a pick based on other years and not that same year.

Let's be clear about what they were doing in the first place: They were trying to establish what a pick would normally be worth. They were doing the same analysis I did, using a different set of metrics.

This is not a wrong thing, in my opinion. For example, the #1 pick in 2003 or 2004 was worth much, much more than the #1 pick in 2005 or 2006.

However, while that may be a good measure of what a team has gotten out of the draft, it is an incomplete metric when you're trying to determine how well a team has done in the draft. That is to say, a team grade should be based on how well they maximized their opportunities in each draft.

So, I'm ok with what they did, but I think trying to grade the teams was silly.

rexnom
06-04-2009, 04:01 PM
Let's be clear about what they were doing in the first place: They were trying to establish what a pick would normally be worth. They were doing the same analysis I did, using a different set of metrics.

This is not a wrong thing, in my opinion. For example, the #1 pick in 2003 or 2004 was worth much, much more than the #1 pick in 2005 or 2006.

However, while that may be a good measure of what a team has gotten out of the draft, it is an incomplete metric when you're trying to determine how well a team has done in the draft. That is to say, a team grade should be based on how well they maximized their opportunities in each draft.

So, I'm ok with what they did, but I think trying to grade the teams was silly.
I agree with you - I just wish they would sell this differently. It's not the value of a draft pick, which is inherently relative, but rather, as you said, the normal worth of each draft spot. For this reason, I had no problem with your analysis but I think that they were rather disingenuous about this.

ChicagoPacer
06-08-2009, 02:13 AM
If someone could get me the EWA of all players eligible in the 2000 draft, for example, I bet I could show that while Kenyon Martin was not a great number one pick in general (i.e. he's not Yao, Olaujuwon, Shaq, etc. quality), he was an excellent number one pick in 2000. It is just wholly unfair to judge a team's drafting by this metric.

Couldn't agree more when it comes to grading teams on drafting ability. In addition to the issue of comparing across drafts, there is a weakness of taking a players EWA out of the starting gate when he's a project player.

For the issue you have pointed out, it would make more sense to compare the drafted player's EWA to the two or three picks immediately following him in the same draft. If you're picking at #4, the guys you might be considering are probably going to go in the next 2 or 3 picks anyway. If you picked well, you walked away with the best or at least the second best of the 4. If you bombed, you selected the worst or second worst of the 4.

To eliminate the developing player bias, it might help to look at a player's cumulative EWA over his first 3-5 seasons to get a better picture of how he panned out. You won't have the history to do this with more recent drafts, but the jury is still out on guys selected in the last couple of years anyway, so I don't know if there is any value in calling them good picks/bad picks right now.

Naptown_Seth
06-08-2009, 07:21 PM
since it takes about 30 points over the course of an 82-game season to add another win.
Interesting.

I really need some space at work so I can return to crunching some of these numbers during the day. Or a long vacation with nothing else to do. It's killing me not to spend more time with some of this stuff. Maybe after the draft because draft clips is my current goal.



BTW, as I point out every year I think 82games and/or DraftExpress has a nice, detailed analysis of the value of each pick and how well teams have drafted. They also factored in what ChicagoPacer just mentioned, players that aren't good right away or are projects or become better for someone else but don't help you.


Frankly I would use the pick as earned each year and grade teams on what they got for each pick, period. So if you're Boston and you trade the #5 for Ray Allen then figure out how much he produced for you in the next, say 3 years vs what you could have had.

Of course there are still other factors, primarily COST. You get a better player but it costs you more money. Is the money linear with wins or a diminishing return. And even if you do win then you have the lost draft slots in future years, which do have value.

This leads us to the old Pacers debate. If Boston doesn't win with Ray Ray but instead limps along into 6th and a 1st round out, even if Ray does better than the expected #5 pick which makes them look smart, they aren't. They paid more for him than the draft pick would have gotten, the pick can pan out and be producing for you long after Ray is gone, and in the meantime you've added a few more high picks to the picture.

So the difference between brilliance and stupidity might be the luck of KG not getting hurt his first season with the Celtics - ie, LUCK.



Don't take this the wrong way, I do think teams can be graded on drafting ability (though as a team it's still dumb since many have swapped GMs, coaches, presidents, scouts, etc along the way), I just think we need to admit that probably every angle on it is still going to have flaws.

I mean even in my version where we do adjust from year to year, shouldn't we credit a team for trading out of a bad draft and into a good one, at least if they did it with actual insight to the situation rather than dumb luck? But who makes that luck/insight call?

Argh.

naptownmenace
06-09-2009, 02:05 PM
The thing I took from this is that the Pacers, since 1989, have only drafted one All-Star/potential franchise player.

To drill home the point even further, from 1989 - 2004 (prior to drafting Granger), the Pacers have only drafted 1 player that became a full-time starter for them - Dale Davis. Antonio Davis, Travis Best, Erik Dampier, Austin Croshere, Al Harrington, and Fred Jones proved to be good role players who started for a limited time for the Pacers but never for a full season.

docpaul
06-09-2009, 02:24 PM
The thing I took from this is that the Pacers, since 1989, have only drafted one All-Star/potential franchise player.

To drill home the point even further, from 1989 - 2004 (prior to drafting Granger), the Pacers have only drafted 1 player that became a full-time starter for them - Dale Davis. Antonio Davis, Travis Best, Erik Dampier, Austin Croshere, Al Harrington, and Fred Jones proved to be good role players who started for a limited time for the Pacers but never for a full season.

I know you didn't say this explicitly, but I'm not sure that this can be an indictment of drafting ability. If you buy into count55's excellent series of posts on the value of a draft pick (I do), then looking at the Pacers history on draft picks would show you that the team has only had two opportunities in the modern draft era to land a pick within the top 10: John Bender in 99 and George McCloud in 89. Bender wasn't even truly a pick (wasn't he a trade with Toronto for Antonio Davis?)

EDIT: Here's an actual list of draft picks by year for IND:

http://insider.espn.go.com/nba/draft/d05/tracker/team?team=ind

We already know that picking beyond top 10 or so provides very few to no "franchise players"... so, shrug... all their history shows me is that they haven't been able to beat the odds. :)