The Not-So-Instant Impact of 1st Overall Picks
Posted by Neil Paine on July 8, 2009
Let me ask you a question: Would you rather have Blake Griffin for the next 3 years, or 3 years of one of the 5 best players in basketball last year? How about one of the Top 10? One of the Top 20? In other words, how does the #1 pick typically fare during his 1st contract when compared to one of the league's best players from the year before he was drafted?
To answer this question, I'm going to use our old favorite SPM, since it's a rate stat and isn't so dependent on team performance as something like Win Shares (so as not to give the #1 pick, who often ends up on a bad team, an unfair disadvantage). We want to look at the cumulative SPM for each #1 pick in the lottery era vs. the top players from the year before, minimum 2,000 minutes (normalized to an 82-game schedule), over the following 3 seasons. Here are the results:
| Player | Ht | Pos | Years | G | Min | Stat+/- | 1-5 | 6-10 | 11-15 | 16-20 | 21-25 | 26-30 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Andrea Bargnani | 82 | F | 2007-09 | 221 | 5943 | -1.60 | 8.83 | 5.89 | 3.57 | 2.61 | 2.59 | 2.94 |
| Andrew Bogut | 84 | C | 2006-08 | 226 | 7326 | 0.60 | 7.37 | 4.71 | 4.94 | 5.56 | 2.10 | 2.04 |
| Dwight Howard | 83 | F | 2005-07 | 246 | 8714 | 2.35 | 6.99 | 6.67 | 4.70 | 2.87 | 3.72 | 3.02 |
| LeBron James | 80 | F | 2004-06 | 238 | 9871 | 6.99 | 7.90 | 5.64 | 3.31 | 3.00 | 3.08 | 1.32 |
| Yao Ming | 90 | C | 2003-05 | 244 | 7522 | 1.81 | 7.44 | 4.74 | 4.64 | 3.28 | 3.38 | 2.63 |
| Kwame Brown | 83 | F | 2002-04 | 211 | 4831 | -1.81 | 4.86 | 6.89 | 4.92 | 3.09 | 3.89 | 3.10 |
| Kenyon Martin | 81 | F | 2001-03 | 218 | 7404 | 0.52 | 6.06 | 5.10 | 3.96 | 3.55 | 1.70 | 2.61 |
| Elton Brand | 80 | F | 2000-02 | 235 | 8926 | 2.83 | 6.39 | 3.46 | 4.90 | 2.87 | 2.54 | 0.94 |
| Michael Olowokandi | 84 | C | 1999-01 | 207 | 5899 | -6.07 | 5.61 | 3.59 | 4.45 | 3.32 | 2.46 | 2.07 |
| Tim Duncan | 84 | F | 1998-00 | 206 | 8042 | 5.18 | 5.77 | 3.85 | 3.53 | 2.97 | 3.01 | 2.11 |
| Allen Iverson | 72 | G | 1997-99 | 204 | 8185 | 3.40 | 7.74 | 4.86 | 4.30 | 2.78 | 2.58 | 1.75 |
| Joe Smith | 82 | F | 1996-98 | 241 | 8251 | -1.91 | 8.27 | 5.10 | 3.55 | 2.52 | 2.70 | 2.97 |
| Glenn Robinson | 79 | F | 1995-97 | 242 | 9321 | -0.53 | 8.36 | 5.42 | 3.94 | 2.52 | 1.92 | 1.29 |
| Chris Webber | 81 | F | 1994-96 | 145 | 5063 | 5.06 | 9.50 | 3.78 | 4.00 | 4.54 | 3.93 | 2.38 |
| Shaquille O'Neal | 85 | C | 1993-95 | 241 | 9218 | 7.69 | 9.45 | 6.57 | 3.58 | 2.10 | 2.76 | 1.67 |
| Larry Johnson | 78 | F | 1992-94 | 215 | 8127 | 1.52 | 8.53 | 5.58 | 4.82 | 2.95 | 3.48 | 2.48 |
| Derrick Coleman | 82 | F | 1991-93 | 215 | 7568 | 3.23 | 10.78 | 7.55 | 4.62 | 2.35 | 1.91 | 2.74 |
| Pervis Ellison | 81 | F | 1990-92 | 176 | 5319 | 1.33 | 10.44 | 5.34 | 3.60 | 3.85 | 3.51 | 2.22 |
| Danny Manning | 82 | F | 1989-91 | 170 | 5416 | 1.33 | 9.89 | 7.47 | 4.29 | 3.54 | 2.49 | 1.95 |
| Brad Daugherty | 84 | C | 1987-89 | 237 | 8473 | 1.31 | 9.07 | 3.94 | 2.82 | 0.75 | 0.59 | 1.48 |
| Patrick Ewing | 84 | C | 1986-88 | 195 | 6523 | 2.98 | 8.40 | 2.72 | 4.26 | 4.05 | 2.02 | 3.43 |
| Average | 1986-09 | 1.72 | 7.98 | 5.18 | 4.13 | 3.10 | 2.68 | 2.24 |
As fans, we like to think holding the #1 pick in the draft is the key to an instant turnaround, pointing to examples like the Spurs in '98 with Tim Duncan, the Cavs with LeBron James, and even the Magic last year with Dwight Howard. However, recent history shows that if given the choice between a proven veteran and a talented rookie for the next 3 years, the vet will give you better production, and it's not really a close contest. 67% of the time during the lottery era, you'd actually be better off with a veteran who had ranked between #26 & #30 in the league the previous season than you would with the #1 overall pick over the next 3 years. Only in the case of a select few first picks (LeBron James, Tim Duncan, Chris Webber, Shaquille O'Neal) would it have been a good idea to trade their first 3 seasons straight-up for the next 3 seasons of a player who ranked between 6-10 the year before, and for zero #1 picks would it have been advisable to trade a Top-5 talent from the prior season.
Meanwhile, the bust rate for #1s has been high, given expectations. It's not unfair to expect the top player in the draft to be at least an average player during his first 3 NBA campaigns, but almost as many of them (5) were below average as were worthy of swapping with the thirtieth-best player in the league from the year before. Some, like 1998 #1 Michael Olowokandi, even rank among the worst players in the game during their 1st 3 pro seasons. What this means is that when picking #1, your odds of ending up with a disappointment are actually better than your odds of grabbing a superstar!
This isn't to say that Blake Griffin, the Clippers' prize #1 pick this summer, is going to be a bust or a disappointment. But the point is that all too often fans expect a highly-touted rookie to be the catayst in a franchise-changing turnaround, when in reality very few high draft picks -- even those identified as the cream of the crop, the absolute most talented prospects going into the draft -- end up being significant impact players in their first three professional seasons. Griffin could always buck this trend like a Duncan or an LBJ, but if history is any indicator, the odds are definitely not in his favor.
July 9th, 2009 at 9:59 am
Neil - Who were the top 1 - 5 in 1990-2 & 1991-3? They averaged out to over +10 SPM. Impressive! Actually 6 - 10 from 1991-3 have a higher SPM than 1 - 5 do eight different times!
July 9th, 2009 at 12:49 pm
It was truly a golden age in the late 80s/early 90s, wasn't it? We had MJ, Magic, Sir Charles, The Dream, The Mailman, and The Admiral contributing to those ridiculously high SPM scores for the Top 5 in '89 and/or '90. Crazy amount of talent in the league back then.
I noticed the #6-10 > #1-5 thing too -- my best guess is that a player makes it to #1-5 only when he hits his absolute peak, from which there is nowhere to go but down; meanwhile, #6-10 is often filled with pre-peak guys who are on their way up. That's just my knee-jerk assessment from glossing over the data for 2 seconds, though, so maybe I'll investigate the phenomenon further in a later post.
July 9th, 2009 at 3:08 pm
There's no question you're at least half-right. Take the best players at any point in league history, and they will definitely regress over the next few seasons. As to whether the next batch of players is _typically_ made up of up-and-comers, I doubt that is true, though it certainly may have been the case for those particular years. (But am I misreading the columns? I only see 6-10 beating 1-5 once, in 2002-04.)
July 9th, 2009 at 3:13 pm
I re-read it and I get what he's saying now -- #6-10 in 1990 were so strong that they would have beaten #1-5 in 8 other seasons, which again speaks to the total depth of top-line talent in those days. So go ahead and totally discount what I was saying about #6-10 being up-and-comers who passed #1-5, because I was looking at my own chart wrong. :)
July 9th, 2009 at 4:35 pm
Sorry for the confusion. I could have phrased that more clearly.
So I guess the guys filling out those 6 - 10 spots in the 1991-93 term would be Clyde, Pat, Nique, Scottie, & Stockton? Something like that?
July 9th, 2009 at 5:42 pm
You were close: Stockton, Drexler, Bird, Ewing, & 'Nique, to be exact. Bird was right at the end of his run by then, and his "91-93" average of +6.43 is actually a 91-92 average, b/c he was retired by '93.
July 9th, 2009 at 6:05 pm
You guys should get together and do something like the Hall of Merit for baseball. It would be interesting to see basketball players evaluated like that...
July 9th, 2009 at 8:33 pm
There already kind of is something like that at APBRmetrics, called the "Ring of Honor".
July 10th, 2009 at 2:49 pm
Neil - I'm looking at these numbers and getting confused again. When you look at the row for 2007-09, how can the group of players ranked 26-30 have a higher average SPM than those ranked 21-25 & 16-20? If the list is being weighted strictly by SPM, shouldn't those ranked 16-20 by definition have a higher SPM than 26-30?
2007-09 is just an example, not the only instance.
July 10th, 2009 at 3:04 pm
Sorry, I knew the format of this table would be confusing, I should have made it clearer. The rankings ("1-5", "6-10", etc.) are from the year before the player in question got drafted (we'll call that year Y). The numbers underneath the rankings, though, are the average SPM scores for that group in years Y+1 through Y+3. So to take your example, the players who ranked #21-25 in SPM in 2006 (the year before Bargnani was drafted) averaged an SPM of +2.59 over the following 3 seasons (2007-2009); meanwhile, those who ranked #26-30 in 2006 averaged an SPM of 2.94 from 2007 to 2009. Does that help it make more sense? I really should have presented this data better.
July 10th, 2009 at 3:09 pm
Got it! Thanks.
July 10th, 2009 at 6:08 pm
Neil, thanks for the link...