You probably have heard that in 2011 there will be a really weak NBA draft.
You also probably have heard about incentive to lose in the NBA, or in simpler word tanking.
It’s a well known phenomenon among observers of this league and yet IMHO two huge points are missing from this discussion. I’ll explain first one today and next one will be in tomorrow’s post.
But first, let’s start with the basics…
My assumption was that in the first 2-3 months of the season teams play up to their true abilities to win games because the possibility of reaching playoffs isn’t out of the question yet but when it’s not going well it creates an incentive to lose more and more games to improve team’s draft pick.
So for the Last 20 Full Seasons I picked 10 Teams Each Year which were the farthest away from playoffs [*], compared it’s wins in the first half of the season to wins in the second half and then sum them up for the whole year. It’s a very simple measure so any kind of discrepancy doesn’t have to mean “tanking” – it can be caused by a lack of luck in the close games or changes in the roster or injuries etc. That’s why I’ve examined every difference bigger than 5 games either way.
[*] – usually it meant simply the worst teams record-wise but in some cases there were ties for 10th or 10th worst team actually qualified for the playoffs.
Obviously, that idea has its flaws, for example:
- it doesn’t include teams which were tanking the whole season so I’ve checked every really bad records,
- it doesn’t include differences in schedule which could be an explanation in some cases,
- I assumed the same motive for every team – to improve draft odds – but in some cases, like Timberwolves in 2006/07, there was different one: they sold their pick to the Clippers, which was Top10 protected, so they had to be sure it will be in the Top10. I couldn’t chart fate of every pick like that,
- We are talking here about 20 years and around 200 teams so it’s possible that I’ve missed some changes, injuries etc. In those cases please let me know in the comments.
OK, that’s enough for the intro, here are the results…
| Season | Wins (1st half) | Wins (2nd half) | Difference |
| 2010-11 | 126 | 134 | 8 |
| 2009-10 | 131 | 117 | -14 |
| 2008-09 | 144 | 144 | 0 |
| 2007-08 | 129 | 121 | -8 |
| 2006-07 | 154 | 144 | -10 |
| 2005-06 | 153 | 139 | -14 |
| 2004-05 | 137 | 129 | -8 |
| 2003-04 | 155 | 139 | -16 |
| 2002-03 | 138 | 139 | 1 |
| 2001-02 | 138 | 146 | 8 |
| 2000-01 | 128 | 126 | -2 |
| 1999-00 | 154 | 143 | -11 |
| 1998-99 | there was a lockout so every team played only 2nd half | ||
| 1997-98 | 115 | 117 | 2 |
| 1996-97 | 124 | 111 | -13 |
| 1995-96 | 133 | 125 | -8 |
| 1994-95 | 127 | 142 | 15 |
| 1993-94 | 127 | 115 | -12 |
| 1992-93 | 143 | 134 | -9 |
| 1991-92 | 145 | 126 | -19 |
| 1990-91 | 132 | 138 | 6 |
Average difference is around -5 wins, which suggest that teams do lose more often in the second half, but it seems pretty random, right? Well, here’s the same table with one additional information…
| Season | Wins (1st half) |
Wins (2nd half) |
Difference | Top prizes/picks especially regarding hype BEFORE draft |
| 2010-11 | 126 | 134 | 8 | Weak at the top |
| 2009-10 | 131 | 117 | -14 | Wall/Turner |
| 2008-09 | 144 | 144 | 0 | Griffin, tie for 10th so 11 teams |
| 2007-08 | 129 | 121 | -8 | Rose/Beasley |
| 2006-07 | 154 | 144 | -10 | Oden/Durant |
| 2005-06 | 153 | 139 | -14 | Weak at the top, IMHO there were some legit injuries |
| 2004-05 | 137 | 129 | -8 | Bogut/Marvin/Deron/Paul |
| 2003-04 | 155 | 139 | -16 | Dwight/Okafor |
| 2002-03 | 138 | 139 | 1 | LeBron/Darko/Carmelo/Wade/Bosh |
| 2001-02 | 138 | 146 | 8 | Weak at the top |
| 2000-01 | 128 | 126 | -2 | Weak at the top |
| 1999-00 | 154 | 143 | -11 | Weak at the top, tie for 10th so 11 teams |
| 1998-99 | there was a lockout so every team played only 2nd half | |||
| 1997-98 | 115 | 117 | 2 | Weak at the top |
| 1996-97 | 124 | 111 | -13 | Duncan |
| 1995-96 | 133 | 125 | -8 | Iverson/Camby |
| 1994-95 | 127 | 142 | 15 | Weak at the top |
| 1993-94 | 127 | 115 | -12 | Robinson/Kidd/Ghill |
| 1992-93 | 143 | 134 | -9 | Webber/Penny/Bradley |
| 1991-92 | 145 | 126 | -19 | Shaq/Mourning/Laettner |
| 1990-91 | 132 | 138 | 6 | Weak at the top |
Now can you see the pattern here? Or am I exaggerating or cherry-picking data?
Because it seems to me that tanking happens only when there are consensus stars at the top of the draft so it’s not a function of improving draft odds, it’s a function of hype around players before draft.
IMHO it also means that weak drafts have one huge positive side: teams are not tanking because there won’t be a price worth doing it – that happened 8 times with a total +12 difference between 1st and 2nd half wins and a staggering -116 difference for 12 other drafts with hyped up prospects.
What’s more, even 3 exceptions to aforementioned observation:
- 2009 with Griffin,
- 2006 with weak draft at the top and
- 2003 with a very strong crop at the top
have a reasonable explanation:
- changes in OKC [+7], especially regarding Durant’s shooting, in 2009
- Blazers [with a worst record] and Knicks [who didn’t owed their pick] had -14 without an incentive in 2006
- Cavs and Nuggets were either really bad or tanked since the opening day and Hawks had a +7… when their inefficient top scorer [Glenn Robinson] missed some games in 2nd half of the season.
So was it a bad example of cherry-picking or am I onto something?
P.S. I have a breakdown of those wins, and case-by-case best guess for explanation of any difference, but it’s too long list to publish so if you are interested in any particular year or team let me know in the comments. Have in mind that I’ll focus more on some of the extremes in tomorrow’s post.
Weak Side Awareness 
DSMok1
June 23, 2011 at 15:59
Very interesting. You might be onto something.
Alex
June 23, 2011 at 16:46
Would you be able to do this again with maybe the worst 3-4 teams at the three-quarter or four-fifths point of the season? If teams are tanking, they should be able to do so more effectively with more information about where they’ll end up. Closer to the end of the season, only the worst teams should really be trying to lose. Teams that are, say, 6th or 7th probably have no shot at getting to worst and should actually try to play to their ability again (although they may be playing rookies or other worse players more than earlier in the season).
wiLQ
June 23, 2011 at 18:15
“Closer to the end of the season, only the worst teams should really be trying to lose. Teams that are, say, 6th or 7th probably have no shot at getting to worst ”
They do! Don’t forget that record is not a deciding factor and there’s a lottery so teams really don’t need worst record to acquire those top3 picks.
I’ve just checked total difference of wins between halves for all teams under 21 wins [40 of them] and their result was… +5. The worst 20 teams were only -1 and most of them had such bad first half they actually had a room to improve in the second . By this measure the biggest tank-fest goes on for teams in the 22-31 wins range and considering which odds landed first pick in the last years [8, 5, 3, 9, 7, 5, 6, 1, 1, 5, 3, 7, 3, 3] IMHO it makes sense.
What’s more, I’m worried that the smaller sample size the bigger problem schedule will become.
EvanZ
June 23, 2011 at 17:03
I have no doubt that teams tank. The lottery was implemented to prevent it, so it’s not like teams didn’t already have that strategy. If you’re an owner with a bad team, the best move is to dump salary and start over. The teams that don’t get it (like Golden State) end up in the mid-lottery year after year after year.
wiLQ
June 23, 2011 at 18:22
“I have no doubt that teams tank”
But that wasn’t my point:
I haven’t seen anywhere a connection between tanking and a quality of the draft at the top.
In mentioned article by Prof Berri he suggested that teams in the NFL and MLB don’t tank because of uncertainty about players’ quality… well, weak draft creates exactly the same situation in the NBA!
So ironically weak quality of draft works as a remedy for tanking ;-)
EvanZ
June 25, 2011 at 17:25
Yeah, I fully agree with you. I apologize if I wasn’t clear. I like what you’ve done here to suggest that causal link.
Chris
June 24, 2011 at 23:45
Nice Article very interesting, as for the top 10 Minn pick “sold” to the Clippers, it was actually traded by Minn with Sam Cassell for Lionel chalmers and a resigned Marko Jaric. It’s top 10 protection has remained in place ever since and finally becomes unprotected next season.
I apologize if this comes off as nitpicking, but being that it is 1 of top 3 trades during the Elgin Baylor regime of many terrible decisions I had to point it out. Keep up the good work! I especially enjoyed your charges drawn article(and many others)but being that I could not find those numbers anywhere on the internet for quite some time before Stathead led me to your site.
wiLQ
June 27, 2011 at 19:49
“Minn pick “sold” to the Clippers, it was actually traded”
It has such a different meaning? Because that’s what I meant.
“Keep up the good work! I especially enjoyed your charges drawn article(and many others)but being that I could not find those numbers anywhere on the internet for quite some time before Stathead led me to your site.”
Thanks! And that’s exactly one of the points of this blog ;-)
Crow
June 25, 2011 at 19:06
Given that you are working with blocks of 10 teams, the early to late season win change never exceeds an average of 2 wins per team. That is pretty small.
That could just be mainly from “playing the young guys’, players giving up (especially on defense), or going for their own stats, or fired coaches / new coaches, or coaches not having a clue how to fix things after the league gets there number or injuries or trades that give up talent to look ahead. It might not be primarily explicit tanking for draft picks.
Like Alex suggested, I’d like to see the full split of the bottom 10 into the bottom 3, then next four, then 3. Given what you already found here and how draft ping pong balls are assigned maybe team projected to be in the middle of the lottery have the most incentive to try to move down and fatten up their number of lottery balls? Something the pingpong ball distribution plan didn’t address because it was primarily concerned about the behavior of teams at the top and bottom of the lottery? .
Ideally it might be worth checking for teams with lottery protected picks involved, coming or going. I’d also like to see when changes in how the assignment of lottery balls are made against the data. If there is no significant average response after vs before that change that made tanking less lucrative in general and extreme tanking in particular then maybe it is not primarily aimed at that. Ideally I’d like to see tank suspicious behavior by team and with awareness of whether the coach and GM employed at the time of the alleged tanking was allowed to continue their job afterwards.
There is lot of room to continue to pursue this interesting topic.