Friday, February 4, 2011

Champions and Drafted Players

New readers, please check out the basics before leaving any comments.


I've been away for a while, so I've been writing a bunch of new posts in the short term. Today I'd like to revisit a classic topic: NBA champions and the draft.

Congraturation!

I've already shown that securing a top tree draft selection is generally not the ideal position to be in for an NBA team. To recap, I found that:
  • in the lottery era (1985-present), 60% of teams that have ended up with a top three draft pick haven't made it past the first round of the NBA playoffs during their draftee's first four years in the league.
  • to put it another way, only 40% of teams with a top three draft selection managed to advance past the first round in one or more of their draftees' first four years.
  • during the lottery era - only four (Duncan, Robinson, Sean Elliott, who was traded away and then reacquired, and Darko Milicic, who shouldn't count because he didn't really play) out of the 78 players (5%) taken with a top three pick have won a championship with the team that drafted them.
  • all that makes sense, as top three draft picks generally start their careers on bad teams. The Spurs lucked out with Duncan because, rather than being a bad team, they were a good team that had injury problems. The Pistons "lucked out" with Milicic because the Grizzlies were dumb enough to trade away their 2003 1st round pick in exchange for Otis Thorpe. Robinson and Elliott won their championships well after they were drafted, and Elliott was actually traded away and reacquired before he got his.
  • top three picks fared a little bit better in the pre-lottery era - ten players (Jordan, Olajuwon, Worthy, Thomas, McHale, Magic, Walton, Kareem, Unseld, and Cazzie Russell) out of the 57 top three draftees (17.5%) eventually won with the team that drafted them, and only 28% of teams couldn't make it past the first round in their draftee's first four years.
  • this also makes sense, as there were fewer teams in the pre-lottery days, which gave each team a better chance of moving up in the standings.

But there's another myth about championship teams that I want to address this time: the myth that championship teams are led by high (as in top three, for example - am I the only one who finds the term "high draft pick" rather ambiguous?) draft picks. I looked at every NBA champion from 1966-2010 and used total regular season Wins Produced to determine the top three most productive players from each team, recorded when these players were drafted, and calculated the averages. I settled on the top three players - instead of the top two or top five, for example - because it seemed to me like teams only go as far far as their top three players. In fact, according to a recent David Berri post, the performance of a team's top three players usually explains around 76% of a team's success, so three players seems like a good choice.

In determining Wins Produced, rather than using the yearly AdjP48 for each position for every individual season, instead I've taken the 1978-2010 average AdjP48s at each position (which - thanks to Arturo - can be found here), which admittedly isn't as precise as possible - but it is easier. Unfortunately, seasons before 1978 posed a few problems due to incomplete stat-keeping:
  • from 1977-1974, there are no individual TO stats kept.
  • from 1973-1971, in addition to the above, steals, blocks, and offensive rebounds weren't tracked.
  • from 1970-1966, in addition to the above, opponent statistics weren't kept.
Going back to 1974 is no problem; basketball-reference has a handy way to estimate TOs for those seasons. But I have no way of estimating all the missing stats for seasons played before 1974, which means that my rankings not be correct. For those years, sometimes it's easy to tell who should be where (Bill Russell, for example, was clearly the Celtics' best player when he was playing), but where it's closer (between, say, Jerry West and Happy Hairston on the '72 Lakers), I'm more likely to have made some mistakes. And remember, the average AdjP48 numbers I used are based on the averages from 1978 onwards - if things were a lot different before 1978, my rankings for older seasons are even more wacky.

Before I start discussing the results, I'd also better address some issues that popped up when it came time to record draft positions. They are:
  • territorial draft picks (Wilt Chamberlain, Oscar Robertson, Dave DeBusschere, Jerry Lucas)
  • undrafted players (Ben Wallace, Udonis Haslem, Moses Malone)
  • really late draft picks (Mario Elie, Bill Laimbeer, George Johnson)
I recorded the territorial draft picks as top picks (#1s); in Robertson's case, the Royals had the #1 pick anyways, so there is no harm there. Chamberlain would probably have been the #1 pick as well. That only leaves two bonus #1s, which isn't too bad.

Undrafted players were dealt with as follows:
  • find the year they made their NBA debut
  • count the number of draft picks made that year
  • add one to that number
Wallace and Haslem made their NBA debuts in years when there were 58 draft picks, which means that they were counted as 59th picks. Moses Malone, however, was a very weird case. Immediately following high school, Malone was drafted in the third round (possibly 27th overall) of the 1974 ABA draft. It's hard to know what number to give him, so Malone is the only player on the list that isn't included in the averages.

Since 1989, the draft has only had two rounds, but before then, the draft was routinely ten or more rounds long. This means that, while the latest a player can be drafted today is with the 60th pick, some players were picked well after that. Mario Elie, for instance, was taken with the 160th pick of the 1985 draft! Bill Laimbeer was taken with the 65th pick, and George Johnson (b. 1948) was taken with the 79th pick. Of these three, only Elie really throws things off. If we considered him undrafted - not too much of a stretch, given that he made his debut five years after he was both drafted and waived by the Bucks - and used the same rules as above, he'd be a 55th pick. The 160 is just such an outlier that I'm going to do just that.

All that being said, these are the results (full spreadsheet here):


What do we see? Several things:
  • the averages do play out in favour of higher draft picks. In every era, the average best players are earlier draft picks than the average second best players, and the second best players were drafted earlier than the average third best players in all but two of the eras (and one of these "eras" was only four years long).
  • however, in the lottery era, the best player on a championship team has been drafted around the 7th pick. That means that all those fans who are hoping that their favourite teams will tank should reconsider. While the number one pick is often head and shoulders above the rest of the draft class, there are always very productive draftees to be had all the way past the 7th spot.
  • the 1st best, 2nd best, and top three averages are all later picks during the lottery years than during the pre-lottery years. The only exception is the 3rd best players, which are still pretty close. While this is probably due to the fact that there are more teams in the league, it also could be that teams have gotten worse at drafting over the years.
  • the mid-late '70s were an interesting time - the third best players were drafted almost as early as the best players, and the second best players were taken really late. The average of the top three players in this era is an outlier compared to the other time periods. During this time, the champions were the Blazers (49 wins, 55 expected wins), the Celtics (54 wins, 47 expected wins), the Warriors (48 wins, 50 expected wins), and the Celtics (56 wins, 51 expected wins), so (and I'm getting tired of saying this) it's not much of a surprise.
  • the late '60s were dominated by the Celtics, making that era an outlier as well. The top three players during this era were picked around eight spots ahead of the top three players from the other time periods.

Very interesting stuff. Well, at least, I found it interesting.

 - Devin

11 comments:

  1. DDignam,

    In looking over your spreadsheet (as well as remembering back to Dr. Berri's blog), I noticed that Dennis Rodman and Ben Wallace skew the overall turnover average down, and know from my stat experience that 27 and 59 would make a much bigger contribution to the average than 1-6. By my guess, if we disregard these points, the average climbs up all the way up to 2-3, which would definitely provide justification for tanking.

    Besides, as an NBA GM, I'm noting that the number of top-5 top players is much higher than those of second-round picks (Rodman) or undrafted players (Wallace). Based on historical data, thus, I would be much more likely to find my top player by tanking now than by accepting the improvement and waiting later. And if I have the worst record in the league, I know I can go no lower than 4th and can draft right around the 3rd spot based on historical data. Thus, I am still going to tank.

    ReplyDelete
  2. I just checked the numbers, and if you look at all the best players in the turnover era other than Wallace or Rodman, the average climbs all the way to 2.13, which is far, far different than 7.

    ReplyDelete
  3. Well true, but Rodman was the most productive player on a championship team three times; it's not really fair to simply remove him altogether. A possibly better way would be to cut each top player only once and them average them. Wallace is pretty important, because his draft number is actually infinitely large. But let's say we take out Wallace and count every player once - the number is still 6.5.

    Like I said, certain lottery-era top threes have been totally worth tanking for - Duncan, Shaq, Robinson, Jason Kidd, Grant Hill, Elton Brand, Pau Gasol, LeBron, Dwight Howard, Al Horford, Durant, and Blake Griffin. But the rest of the list is filled with forgettable names...just take a look: http://bit.ly/hpvUxn

    ReplyDelete
  4. Devin:

    Although I'm loathe to quote him, Bill Simmons reported in his Book of Basketball that Moses Malone was drafted 5th in the ABA dispersal draft on 8/5/76 when the NBA & ABA merged. Don't know how that changes your analysis, but just wanted to put it out there - http://books.google.com/books?id=oLCSBeuStRcC&lpg=PA126&dq=%22book%20of%20basketball%22%20%22aba%20dispersal%20draft%22&pg=PA198#v=onepage&q&f=false

    ReplyDelete
  5. Devin:

    Very interesting post... Another reason for the best players being selected later in the lottery years could be that the worst teams have the worst management. Perhaps the teams with the better records performed better because they had better GMs running their drafts.

    ReplyDelete
  6. Mosi:

    Yes, thanks for reminding me - that was another point I meant to make. The same teams tend to show up in the lottery over and over, and the best teams usually stay out of it.

    Thanks for the dispersal draft info; we could use that and make him a 5th pick...but I don't think that's totally fair, because he had already been playing for a couple of seasons. Certainly he would not have been a NBA 5th pick when he was drafted into the ABA. I'd rather just keep him out and treat him as a special case.

    ReplyDelete
  7. DDignma,

    One problem that I see with a lot of the averages that you use for this explanation is that the NBA is a league where one great player will win multiple titles. Because of this certain drafts are significantly more valuable than other drafts. Since 1980 only 8 teams have won a title:
    LA Lakers - 10
    Chicago Bulls - 6
    San Antonio Spurs - 4
    Boston Celtics - 4
    Detroit Pistons - 3
    Houston Rockets - 2
    Miami Heat - 1
    Phildelphia 76ers - 1

    And since the favorites for this years title are (with margin of victory):

    Miami +7.8
    Boston +7.1
    San Antonio +7.0
    LA Lakers +6.7
    Chicago +6.0
    Orlando +5.7

    That group of eight is unlikely to expand.

    However if you look at those eight teams, the KG Celtics are the only one to win a title without a top 3 pick in their top 3 players.

    The Wallace and Rodman lead Piston teams teams featured high profile top 3 pick lead guards.
    Wallace had #3 pick Billups and Rodman has #2 pick Isiah Thomas. While those guys were not the best player on the team, they were crucial to the team being good enough to win the title.

    As for the Rodman lead 1998 Bulls, that team also featured Michael Jordan, a #3 pick who was the best player in the league for most of his career and was still among the top 5 best players in the league in 1998.

    I guess it comes back to the issue of how to get a top talent. As I see it, you have three options: draft, sign, or trade. In the draft, key players are likely to be selected in the top 3. Some fall past that but very few reach the second round or undrafted levels like Ben Wallace. To sign a key player, you need to have cap room and be an attractive destination. This generally excludes most of the NBA because it is a rich get richer scenario. The best example being the number of high profile centers that join the Lakers and win titles as a the best player in LA (Shaq, Kareem, Wilt, etc.) When Dwight Howard becomes a free agent in 2012, he is not going consider cities like Cleveland, Memphis, and Toronto. He will either stay in Orlando or consider teams that are already good and in big markets. Trading for a key player is very difficult as well since you generally have to give away valuable assets like high picks to get that player.

    If you are a small market franchise with no championship history, it is very difficult to attract top tier talent in free agency and for the same reason trades are difficult because you have to convince the player to resign in order for a trade to be a good idea. By contrast, if you draft a top tier guy then he has to play for you for 4 years and you just might be able to convince him to resign after that time period.

    ReplyDelete
  8. Anonymous:

    Thank you for the comment - I wish all of my comments were like that.

    I agree with almost everything you said! You're right that smaller-market teams have to build through the draft. You're right that the "top guys" - the best players - from every draft are generally really, really good, and that having one of those players is critically important to being successful in the NBA.

    BUT I disagree that having a top three pick is necessary to draft one of those players. Remember, here's the list of top three picks from the lottery era: http://bit.ly/hpvUxn

    Let's take 1985. Patrick Ewing was clearly better than Wayman Tisdale and Benoit Benjamin. But do you know who lasted until the 13th pick? Karl Malone. Oakley was picked 9th. Arvydas Sabonis was drafted 77th.

    How about 1986? You've got Daugherty, Bias, and Washburn. Dennis Rodman lasted until 27th, and Sabonis was back again for #24, and Hornacek at 46.

    1987? David Robinson was an extremely productive #1. Pippen was 5th, Kevin Johnson 7th, Horace Grant 10th, Reggie Miller 11th, Mark Jackson 18th.

    1988? Pretty weak year, but Rod Strickland was 19th and Anthony Mason was taken 53rd.

    I could go on like this for every year. Now, some of these later drafted players aren't in the same category as David Robinson, LeBron, Shaq, Duncan, Griffin, etc, but some of them - take Rodman, for instance - are, and there are always really good players available up into the teens and beyond. If I'm a small market team, I play as best as I can, try to draft the most productive players available, and hope to make smart, smaller trades that help my team win. Do that and the team will thrive, and maybe create a culture that will eventually start to entice big-name players. Relying on ping-pong balls isn't a good management strategy, because even the team with the worst record can fall as far as 4th.

    ReplyDelete
  9. Devin, thanks for this post. Sorry I'm a little late replying. I was looking around at good 2nd round picks recently. Here they are by year, along with their WS40 and PAWS40 for their final year of college. To make it fair, I included only players from American colleges. Anybody can be wrong about foreign players (and high schoolers in years past), so I limited the search to guys who should have been obvious.

    Name Pos. WS40 PAWS40
    Landry Fields SG/SF 11.7 13.5
    DeJuan Blair PF/C 21.7 19.2
    Chase Budinger SF/SG 9.5 9.7
    Joey Dorsey PF/C 16.2 13.8
    DeAndre Jordan C 10.9 8.5
    Luc Richard Mbah a Moute SF 7.6 7.8
    Carl Landry PF 12 9.6
    Josh McRoberts PF 10.5 8
    Ramon Sessions PG/SG 7.8 10.6
    Paul Millsap PF 18.3 15.9
    Trevor Ariza SF 5 5.2
    Luke Walton SF 7 7.2
    Steve Blake PG 7.7 10.5
    Keith Bogans SG 10.3 12.1
    Matt Bonner PF 11.1 8.7
    Mo Williams PG 5.2 8
    Kyle Korver SG/SF 13.7 13.9
    Roger Mason SG 5.2 7
    Carlos Boozer PF 16.8 14.4
    Matt Barnes SF 8.4 8.6
    Gilbert Arenas PG/SG 6.9 9.7
    Earl Watson PG 5.8 8.6
    Eddie House SG 7.8 9.6
    Eduardo Najera SF 10.4 10.6
    Michael Redd SG 7.4 9.2

    The lessons? First of all, if a superstar is available in the second round due to injury (Boozer and Blair), take him. Second, take the best player available: Fields, Dorsey, Blake, Bogans, and Najera all put up above-average numbers at big schools. They should have been no-brainers in the second round. Third, take stars from mid-majors in the second round: Paul Millsap, Ramon Sessions, and Kyle Korver were dynamite at mid-major schools. If you're keeping score at home, that 10 second rounders in 10 years that were above average in college and went to become above average in the NBA. The rest were legitimately second rounders - below average college guys with "potential" which they actually realized. So the lesson stands that good players are actually available deep in the draft, if you're just open-minded enough to look.

    ReplyDelete
  10. Awesome comment!

    I agree with your guidelines - time and time again, whenever anyone looks at the stats, they find that successful second round picks are often predictable. As a GM, I'd follow your recommendations.

    And I think you make an important distinction: high schoolers and international players are hard to evaluate, and so teams should be very wary of picking them with a top 5 pick (Bargnani, Milicic, Kwame, Tskitishvili, Eddy Curry, and Jonathan Bender immediately come to mind). Those are the type of players you want to pick later in the draft. The number of successful high schoolers and (exclusively) international players taken top 5 is pretty small (Garnett, Tyson Chandler, LeBron, Dwight Howard, Yao Ming, and Pau Gasol) .

    ReplyDelete