The Leonsis Interview with Mike Wise: Some Thoughts In Response

Washington Post columnist Mike Wise sat down with Wizards owner Ted Leonsis for a 45-minute interview. The transcript is here, and I recommend reading the whole thing. I’m going to pull out a few of Ted’s comments that I think are worth commentary and/or analysis.

First, though — kudos to Ted for doing the interview at all. I disagree with much of what he says, but I appreciate his openness and willingness to face reporter questions. It’s more than a lot of owners do.

Leonsis: Wall, Crawford and Beal, that’s a pretty good three-guard rotation. We want to bring in a seasoned backup point guard. … And so what our belief is, we’re hoping John Wall and Beal become real stars that we keep and kind of build around them. Can Wall and Beal and Crawford one day be Isiah, Dumars and the Microwave? Right, I mean, that would be a pretty good backcourt.

I wish Ted had included “potentially” in this comment because it’s not a good three-guard rotation. And yeah, Isiah, Dumars and Vinny Johnson did make a good backcourt. But let’s be real: none of the Wizards “trio” have proven to be “good” NBA players. Wall has been average. His “glory stats” (per game points, rebounds and assists) look pretty good, but his shooting has been atrocious and his overall efficiency (including turnovers) has been awful.

Crawford is bad — one of the least efficient high usage players in league history. Since the NBA implemented the 3pt shot in 1979-80 there have been 930 player seasons in which a player received at least 1500 total minutes and had a usage rate of 25% or higher. Crawford ranks 907th in  individual offensive rating (individual points produced per 100 individual possessions).

During that span, there have been 851 player seasons with an individual offensive rating at least 3 points per 100 possessions better than Crawford. If we limit the look to players in their first two seasons, there have been 121 player seasons fitting the criteria above. Crawford ranks 108th on that list. Virtually anyone in the NBA could score as many points as Crawford did last season if he shot as frequently.

Beal, of course, has played zero minutes in the NBA. I predict that he’ll be a good player. In my draft analysis, Beal’s score in YODA was similar to freshman SGs like Michael Jordan, Clyde Drexler and Vince Carter. So there’s hope.

Leonsis: That’s why I looked at are we better at using our money in space to get Okafor and Ariza, then hoping and praying that we can get a free agent that believes and wants to come here. Then you hope while making a free-agent deal, those deals are pretty high-priced. Two years.

I understand what Ted’s hoping to do by using the cap space on Emeka Okafor and Trevor Ariza, but I think it’s a bad strategy. As I wrote a few days ago, they would have been smarter to buy out Rashard Lewis and amnesty Blatche, and then used the resulting cap space to outbid Dallas for Elton Brand (amnestied by Philly and picked up by the Mavericks for $2.1 million), outbid San Antonio for Danny Green (hugely underrated and re-signed for $4 million per season — an absolute bargain), and outbid Atlanta for Lou Williams (an efficient, 3rd guard type who signed for the mid-level exception). And they still could have had $8-9 million in cap room next offseason.

What the Wizards have done is buy a couple years of mediocrity at best. And at the end of those two years, they won’t be in position to pursue free agents because they’ll need to re-sign John Wall (they hope) and they’ll have other contracts coming due for extensions and renewals. This was their shot to use their cap space to add young players who fit their rebuild — they spent it on Okafor and Ariza.

Leonsis: We’ve made big investments in the analytic side and the technology side. … Besides our in-house guys, we have one cool guy: Joe Sill. Joe presents on occasion at that stats thing at MIT. Double-math PHd. He’s almost like a technical trader on Wall St. I can pick a company you should invest in. He’ll never meet the CEO, but he knows from the numbers which ones to pick. …

I do think there is a big, big role in informing some decisions. Also, that the little things have value. Our defensive rebounding — and the defensive rebounding stats of our guards — improved dramatically when Nene came and JaVale left. So, getting guards who can rebound becomes important. If your forwards are pushing their men out, that’s not a stat. That’s something you follow. That means the guards have the opportunity to get the rebounds and initiate their own break. Teaching rebounding becomes important. So Beal is a real good rebounder as a guard who fits really good with what we’re trying to do.

I’m a stat guy, so I like that they’re investing in analytics. Joe Sill is a “regularized adjusted plus minus” guy, and definitely a smart guy. The “promise” of RAPM is that it’s results are more accurate than straight (APM) because of something called “ridge regression,” which reduces standard errors. RAPM and APM are numbers I pay attention to, but I’m dubious about their utility because I’m not persuaded that the proponents fully understand all the factors that go into the “adjustment.”

Leonsis:

We’re also one of the few teams have installed this super heat-seeking missile cameras. Have you heard about this? We have these HD cameras. Another Stanford kid does it for us. This thing creates real-time heat maps. Literally you can get down to the pixels on the floor. Where are the shots being taken, where are the shots being made, where are the picks being made. It does interesting things like, how many dribbles on a fast break does your guard hold the ball before he dishes off, and was their a good shot made versus other guards in the league.

How does this work in practice: You tell a guard you were negating your speed by dribbling two more times. And then, when you dribble only three times and then you dish, we convert 70 percent of the time. All of this data then gets used in practice, like, in coaching sessions.”

I’ll tell you a lesson I learned 10 years ago with Ron Wilson and Adam Oates. I’ll never forget this. Adam Oates was the [quarterback] of our power play. Adam didn’t even know he was doing it, but he would put his skate up against the wall and bring that skate down and then get that pass. When the [power play] became not productive, he stopped doing that, and he was collecting the puck just a little further away, seven or eight inches away, from the half board.

The entire geometry of the ice changed, six or seven inches. Rob Wilson showed him, I remember, and said, ‘Just do that: Put your skate against the board.’ The power play came back. So that to me, as an owner, was the first indication that a little thing reviewed, fixed, coached can have unbelievable, big positive impact.

So we wanted to use as many tools as we can to try to give us advantage, and bringing in some of these coaches from winning programs. Bringing in these analytics, bringing in high IQ, good people it’s all a part of trying to change a losing culture to a winning culture.

This sounds really cool, and I’d love to analyze the take. The information is ultimately going to be as good as the analyst, and hopefully the team has good ones in place.

Interesting exchange between Wise and Ted about Blatche:

Q. I was reading quotes from two years ago about how much you thought signing Blatche to an extension was a great idea. Two years later, does this qualify as your biggest disappointment of owning the Wizards’ thus far?

A. Yes — we made a mistake — although the NBA has had close to $250 million of amnestied players to date — sometimes you get a chance to take a mulligan under the new rules and that is what we did.

Q. Who bears the most responsibility for the fact that he didn’t work out in Washington?

A. We are all in it together — so we are all to blame. Buck has to stop with me though as owner. I appreciate Andray’s apology to the fans and I hope he is able to turn around his career.

Q. Given your belief in redemption, was it particularly hard to cut him loose?

A. No, it wasnt. It was in best interest of franchise.

Ultimately, the person who bears responsibility for Blatche’s failure in Washington is Blatche himself. Had he worked hard, played hard and had a better attitude, he’d still be with the team and he’d be productive. He had ability. He just never put in what was necessary to come close to maximizing it.

On the (to me) difficult to understand decision to give Ernie Grunfeld another two years running the Wizards:

Leonsis: With Ernie what I found was, could we be on the same wavelength? Would he build team with eight or nine first-round picks? Could he make trades? I thought trading Gilbert was impossible. I thought trading Rashard was impossible.

It’s hard to even unpack what I think are faulty assumptions by Ted. Any player can be traded if you’re willing to pay the other guy’s price. They could trade Arenas because they took back a contract almost as bad. They were able to trade Lewis because they could solve all of New Orleans long-term salary cap issues with a single trade. And somehow, Grunfeld also gave up a 2nd round pick to make it happen. And New Orleans then used the cap room they acquired from the Wizards to trade for a 23-year old PF who happens to be supremely underrated.

I like several of the guys working for Grunfeld, but I have little belief that Grunfeld will build a title-contending team in Washington. The moves they made this offseason seem designed to get the Wizards in contention for the playoffs for the next two years. That is, contention for the 7th or 8th seed, not for a division title and a top 3-4 seed. To me, it’s a disappointment to see the team go for a sacrifice bunt when they had an opportunity to swing for the fences.

That said, I hope I’m wrong.

2012 NBA Draft Position Rankings from YODA

Image

Earlier today, I posted overall rankings for the 2012 NBA Draft using my draft analysis system, which has been dubbed YODA. Here’s the same information, but this time by position. In this post, I’m adding in a standardized score to show the differences between players.

The best score in YODA history was Shaquille O’Neal’s sophomore year at LSU. In the standardized score I’m showing, Shaq’s season scored a 36. Everyone else scales below that. Zero means the player rates as an average prospect.

Here are all players with a draftable score in YODA:

Point Guards

  1. Damian Lillard — 7
  2. Kendall Marshall — 4
  3. Tony Wroten — 0
  4. Reggie Hamilton — 0
  5. Jordan Taylor — 0
  6. Casper Ware — -1
  7. Scott Machado — -1
  8. Lazeric Jones — -3
  9. Marquis Teague — -3
  10. Scoop Jardine — -3
  11. Tyshawn Taylor — -4

Shooting Guard

  1. Marcus Denmon — 12
  2. Bradley Beal — 8
  3. Will Barton — 5
  4. Dion Waiters — 5
  5. Darius Johnson-Odom — 2
  6. Jeremy Lamb — 2
  7. John Jenkins — 1
  8. Orlando Johnson — 1
  9. Devoe Joseph — 0
  10. Doron Lamb — -2
  11. J.R. Cadot — -2
  12. Kim English — -3
  13. Matt Gatens — -3
  14. Tanner Smith — -3
  15. Jared Cunningham — -4

Small Forward

  1. Jae Crowder — 16
  2. Michael Kidd-Gilchrist — 8
  3. Harrison Barnes — 3
  4. Jeff Taylor — 1
  5. Quincy Miller — 0
  6. John Shurna — 0
  7. Ken Horton — -1
  8. Moe Harkless — -1
  9. Chris Johnson — -2
  10. Darius Miller — -3
  11. Robbie Hummel — -3
  12. Chace Stanback — -3
  13. Wendell McKines — -4
  14. Alex Young — -4
  15. Terrence Ross — -5

Power Forward

  1. Anthony Davis — 28
  2. Draymond Green — 7
  3. Thomas Robinson — 6
  4. Jared Sullinger — 3
  5. Terrence Jones — 3
  6. Ricardo Ratliffe — 2
  7. John Henson — 0
  8. Kevin Jones — 0
  9. Drew Gordon — 0
  10. Quincy Acy — -1
  11. Trevor Mbakwe — -1
  12. Perry Jones III — -2
  13. Arnett Moultrie — -2
  14. Mike Scott — -3
  15. JaMychal Green — -3
  16. Bernard James — -3
  17. Jack Cooley — -3

Center

  1. Tyler Zeller — 15
  2. Miles Plumlee — 3
  3. Andre Drummond — 3
  4. Meyers Leonard — -1
  5. Garrett Stutz — -3
  6. Fab Melo — -3

The Draft According to YODA

Image

Over the past couple months, I’ve been posting regularly in the draft threads on the Wizards board at RealGM about my research into evaluating NCAA draft prospects. The project is an effort to use college stats, as well as measurements, times and scores from draft combines to project which players are most likely to become good professionals.

The system I’ve come up with involves using per minute stats, includes an accounting for level of competition, and adjusts for age — a great performance from an 18-year old freshman is more impressive than the same performance from a 23-year old senior.

After referring to “the system” several times on the boards as Ye Olde Draft Analyzer, someone dubbed it YODA and the name stuck. The results? Time will tell on this year’s draft class. In previous years, YODA’s predictions have been good, but that’s the subject for a later post at some point in the future.

Below are the players as they’re rated by YODA. I’ve divided them into “tiers” — basically groupings of players by quality. The separation between Tier One and Tier Two is significant; the differences between players in the same tier are insignificant.

Here’s how I think about the tier system. Teams generally give themselves a better chance of succeeding when they pick best player available rather than focusing on specific roster needs. This should limit “reaches.” Plus, getting the best player gives the team another asset for trade purposes. Of course, it’s not always clear who the “best player” actually is.

That’s where the tiers enter the picture. Players who are “about the same” fall into the same tier. When a team’s pick comes around, they can use this process to pick whichever player on the same tier best fits their needs. They should avoid “reaching” into a lower tier to pick for need.

Anyway — the prospects according to YODA (along with sample players through the years who rated on the same tier):

NOTE: The tiers are numbered for this year only. They’re not historical tiers — that’s a work still in progress.

Tier One

Historical: Kevin Durant, Greg Oden

  • Anthony Davis, PF, Kentucky — Davis posted the most impressive freshman season I’ve analyzed. He’s the second rated prospect in YODA behind Shaquille O’Neal’s sophomore season. Similar YODA scores: Kevin Durant and Greg Oden.

Tier Two

Historical: Allen Iverson (SO), Zach Randolph (FR), Carmelo Anthony (FR)

  • Jae Crowder, SF, Marquette
  • Tyler Zeller, C, North Carolina

Tier Three

Historical: Jason Kidd (FR), Josh Howard (SR), Tim Duncan (JR)

  • Marcus Denmon, SG, Missouri

Tier Four

Historical: Paul Pierce (JR), Tony Allen (SO), Gilbert Arenas (FR), Ty Lawson (FR)

  • Bradley Beal, SG, Florida
  • Michael Kidd-Gilchrist, SF, Kentucky

Tier Five

Historical: Caron Butler (SO), Brandon Roy (JR), Rudy Gay (FR), Mike Miller (SO)

  • Damian Lillard, PG, Weber State
  • Draymond Green, PF, Michigan State
  • Thomas Robinson, PF, Kansas
  • Will Barton, SG, Memphis
  • Dion Waiters, SG, Syracuse

Tier Six

Historical: Ryan Anderson (SO), Ben Gordon (JR), Roy Hibbert (SR), Karl Malone (FR)

  • Kendall Marshall, PG, North Carolina
  • Jared Sullinger, PF, Ohio State
  • Miles Plumlee, F/C, Duke
  • Harrison Barnes, SF, North Carolina
  • Terrence Jones, F, Kentucky
  • Andre Drummond, C, Connecticut

Tier Seven

Historical: Antawn Jamison (SO), Deron Williams (SO), Jordan Crawford (SO)

  • Darius Johnson-Odom, SG, Marquette
  • Jeremy Lamb, SG, Connecticut
  • Ricardo Ratliffe, PF, Missouri

Tier Eight

Historical: Chris Singleton (JR), Jarrett Jack (SO), Harold Miner (FR), Charlie Villanueva (SO)

  • Jeff Taylor, SF, Vanderbilt
  • John Jenkins, SG, Vanderbilt
  • Orlando Johnson, SG, UCSB
  • Tony Wroten, PG, Washington
  • Quincy Miller, SF, Baylor

Tier Nine

Historical: Jared Dudley (JR), Gerald Wallace (FR), Hakim Warrick (FR), Lester Hudson (JR), Morris Almond (SR), Michael Redd (SO)

  • John Henson, PF, North Carolina
  • Reggie Hamilton, PG, Oakland
  • Kevin Jones, PF, West Virginia
  • Jordan Taylor, PG, Wisconsin
  • John Shurna, F, Northwestern
  • Devoe Joseph, SG, Oregon
  • Drew Gordon, PF, New Mexico

That’s enough to get through the first round. Tier 9 players are average prospects historically, according to YODA. In the “historical” sections, I’ve included a few representative players. It’s worth nothing that good pros have come from many “tiers,” which is demonstration of the reality that a) evaluating prospects is an inexact science; and b) that MANY young players have the capability to be quality professional players if they work hard enough and smart enough. The reality is that EVERY player arriving in the NBA needs to improve. The difference between a kid who becomes an All-Star and a kid who’s out of the league in three years is very much about the amount of purposeful, deliberate, well-considered work each player puts into developing his game and his body.

Who Do You Want Shooting In The “Clutch”?

Here’s the situation: Your team is involved in a close game (margin no farther apart than 5 points either way) in the 4th quarter or overtime. The question: Who do you want shooting the ball for your team?

The popular answer over the past decade has been Kobe Bryant. According to legend (and even polls of general managers), no one is more clutch than Kobe. No player of recent vintage can surpass (or even match) Kobe’s incredible, incomparable, sublime skills as a “closer” of games. Right?

Kobe fans would love the chance to regale us with details of why he’s such a great closer. They’d list great performances and incredible buzzer beaters, and we could all sit around and marvel at the incomparable daring-do.

Or, we could go to the Play Index and Basketball Reference and look up the numbers.

This season, in the 4th quarter and overtime when the scoring margin is 5 points or fewer (in either direction), Bryant leads the league in shot attempts. But, among the 37 players with at least 40 field goal attempts in the 4th quarter or overtime with the game margin within 5 points, Kobe has the has the 4th WORST shooting percentage — .327. Only Kevin Garnett (.325), OJ Mayo (.308) and John Wall (.238) have shot worse this season in clutch situations.

Well that’s just this year, right? Small sample size.

Since 2000-01, Bryant has 2094 field goal attempts, including 434 three-point attempts. His FG% on those attempts: 41.5%. His 3pt%: 29.7%. His eFG: 44.6%.

Hmm.

I went back through the records and found 246 individual players who generated 768 player seasons since 2000-01 in which the player averaged at least 1.0 FGA per game in “clutch” situations (4th quarter or OT, margin of the game 5 points or fewer). The averages:

  • FG%: 42.4%
  • 3pt%: 34.1%
  • eFG: 46.4%

In other words, the greatest closer of the past decade has actually been a below average shooter in clutch situations.

If we slice off everyone with fewer than 900 FGA (to find a group most like Kobe in terms of longevity and volume), Kobe still ends up below average in shooting from the floor.

  • FG%: .422 vs. Kobe: .415
  • 3pt%: .333 vs. Kobe: .297
  • eFG: .459 vs. Kobe: .446

Make the cut 200 total games in this situation and we get the same story.

The point is not to bash Kobe, but rather to highlight the utility of stats. Fans (and at least some basketball decision makers) “know” what they see. They don’t need stats — those are just for a bunch of geeks.

Except, subtle distinctions are virtually impossible to see. Take two hypothetical players. Both play exactly the same number of minutes for a team. They’re identical players in every way (including attempting 10 field goals). Over 82 games, that’s 820 FGA each. The only difference: One is a 50% shooter, the other a 45% shooter.

Over the course of a season, the difference is 41 made shots — one additional made shot every other game. If you say you can see that difference, you’re lying.

One of the values of stats is to check the claims that slide off the tongue so easily. Did I really see what I thought I saw? Did what others say happened really happen?

We get to find out whether Kobe really is a great clutch shooter, or whether he’s average (or below average) in the clutch. There’s value in that.

The underlying point to the Kobe hagiography is that any story about a basketball player’s successes is incomplete without accounting for the failures as well. When I hear that a player averaged 18 points and 8 rebounds, I don’t have enough information. How often did he shoot? How often did he get to the line? How many assists? How many turnovers? How many offensive rebounds?

Efficiency matters — a lot.

At this point, let me pause and say that I think Kobe is a terrific basketball player. He’s one of the better players of his generation, and he’d almost certainly go on my Rushmore of all-time NBA shooting guards. He’s not Jordan, but that’s not much of a criticism — Jordan was outlandishly, preposterously dominant.

But we’re all done a disservice when legend and hyperbole take the place of reality. Kobe’s done enough to be recognized as a great player. His fans don’t need to make stuff up to prop up his legacy. His overall meh clutch shooting needs to be considered, but there are more than enough accomplishments for Kobe to go down as an all-time great.

Should Basketball Players Be Evaluated Behind A Screen?

In Blink, Malcolm Gladwell tells the story of high-level music auditions. It seems that for eons, those who ran the music world thought that women didn’t have the strength to perform in orchestras. Women were deemed not to have sufficient strength to play certain piano compositions — and to lack the lung capacity and lips to compete with men on brass and woodwind instruments.

In the 1960s, leaders of the music world finally implemented The Screen. Where the auditioning musicians once walked onto a stage or into a room in full view of the judges, they began auditioning from behind a literal screen that prevented judges from seeing who was playing. Instead of assessing auditioners based on visual cues like race, gender, height or attractiveness, judges were forced to focus exclusively on the sound coming from the instrument. In most trials, the musician auditioning was not permitted to even speak.

What happened? The number of women winning auditions quintupled. The very first time the Metropolitan Opera in New York had blind auditions, they were seeking four violinists. Women won all four spots.

I’m reminded of Gladwell’s story by the recent stellar play from Knicks point guard Jeremy Lin. As you can see from the picture above, Lin is of Asian descent. And let’s face it — there haven’t been many Asian-American NBA players in history. I’ve been following the NBA since the late 70s, and I can’t remember even one. The only Asian players to reach the NBA I can recall are bigs from overseas (like Yao Ming).

We all know what basketball players look like, and Lin doesn’t fit the mold. Except that he’s 6-3 with long arms, and that he’s quick, and that he has good speed, and that he can jump a little.

Thrust into the starting lineup by injuries to other players, Lin has performed spectacularly. Nate Silver found 41 players who had similar results to Lin in their first four starts, and concludes that Lin is no fluke. The “screened” eye of stats suggests that Lin is likely to be a solid NBA player — at worst.

How did EVERYONE miss on a player who looks this good? How did Golden State and stat-savvy Houston decide to cut him? How did at least two pro teams take a smart, athletic, skilled player and dump him when he almost certainly could have helped either team?

I’m not about to accuse either franchise of overt anti-Asian discrimination. But I do believe that prejudice played a role in their evaluations of Lin as a player. Not in a conscious, “Asian people can’t play basketball” way, but at the deep-down, subconscious implicit bias level. He doesn’t look “right” to folks who have been conditioned for decades to “know” what good basketball players look like.

And by the way, I don’t think Lin’s ethnicity is the only “blinder” at work in his case. NBA talent evaluators have long overlooked offensive efficiency as a critical factor. The ability to shoot the ball with accuracy lags on the priority list behind height, speed and leaping ability. Never mind intelligence and skill — what’s the guy’s vertical? What’s his standing reach? A great athlete can “develop” the skills and learn the game. At least that’s the theory.

Start researching the NBA and you’ll find an array of blind spots and faulty assumptions from decision-makers. Playing time and salaries are determined mostly by per minute scoring — even though that’s not what wins. Minutes and length of a player’s career are largely determined by draft position — players picked high keep getting chances when more productive players picked later languish.

Want to know who’s going to make the All-Star team? Try picking the leading scorer on the teams with winning records. And no, the coaches (who pick the reserves for the All-Star game) do no better rewarding the game’s best players than the fans do in voting for the starters.

The All-Rookie team is determined almost exclusively by per game scoring.

Despite all the tools available to assist in evaluation, we keep getting “surprises” like Jeremy Lin.

Maybe it’s time NBA talent evaluators took a cue from the music world and started looking at players from behind a screen. Maybe it’s time to worry less about what people look like, and spend more time paying attention to what they’re actually doing.