*************

SOLUTION TO NBA OFFICIATING: I’m serious, this is a game-changer, literally

The solution to the godawful NBA officiating starts with the creation of a complete and completely objective NBA officiating website—details below .  I floated the idea to…a senior-level executive of an NBA team, and he said “great!” but said they couldn’t fund it because it needs to be independent to have validity.  I don’t have the money, time, or the expertise to make this happen, but perhaps some of your readers do—it’s definitely a group project.

The officiating website I envision would be a resource for data without subjective content.  It would therefore *only* include:

Referee calls: video of every call—and every type of call—made by refs of every NBA game, in real speed, and slow motion, from every angle available from broadcasts. (Without audio.)
Statistics: raw data, counting stats—no percentages, ratios, graphs, etc.  Even choosing which metrics to post is subjective.  The site would only provide raw data to give users the means to conduct their own studies.

There would be no commentary, no chat boards, no links, etc.  There would also be no non-calls content.  It won’t be necessary.

If the raw data is available and made easily accessible to site users, fans, students, academic institutions, bloggers, et. al. they will provide the secondary content—the commentary and analysis.  If the site gives them something to work with, they’ll embed the video on their blogs and cite the statistics, and provide the commentary that will spur debate.

Furthermore, non-calls are as important an issue as whistled action, but eventually (and swiftly, I believe) non-call sites will pop up.  Tons will be partisan, but I think you’ll see objective-as-possible Sloan-academic-type sites emerge.  Objective-as-possible non-calls means trying to show all plays of all NBA games that 1) violate rules from the official NBA rulebook (which I believe is also available in video on nba.com) and 2) non-calls can be compared with whistled action from ‘my’ objective officiating site.  The non-calls issue is much more complicated and much more subjective—a longterm project, but the inherent subjectivity is why it must be left to the public, and not appear on the objective officiating site I envision.

I think this site is necessary regardless of the way the NBA league offices handle officiating review, and it’s understandable that the NBA league offices could never produce a site like this, for various reasons, but their infuriating behind-closed-doors policies of dealing with officiating just begs for public response—a cogent, progressive response.  NBA officiating simply *must* be subjected to independent analysis.  And I truly believe the site I propose is the beginning of the solution.  As Justice Brandeis noted: sunlight is the best disinfectant.

So consider just a few aspects of the site’s utility.  Besides all the conspiracy theories and beliefs that refs intentionally favor or seek to harm certain teams, unconscious bias is simply an inherent aspect of the human condition.  However well intentioned these guys are, refs aren’t robots.  But being forced to discover and confront those biases, I like to think refs will likely feel impelled to improve.  And either way, they’ll definitely feel the heat when the whole world can point to objective measurements of their biases.  And who knows, maybe the data will vindicate refs of certain accusations.  And of course, the data won’t just be used to confirm or invalidate popular beliefs about officiating; surely it will elucidate other matters.  For example, I suspect the data will assist in extricating the ‘noise’ from analytical studies of the NBA game.  But yeah, it will also help address accusations of malintent, “Everyone knows that this or that Ref X totally screws this or that team”—fighting, urge, to, use, Spurs, as, completely, randomly, chosen, example.  And my hope beyond hope is that someday, studies of the data and non-calls data will show *definitively* which refs just suck.  Early retirement!

And there are as many other possibilities as there are NBA fans who have the curiosity and creativity to ask questions and the dedication to pursue answers.

Best,

Michael

**********
That was long enough, but if you want an idea or two of how to get the ball rolling…

Obviously the builders of the site would need to watch every single NBA game each season as the season progresses, record them all, and post and categorize every, every charge, every moving screen, every travel, every palm, even every out-of-bounds.  And yes, they’d even have to watch Pistons games.

Also, it’s quite plausible to start with the ’09-’10 season, though it might take a bit of organizing and persuasion…  I’d contact every team in the league by email and phone, requesting one copy of DVDs of all of that team’s games from the ’09-’10 season.  Consider this idea:

First a polite email is written that 500 fans (or 5,000 fans) electronically sign.  Then a PR rep for each team is contacted at the same time (approximately, no watch synchronizing necessary I think) by phone just to let them know the email is coming, so it doesn’t get lost in the shuffle.  It’s easy to get teams’ PR reps on the phone—I did it with almost every team last season for a project—and their email contact info is available, too.

Then the email is sent with every team carbon copied—and it can be made known, tactfully, that their response will be public: “our 5,000 petitioners eagerly await your reply!”  I don’t even know how much cajoling would be necessary.  I suspect fan-interactive teams like the Mavericks would be happy to send the DVDs (especially considering the umbrage they’ve taken with the quality of officiating), and maybe other metric-centric teams will too.  But if a few teams send the DVDs (maybe even just one team) others might feel compelled to comply with the request.

That may not work, but once the site is established, if it gets serious traffic…well, traffic is everything.  If the site gets big traffic and publicity of any sort, grassroots or otherwise, I imagine it could build the clout to get teams to send DVDs of all games going back 20 years.  How cool would it be to use officiating data to analyze (umm, for someone else to analyze) how the game has changed through the years–that is, as the rules, trends, and basketball culture have changed?

Hey, if you’ve gone all the way through this email, thanks, that’s awesome.  I truly believe that the site I envision could be the beginning of a better League.  If you agree with me, well Mr. Huge Audience Blogger…

Best, again,

Michael

This makes me feel weirdly guilty, like I’m copping out.  And I hear this guilt-inducing voice in my head—specifically the voice of Han, the badguy from Enter the Dragon.  After Bruce Lee is captured by Han’s minions  and brought out to the courtyard, Han wants Roper (the obligatory white good guy) to fight him.  But Roper refuses (in a way that doesn’t make him look wussy, a sort of “I won’t be your puppet” thing) and Han says to him, “So!  There is a point you will not go beyond.”  That’s what I hear each time I don’t want to go what feels like the extra mile.  And the thing is, I don’t know if I can answer that in a way that doesn’t make me sound wussy.  I mean, right after Roper refuses to fight Lee, he whups like twenty guys, fighting with Lee.  There’s nobody for me to whup.

I was thinking of these things because I was just writing about the Mavericks and why they aren’t contenders.  Some of it is observation, but I think some of the raw numbers support these observations, and I’d like to take a slightly (slightly) deeper look into it.  But each step you take begs more questions, and more and more, and the whole thing makes me sleepy.  But maybe I’ll take a nap, and when I wake up (like, tomorrow) I’ll try to find the patience to address some of the questions.  We’ll see.

Billy Beane’s thinking and process of deliberation might be the same, but valuations are often conditional, and they change all the time.  Lewis would have you believe that success is all to be found by signing high OBP players at the expense of anything else.  The Athletics, rather, look not just for the most significant ways of valuating players, but (having relatively limited resources) the best deals available—undervalued players.  I haven’t paid attention in a few years, but last time I checked, Oakland was acquiring cheap, excellent defenders, and drafting high school pitchers.

Lewis wrote this Times article about analytical thinking in basketball, centered around Shane Battier, one of my favorite players, ever—I have always maintain that the Rockets got the better of the deal, acquiring him for (essentially) Rudy Gay.  And at one point he discusses the plus/minus stats that annoy me so much—which I go into here, and here.  One thing that’s very interesting to me, quoted below, is the talk on +/- stats.  Lewis gracefully calls the raw numbers “hardly perfect”, but I think it’s more accurate to say “utterly useless, absent more data to provide context.”  Daryl Morey says they can do that, isolating a player’s value by using +/-, adjusted in certain ways.  Cool:

“One well-known statistic the Rockets’ front office pays attention to is plus-minus, which simply measures what happens to the score when any given player is on the court. In its crude form, plus-minus is hardly perfect: a player who finds himself on the same team with the world’s four best basketball players, and who plays only when they do, will have a plus-minus that looks pretty good, even if it says little about his play. Morey says that he and his staff can adjust for these potential distortions — though he is coy about how they do it — and render plus-minus a useful measure of a player’s effect on a basketball game. A good player might be a plus 3 — that is, his team averages 3 points more per game than its opponent when he is on the floor. In his best season, the superstar point guard Steve Nash was a plus 14.5. At the time of the Lakers game, Battier was a plus 10, which put him in the company of Dwight Howard and Kevin Garnett, both perennial All-Stars. For his career he’s a plus 6. “Plus 6 is enormous,” Morey says. “It’s the difference between 41 wins and 60 wins.” He names a few other players who were a plus 6 last season: Vince Carter, Carmelo Anthony, Tracy McGrady.”

I can understand them “being coy about how they do it” but it’s annoying for outsiders like me, especially those who don’t have the stats chops or data (like me) to dig out answers, by themselves.  Their methods are “black box” by intent, rather than Hollinger’s condescending tendency (paraphrased) “the way I put this together is complicated, so I’ll just give you my results, not my methods.”  But either way, it just leaves me with the choice of accepting that their valuations or not accepting them.  Hollinger’s thinking is so blatantly flawed that his PER is easy to discard, but the Rocket’s valuations?  I don’t know, except that Morey seems to have a better idea of what he’s doing, and my feeling there is partially influenced by his credentials, academic and professional.  But I don’t know, and I don’t like accepting things as given without understanding them.

The real problem is that basketball just doesn’t seem to have the massive stats community that helped push the sabermetric movement, in baseball.  Yes, Bill James started it, and Rob Neyer and Michael Lewis re-broadcast that thinking to a much wider audience, but I think it’s the stats community outside baseball and popular media that finally cemented it’s place in baseball, in part by making it an open discussion for anyone to see.

I sort of  look at 38% shooting from three as a “Hey that guy’s a really good shooter” with 40% being elite and anything considerably above that as just gradually more amazing – and I think it’s always worthwhile to consider reps, and who’s guarding the guy.  Ray Allen draws the best perimeter defender every night, and the acute awareness of every other defender on the floor.

These numbers, by the way, are out of thin air.  I think of 38% as being really good because I just do.  I don’t feel like finding a better rationale.  Well, OK, just a bit: you have to shoot 57% on two pointers to equal the value of shooting 38% on 3-pointers.  57% on 2s is wicked rare.  There are currently 4 guys in the league who score regularly who shoot over 57% (all four are second or third options, none shoot 3s at all).  Meanwhile, there are about 30 guys shooting over 38% from 3 who are top three options for their teams (quick and dirty numbers).  Is this a lot?  I dunno.  Seems like a lot to me.  Ummm…let me see if I can look at about 4 years ago.  OK, ’04-’05 had about half that number for the year.  And it isn’t just the top shooters.  3-pt shooting percentages all over the league are up.  You can do a more comprehensive study if you like.  It’s enough for me to see that 10 teams are currently shooting 38% or better.  The NBA is just lighting it up.

Meanwhile, those international league guys are shooting from WNBA distance and I’ll state without a doubt that there are currently more guys in the NBA that can shoot well (arbitrarily 36% and up?) from NBA 3 distance, against NBA defense, than the rest of the professional leagues in the world, combined.  And this isn’t nothing to me, if you read about my thoughts on international leagues (I’ll link, tomorrow.  Remind me if I forget.)

Oh, and while I suspect that 90% of people with European Union citizenship could shoot over 85% from the line, with sprained fingers and a patch over one eye, standing on hot coals, I still think it’s not insignificant that there were 20 scoring-threat NBAers who could do it 4  years ago and 30 this year.

OK, we’re only half way through the season and this is all just me taking a quick glance.  But whatever, I feel pretty confident that I’m right.  And you do too, you cheeky bastard.

And yes, the time stamp is correct, if it says that it’s 4:30 something AM.  So yes, basketball is one of the top 5 things I think of when I can’t sleep in the middle of the night.  May god protect us all from the other 4.  Especially erosion.  Terrible, terrible.

“2. Welcome to the box score, plus/minus! ESPN’s box scores now include plus/minus in the far right-hand column, which it makes it easy to show, for instance, that C.J. Watson made quite a difference Sunday for the Warriors…”

Plus/minus statistics for individual players are so useless.  They tell you what happened, but not why they happened.  That the Warriors were successful when Watson was on the court doesn’t mean, in and of itself, that Watson was responsible for that success.  It doesn’t mean he wasn’t, either.  It doesn’t mean anything without more information.

Without context, a plus/minus score doesn’t tell you anything, and I’m not sure what context one would need to make plus/minus scores meaningful. 82games.com uses 5-man unit plus/minuses, and I think that can be useful, with enough data and context.  But if, for instance, you follow Hollinger’s thinking, San Antonio, which lost in Orlando the other night,  105-98, would have benefited from benching Tim Duncan (-13) for Kurt Thomas (+6) and Manu Ginobili (-19) for Michael Finley (+3)

Oh, and notice in the Warriors/Pacers game, Hollinger references, Danny Granger scored 42 points on only twenty-seven shots, grabbed 7 rebounds, and only turned the ball over twice in 41 minutes.  Granger (also a better than average defender) had a plus/minus of -7.  Ugh.

I don’t feel like going off on a whole diatribe against Hollinger’s PER statistic, because I don’t have the exact formula (I can’t find it in entirety on the web; I think I’m going to have to buy one of his old stats annuals) but among the statistics it counts are blocks and steals. Neither blocks nor steals are an indication of quality defense. In fact, Hollinger regularly writes in his scouting reports on steals leaders “This guy was ranked high in steals, but that’s just because he gambles; he’s actually a really bad defender.” So why should steals contribute, positively, to a player’s PER?  That doesn’t make sense.  He says high steals rates doesn’t make you good but high steals rates contributes to high PER and high PER makes you good?  Nope.

The same can be said about blocks, except that guys who block a lot of shots tend to be decent or better defenders.  But sometimes that isn’t the case; at least most of them aren’t really very good one-on-one defenders.  Again, the block itself isn’t a strong indicator of good defending.

Another stat of Hollinger’s with which I take umbrage is Usage Rating. UR essentially counts how many offensive possessions are “used” by a particular player, where “used” means he took a shot, made a pass that became an assist, or turned the ball over. This is a halfway decent stat I guess, just to describe what’s going on in a team’s offense, but again, it’s not a strong indicator of player effectiveness.  Theoretically, a team runs plays through a guy doesn’t mean that guy is good. Yes, it’s likely he’s one of the best players on the team, because he’s the guy they want getting touches, but the stat doesn’t indicate how well he did with that responsibility.  It could also just mean he’s a ball-hog who doesn’t play within a team concept, like Jerry Stackhouse.

Stackhouse is currently 9th among active players in UR, right behind Tim Duncan, and ahead of (in order): Richard Hamilton, Michael Redd, Dirk Nowitski, Jermaine O’Neal, Grant Hill, Ray Allen, Kevin Garnett, Corey Maggette, Jason Richardson, Elton Brand, Pau Gasol, Baron Davis, and Tony Parker.  It’s pretty hard to make an argument that he contributes more, offensively, than any of those guys except, perhaps, Corey Maggette and Jason Richardson (and O’Neal, since his body fell apart, a few years ago.)  Stackhouse is a low-percentage scorer, but is a huge ball-hog, so he has a very high UR.  Does that makes him an excellent offensive player?  Hardly.  Yes, most of the best offensive players in the league  have high URs, but plenty of mediocre players have pretty high URs, especially when they’re on a crappy team and have to handle more offensive responsibility than they would on a good team.

[It's not even a great stat to show how much a player is "used" because so often the pass before the assist is the one that really made the play. Duncan gets the ball in the post, they double him, he passes to the top of the key, that player swings it to a guy spotting up in the corner, that guy hits a three, and it was Duncan's potency as a scorer that made it happen, not that guy at the top of the key who made the second pass. But that's the guy who's usage rating is increased by that assist.]

I’m writing about this because Hollinger predicts a decline this season for Allen Iverson, the most gracefully aging little man in NBA history. Of AI Hollinger states:

“I know it’s uncouth to question “The Answer,” but Iverson is a small speed demon with a shaky stroke, making him a strong candidate for decline at age 33. His past two seasons weren’t nearly on the level of his 2004-05 and 2005-06 campaigns, so the descent appears to be already in progress. With Iverson battling knee and ankle problems, expect less rather than more.”

Actually, it’s the opposite. Three years ago he was excellent; two years ago he was excellent; last year he was sensational, probably the best year of his career. And yet his PER went from 26 to 19.6 two years ago, to 21 last year. Why? Well for fuck’s sake, he was traded to Denver and plays next to Carmelo Anthony, one of the best scorers (and shot-happiest players) in the league.  Iverson’s shooting accuracy increased each of those years in Denver, next to Anthony, last season being just .03 off his career peak, while his 3-pt shooting marks of the past two years are the best of his career.  However, to accomadate Anthony’s need for scoring opportunities, AI takes 6 shots less per game than he did in Philly (25 in Philly, 19 in Denver.) Does that mean he’s declining? Hardly, he’s adapted to a new team with a new gameplan, and has learned to be more efficient with less opportunity.  I don’t know if Iverson has slowed down at all; I can’t tell, but you sure can’t see it in his statistics.

And Hollinger should be too smart miss something so blatantly obvious.

He notes that the Celtics had an extraordinary offensive rebounding edge over the Lakers in the Finals and wondered “whether Lakers’ low number of offensive rebounds was just a matter of bad luck.” This comment assumes that getting offensive rebounds is valuable, and an important factor in teams’ success.

Offensive rebounds, he notes, account for about 30% of missed shots in the NBA, and even without fancy schmancy research “simple rebound rates can let us see things about the game for the first time.”

What things? More importantly, what’s the significance of these…things?

Ayres goes on:

OK, so what? Not that it’s relevant to this discussion (it’s not) but Philly and Miami were both crappy teams.

“Almost a 10 percent difference in getting the ball back when you miss can have a huge impact on games.”

Yeah? Prove it.

“2. Rebound rates give you a better sense of who are the best individual rebounders. During the playoffs, Tim Duncan had the most defensive rebounds per game“

This is incorrect. Rebound rates simply show who collected the most rebounds per game/minute, not who is best at rebounding.

“3. Finally, rebound rates show that the art of the offensive rebound is distinctly different than the art of the defensive rebound. Camby and Duncan, for example, rank only 35th and 21st among players in the playoffs in terms of offensive rebound probabilities.”

This is also fallacious. With the exception of freethrow shooting—isolated occurrences with universal conditions for all players—individual player statistics are highly influenced by context and can give you, at best, an indication of players’ talents. These statistics are often misleading without context and are never conclusive.

That the Celtics out-offensive-rebounded the Lakers in the Finals is not an indication that offensive rebounds are particularly valuable, nor that attacking the offensive glass is a strategy for success. Many coaches (like coach…me, and Jeff Van Gundy too, I think) deemphasize offensive rebounds, in order to get their defense set up and avoid giving up easy transition baskets. In fact, if you check out the stats, year after year, it seems pretty evident that there is no correlation between offensive rebounds and success.

For what it’s worth (nothing) on their way to a championship, the Boston Celtics ranked 23rd in offensive boards this season.