Why the Lakers are playing so poorly

I enjoy Zach Lowe's basketball analyses on Grantland. He's delves a bit more into X's and O's than some of the pure statistical analysts, and I find that mix more illuminating considering the complex interaction effects in basketball.

His latest article is one of the best explanations yet as to why the Lakers are doing so terribly this year. Since I hate Kobe Bryant, I take perverse pleasure in reading chronicles of their shortcomings, especially when Kobe's poor defense is one of the Lakers problems.

But what's more interesting to me, in a way, is all those short YouTube video clips Lowe uses to break down plays from games. Who is uploading these videos of single plays? Is it Lowe himself, or someone else? If it isn't Lowe, how does he find the right clips of the right plays to use given the scarce metadata? Is it legal to upload these short clips from NBA games?

In the NFL you have to pay to watch NFL Game Rewind, but Lowe relies exclusively on these short YouTube clips which don't appear to have an NBA-sanctioned equivalent.

However these clips are produced, and however Lowe finds them, they're turning into the best option for play by play analysis of the NBA.

The Oath

I just finished The Oath: The Obama White House and the Supreme Court by Jeffrey Toobin. it's his follow-up to his best-seller The Nine: Inside the Secret World of the Supreme Court. Both are great.

I'm generally not too interested in law, but Toobin is covering only the most interesting cases of the highest court in the land, and I found both books engrossing. For this layman, what's eye-opening and somewhat shocking is just how powerful the nine Supreme Court justices are and how politicized the appointment process has become. In many ways, which nine people make up the Supreme Court should matter much more to the the average American citizen than who is elected President. A President can serve only up to 8 years in that office, but Supreme Court justices have decades to shape American life in the most fundamental ways. In fact, one of the ways it really matters who we elect President is that they get to appoint new Supreme Court justices as previous ones retire or pass away.

I've spoken to some of my friends who work in law, and some disagree with Toobin's legal assessments, but for someone without a deep knowledge of law, the book is written at the perfect level. Call it the "New Yorker" level of insight into a topic, as Toobin is one of their writers and an exemplar of the New Yorker idea presentation style.

Related: Critical Legal Studies

Compensation market inefficiency - sports vs. tech

Here's a mystery: in sports, some players on a team make more than their coach, but in the tech world (and the business world, for the most part), almost no employees make more than their manager. Which of the compensation distributions is more equitable?

First, there is a similar inefficiency in both markets, and that is that salaries for those fresh out of school tends to be bounded. In sports it's because of the agreed-upon arrangement between owners and players in each sport. In MLB, for example, a player is under team control for the first 6 years until they become free agents and can hit the open market. They have three years before they can even start to go to arbitration if they don't like the team offer. All of this is a huge suppressor for player salary. Albert Pujols will be paid more over the final 10 years of his career than the first 10, but it's almost certain that he produced more the first half.

In the tech world, most employees come out of school and get slotted into rough salary bands. In the beginning, that's largely fair. You can't tell how someone out of school will work with others or adjust to the more relentless rhythm of corporate life as opposed to the more lumpy work distribution in college.

But young developers, to take a group of people I think are particularly underpaid, typically produce a ton very early in their career, certainly more than they're paid for. It's not always evident right away, but within a short period the best quickly rise to the top. A lot of this has to do with the amazing leverage one software developer's code can have in the marketplace. Code can be made almost infinitely scalable. Even the greatest sports player can only have so much impact. The maximum is probably either a sport with few players on the floor at once (like basketball) and enormous marketing power in the global marketplace, or an individual sport where the player is responsible for nearly all of his/her performance (coaching being the other factor in an individual sport).

A huge difference between sports and the technology industry is thatInformation transparency on the true worth of a tech employee is much lower. Companies have a huge information asymmetry advantage over employees. In sports, the contracts of players are public knowledge, you know how much the guy in the locker next to you makes. Everyone's performance is highly visible, analyzed by thousands of professional and amateur analysts and fans.

In the business world, information about what different people are working on and how productive they are is not publicized that well within a company, and it's even less discoverable outside a company. You are typically limited to what you can put on your resume and who you can list as references if you take yourself on the open market.

[Incidentally, the lack of easy ways to quantify an employee's impact with much objective precision is exactly the reason that tech companies, who love to preach the virtues of transparency, can't be transparent with things like compensation. The disparity between one's compensation and one's impact, which will vary widely depending on who's judging, would cause way too much friction and gridlock. Imagine employees, like sports player unions, going on mass strike every so often.]

My sense is that both the sports world and the tech world are inefficient in their compensation schemes, but in different ways. In sports, I suspect many coaches are paid too little. Bill Barnwell makes a strong argument for Jim Harbaugh as a highly underpaid asset. In this article behind the ESPN Insider paywall, Bradford Doolittle makes a similar argument for the value of Tom Thibodeau to the Bulls.

In the tech world, I suspect the opposite, and that is that many employees are underpaid relative to their managers. Hiring great developers straight out of college and owning them during their equivalent of their pre-arbitration or pre-free agency years (to use a sports analogy) is one of the most important things a tech company can do, and Google, in particular, was the first to really exploit this inefficiency by quickly raising compensation and perks across the board. They drove the compensation bar up towards what is likely a more equitable equilibrium.

It's difficult for young developers to really assess their true worth because they typically have a very limited view on their impact. Even if they had a more global view, the ability to quantify the impact of a young developer versus a hypothetical replacement, to calculate the equivalent of WARP (to use a baseball term meaning Wins Above Replacement Player), in other words, remains very difficult. Will things like Github change this? Perhaps over time, we'll have more quantitative measures on something that will serve as a replacement for a resume, something more like the statistics on the back of a baseball card.

A related inefficiency that many claim is rampant in Silicon Valley compensation markets is age discrimination. It feels like some reporter needing to file a story will write a story on this once a year. Some recent examples include this one in Reuters and this one in the Mercury, but you can go back a few years and find earlier articles that say the same thing. This phenomenon is not unique to technology, you'll see it in sports, too. The brutal truth is It's partially the number of hours a young, single worker is willing to put in versus a married, middle-aged worker with kids, but it can also correlate to the willingness or familiarity of young developers with newer programming languages and techniques. Given how ruthless the technology recruiting battle is and how valuable great developers are, I suspect this inefficiency, if it exists at scale, will be ruthlessly exploited by some tech companies and closed over time.

Employees have some recourse for getting closer to market value. One way is to shop around frequently, like a free agent in sports, to determine true market value. Since there isn't an equivalent of a team's exclusive rights to your pre-arbitration or pre-free-agency years as in sports, you are essentially always a free agent. The downside is the increased stress from spending time selling yourself. Another option is to go do a startup, where, in a smaller team, you earn more visibility and credit for your output than you would within a giant corporation. One downside is that the nature of the work can be much different than in a larger company. Also, it's a much higher risk proposition, and it's not for everyone, but for entrepreneurial, risk-taking folks it's usually worth trying at some point if for no other reason than self-education.

Today we have headhunters in the tech world, or placement agencies, but at some point I wonder if we'll have the equivalent of a CAA in tech, with some knowledgeable agents representing the strongest developers and finding the most interesting work and highest compensation for them.

Miscellaneous

From Moonwalking with Einstein, I learned about using memory palaces as a mnemonic to help memorize long lists of things. Now some researchers have tested and validated the technique by having people use unfamiliar virtual environments as memory palaces.

In a NYTimes op-ed, David Agus asks "when does regulating a person's habits in the name of good health become our moral and social duty?" He has one suggestion, and that is to make it public policy to encourage middle-aged people to use aspirin. 

The most tweeted movie of the year? Think LIke a Man.

This link is a bit math-heavy and abstruse, but less so than you'd think from scanning it. Stein's Paradox in Statistics (PDF) by Bradley Efron and Carl Morris is a famous and fascinating article in which the future batting averages of 18 major league baseball players after their first 45 at bats in 1970. It is a useful introduction to the James-Stein Estimator and concepts like regression to the mean and how to quantify it. In the tech business world, managers tend to be rated on many qualities, but rarely on the quality of their forecasts. Given the value of forecasting in such a fast-paced industry, it's interesting how much people in tech rely on gut instinct.

Your site has a self-describing cadence

A problem lots of websites wrestle with is driving repeat visitation. To some extent, it's in your user or customer's hands. Sometimes that person needs to buy a book, and so they end up at Amazon. Or maybe they want to catch the latest breaking news on some developing crisis, so they visit the NYTimes or their news site of choice.

But a lot of it is in your control, too. An often overlooked tool in this struggle is design. Your site design is a visual metronome from which visitors learn the proper cadence of their visits.

As a test case, let's take the waterfall of news design so popular across so many sites (Dave Winer calls it a river of news, but my mind has always pictured rivers traveling horizontally, so even before I'd read Winer's term I'd thought of this single vertical column design as a waterfall). The design refreshes constantly, and the message is, "Visit often for the latest and greatest, this waterfall won't ever run dry." The newest items are at the top, signaling a site optimized for repeat visitors. If you look at the timestamp between entries in the waterfall, you get a sense for how often new content flows down from the top. Combine that with your average interest level in the average item in the waterfall, filter it through your own interest in the range of content covered at that site generally, and your mind will set an internal attention clock that gives you the itch to revisit the site at specific intervals.

It's a design at the heart of Facebook, Twitter, Instagram, most email clients, and most blogs, and it's one of the most comforting of all web designs. I also suspect it's not a coincidence that these are among the most frequently visited services on the internet. Their very design screams for frequent attention, though email is a bit more insidious in that it confronts the user with a persistent stack of items that accumulates over time, that taunting unread count in the inbox.

[Blogs are the exception though it's usually because single-author blogs can't update frequently enough to demand user visit every day. The blogs that I suspect do get daily visitors either have multiple authors (like Grantland or Boing Boing) or just update multiple times daily through sheer effort of their authors who treat it almost like a full-time job (like Daring Fireball or Kottke).] 

Twitter and Facebook feeds, unlike your email inbox, only show you a very recent slice of items, the rest just wash away down river, forever lost to the past. Whether or not you mourn missed items in your Twitter or Facebook news feed, the way each sites treats that content suggests you shouldn't really mourn what you missed. I'm not a fan of Inbox Zero because it feels like a suboptimal allocation of time to be beholden to the arbitrary number of emails other people send you. I treat my email more like a Twitter/Facebook feed, flagging important items for follow-up and letting the rest just wash downstream. My inbox has tens of thousands of messages, but it doesn't bother me in the least.

Twitter interfaces that auto-refresh and actually put the news feed into motion are the most awesome and terrifying implementation of the waterfall design yet (I recall watching one during a McCain-Obama debate in 2008). It conveys not only as near a real-time content stream as there is (next to a live video chat), it demands focused, unbroken attention from you right now. Reading your Twitter feed normally feels like you're following 100 yards behind someone, picking up crumbs of thought they've left on the sidewalk, but these live Twitter feeds that refresh automatically and keep pushing the latest tweets to you in real-time feel like standing next to someone, walking step for step with them, listening to them talk.

Let's take another site's design to see what it communicates about how often to visit. Techmeme is a really popular news among tech industry followers. It has some qualities of a waterfall design, borrowing from it a single dominant column of items for its left column. However, unlike pure waterfall designs, it doesn't always put the newest news up top. Instead, it tries to put the biggest tech story up top, whatever that might be at that moment. If you visit Techmeme multiple times in a day, the ordering of stories in that column might shift, and over time, you start to learn from those differentials how frequently the site shifts, and within a single story cluster, the top item in that cluster might shift as the story develops. For those who are tech news junkies, the site design efficiently cues users how to prioritize their attention among stories and within stories at any moment in time. Even leaving out the other parts of the Techmeme homepage, the left column, as a sort of waterfall variant, is a highly efficient traffic cop for your attention.

This might all seem obvious, but you only need to visit a site whose design is muddled on this topic to see how important it can be. As an example, take the homepage of Time. Like many news sites, rather than a waterfall design, Time has a complex hybrid column/grid design. The NYTimes has a similar design on its website homepage. The cues as to how frequently you should visit are muted in favor of offering up a sense of everything there is to offer from the site. The reader has to process a ton of stimuli in any single screen, with a wide variety of typeface sizes and weights, interspersed with photos.

It's difficult to parse the rate of change on pages like this because the one single item within the grid can change but the rest of the page can remain unchanged and the user will have a hard time remembering what was where the last time they visited. In a waterfall design, change affects the entire page, shifting everything down. It means the user might miss something of importance, but it is unequivocally clear in signaling both cadence and priority to the user (many of these sites cope with the issue of the user missing critical items by adding a thinner column to the side of the main waterfall where they pin the top items for whatever time interval they treat as primary, which for most news sites is a day).

A site like Time or the NYTimes might offer the same signal to noise ratio as Twitter or Facebook, but the waterfall design of the latter feels like a more efficient way to seek out the signal. Sites like Time and the NYTimes explode out all the stories along multiple axes in different blocks of content, and that adds axes along which visitors must parse out that signal. The result is that I rarely if ever visit the homepage of the NYTimes; I use other sources of signal (for example, Twitter or Techmeme links) to send me directly to what I need within those sites. I don't think it's a coincidence that waterfall design sites sit at the top of my attention funnel. They both cry for my attention all the time and are hyper efficient designs for presenting me with high information density.

If you have a realistic target for visit frequency for your site, think about how to communicate it with your site's design. All sorts of cues are telling your user how often to return. As an exercise, you might look at these three sites.  What does their design say to you about how often to visit and why? What do you imagine their realistic visit frequency is, and and how good a job you think they do at communicating that with their site design? Note that it's quite possible that some services don't even try to deal with the issue with their homepage design because they don't intend for most users to use their service by visiting the homepage regularly. For example, some services  rely heavily on email to establish a cadence with users, like Dave Pell's Next Draft or Rex Sorgatz's ViewSource.