It’s a couple of weeks old, but I just now got around to finishing Austin Carr’s detailed and incredibly well-sourced story on the making of Amazon’s Fire Phone. Scathing take on Bezos’s involvement:

And team members simply could not imagine truly useful applications for Dynamic Perspective. As far as anyone could tell, Bezos was in search of the Fire Phone’s version of Siri, a signature feature that could make the device a blockbuster. But what was the point, they wondered, beyond some fun gaming interactions and flashy 3-D lock screens. “In meetings, all Jeff talked about was, ‘3-D, 3-D, 3-D!’ He had this childlike excitement about the feature and no one could understand why,” recalls a former engineering head who worked solely on Dynamic Perspective for years. “We poured surreal amounts of money into it, yet we all thought it had no value for the customer, which was the biggest irony. Whenever anyone asked why we were doing this, the answer was, ‘Because Jeff wants it.’ No one thought the feature justified the cost to the project. No one. Absolutely no one.” […]

According to three sources familiar with the company’s numbers, the Fire Phone sold just tens of thousands of units in the weeks that preceded the company’s radical price cuts. The $170 million write-down confirmed that the launch has been a dud.

I disagree with Carr’s assessment that Fire Phone was doomed from the outset because it didn’t fit within Amazon’s brand. Carr writes:

What makes the Fire Phone a particularly troubling adventure, however, is that Amazon’s CEO seemingly lost track of the essential driver of his company’s brand. It’s understandable that Bezos would want to give Amazon a premium shine, but to focus on a high-end product, instead of the kind of service that has always distinguished the company, proved misguided. “We can’t compete head to head with Apple,” says a high-level source at Lab126. “There is a branding issue: Apple is premium, while our customers want a great product at a great price.”

Brands are the result of products and services, not the other way around. The problem with the Fire Phone is that it’s a shitty phone. That’s it. If Amazon had made a phone with compelling features — an iPhone-caliber phone — it would have done just fine, and Amazon’s brand would have grown. If you set out to make a premium quality phone, you have to deliver a premium quality phone.

Consider a forgotten game in April 2010 between the Cleveland Indians and the Chicago White Sox. The White Sox were up a run with two outs in the eighth. Their set-up man, Matt Thornton, was on the mound, protecting a lead with a runner on first and the right-handed Jhonny Peralta at bat. Ahead in the count with one ball and two strikes, Thornton froze Peralta with a slider on the outside half of the plate, a couple inches below the belt. For a pitch like that, the umpire, Bruce Dreckman, would normally call a strike — 80 percent of the time, the data shows. But in two-strike counts like Peralta’s, he calls a strike less than half the time.

Sure enough, that night Dreckman called a ball. Two pitches later, Peralta lashed a double to right, scoring the runner and tying the game. Neither team scored again until the 11th, when Cleveland scored twice to win the game. Had Peralta struck out to end the top of the eighth, Chicago almost certainly would have won. ^{2 1}

This one call illustrates a statistical regularity: Umpires are biased. About once a game, an at-bat ends in something other than a strikeout even when a third strike should have been called. Umpires want to make the right call, but they also don’t want to make the wrong call at the wrong time. Ironically, this prompts them to make bad calls more often.

That’s according to research I did with David P. Daniels showing that the strike zone changes when the stakes are highest. We looked at more than 1 million pitches, almost all ball and strike calls from the 2009, 2010 and 2011 regular seasons, and found that the strike zone expands in three-ball counts and shrinks in two-strike counts. ^{3 2} It also shrinks again when the preceding pitch in the at-bat was a called strike. To put it another way, on close calls, umpires are unlikely to call a fourth ball, a third strike, or a second strike in a row. Umpires call balls and strikes as if they don’t want to be noticed.

The umpire’s job is simple: Call a strike when the pitch crosses the official strike zone; call a ball when it doesn’t. When the right call is obvious, umpires make it almost every time. One way to see this is to look at the probability of a called strike by pitch location.

Probability of a Called Strike

The plane at the bottom of the figure is the plane that rises from the front of home plate — the same one on which the official strike zone is occasionally rendered in television replays. The thick red lines on the axes denote the strike zone. The red on the horizontal axis is the width of home plate; the red on the vertical axis is the normalized distance between the batter’s chest and the bottom of his knees. ^{4 3} If you were a home-plate umpire, you’d be looking down through the plane, over the catcher’s head and towards the pitcher.

The 3D heat map rising from the plane measures the probability of a called strike at each location on the plane. Home-plate umpires are good at calling the obvious. Pitches that travel right down the center of the official strike zone — through the red at the top of the heat map — are called strikes more than 99 percent of the time. Pitches that cross the plane well outside the official strike zone — where the heat map is its deepest blue — are called strikes less than 1 percent of the time.

Umpires are inconsistent at the edges of the official strike zone, where the heat map turns green. Here, pitches that cross the plane in the same location are sometimes called strikes and sometimes called balls. This band of uncertainty is wide: about six to eight inches separate pitches that are called strikes 90 percent of the time and pitches that are called balls 90 percent of the time.

There’s a difference between an umpire being inconsistent and an umpire being biased. Inconsistency usually takes place within that band of uncertainty, when the umpire makes different calls on pitches at the same location. But he is biased when those differences correlate with factors other than pitch location, like the count. Where umpires are inconsistent, they also happen to be biased. To see this, consider two versions of the figure above: one for when the count has three balls, and one for when the count has fewer than three balls. These heat maps should be the same. Whether there are three balls in the count shouldn’t matter. All that should matter is the location of the pitch.

When we look at the difference between these two heat maps, we should see no difference — a flat plane. But we don’t. We see an expansion of the strike zone in three-ball counts.

Change in the Probability of a Called Strike With Three Balls

The official strike zone is the red rectangle beneath the heat map. The color and height of the heat map measure the change in the probability of a called strike when the count has three balls versus when there are two or fewer balls. The deep blue signifies no change — these are the pitches that are so obviously a ball or strike that not even a three-ball count changes them. In the center of the official strike zone, obvious strikes are still strikes; on the periphery, obvious balls are still balls. But on the edge of the official strike zone — in the band of uncertainty — a ring of mountains rises from the plane. The strike zone expands in three-ball counts, particularly at the top and bottom of the zone’s vertical axis. Borderline pitches, which are normally called strikes 50 percent of the time, are called strikes about 60 percent of the time with three balls in the count. Umpires act as if they would rather keep an at-bat going on a borderline pitch than issue a walk.

In two-strike counts, we see the inverse effect. For close pitches, a strike is now less likely to be called, which makes our heat map look like a moat.

Change in the Probability of a Called Strike With Two Strikes

The strike zone shrinks by as much as 20 percentage points in the top and bottom. With two strikes, borderline pitches — those that are ordinarily 50/50 calls — become 30/70 calls (30 percent strikes, 70 percent balls) for the average umpire. And with two strikes, the most biased umpire calls balls on borderline pitches almost every time. On close calls, umpires act as if they would rather give the batter another chance than call a third strike.

In both maps, the biases are greatest where the boundaries of the official strike zone are least apparent. What matters most is the vertical location of the pitch. Standing behind the plate, the umpire can easily tell whether a pitch is too far inside or outside. But it’s harder to know where the pitch is relative to the batter’s knees and chest. We would expect this uncertainty to breed inconsistency. But it also seems to induce the greatest bias. The highest peaks and the deepest parts of the moat are at the top and bottom of the strike zone.

Finally, we see that the strike zone shrinks again when the previous pitch in the at-bat was a called strike.

Change in the Probability of a Called Strike When the Previous Pitch Was a Called Strike

Here, the shrinkage is more uniform — about the same on the sides as on the top and bottom. The blue tips of the moat are about 15 percentage points deep: 50/50 calls become 35/65 calls when the last pitch in the at-bat was a called strike. Umpires appear reluctant not only to end the at-bat but also to call two strikes in a row. (Interestingly, there is no change in the probability of a called strike when the last pitch was called a ball.)

These mistakes are frequent — pitchers tend to pitch to the borders of the official strike zone. And they are consequential — they happen in the most pivotal calls. When a 50/50 call becomes a 60/40 call, as it does with three balls, umpires are mistakenly calling strikes on 10 percent of borderline pitches. When a 50/50 call becomes a 30/70 call, as it does with two strikes, umpires are mistakenly calling balls on 20 percent of borderline pitches.

Major League Baseball has embraced technologies that are meant to make calls on the field more consistent. The league has long used pitch-tracking technology to encourage home-plate umpires to behave more like machines, evidently without complete success. This past offseason, the MLB extended replay review to cover essentially all umpire decisions — except ball and strike calls. Now as before, no justice will be served when a pitcher throws a strike and the umpire drops the ball.

This article is adapted from “What “What Does it Take to Call a Strike? Three Biases in Umpire Decision Making,” which the author wrote with David P. Daniels.

What might have been.

Lee Hutchinson / NASA / NOAA

If we die, we want people to accept it. We are in a risky business, and we hope that if anything happens to us, it will not delay the program. The conquest of space is worth the risk of life.

—Astronaut Gus Grissom, 1965

It is important to note at the outset that Columbia broke up during a phase of flight that, given the current design of the Orbiter, offered no possibility of crew survival.

—Columbia Accident Investigation Board Report

At 10:39 Eastern Standard Time on January 16, 2003, space shuttle Space Shuttle Columbia lifted off from pad 39A at the Kennedy Space Center in Florida. A mere 81.7 seconds later, a chunk of insulating foam tore free from the orange external tank and smashed into the leading edge of the orbiter's left wing at a relative velocity of at least 400 miles per hour (640 kph), but Columbia continued to climb toward orbit.

The foam strike was not observed live. Only after the shuttle was orbiting Earth did NASA's launch imagery review reveal that the wing had been hit. Foam strikes during launch were not uncommon events, and shuttle program managers elected not to take on-orbit images of Columbia to visually assess any potential damage. Instead, NASA's Debris Assessment Team mathematically modeled the foam strike but could not reach any definitive conclusions about the state of the shuttle's wing. The mission continued.

In reality, the impact shattered at least one of the crucial reinforced carbon-carbon heat shield panels that lined the edge of the wing, leaving a large hole in the brittle ceramic material. Sixteen days later, as Columbia re-entered the atmosphere, superheated plasma entered the orbiter's structure through the hole in the wing and the shuttle began to disintegrate.

Read 89 remaining paragraphs | Comments

BBC Radio 1 recorded David Attenborough doing nature-style commentary for curling, but the YouTube video isn't available in the US, but luckily there's a copy on LiveLeak:

For the curious, here are the rules of and other assorted information about curling.