The Tesla Motors v. New York Times story reminds us of two things: one, if you tell someone you just took a road trip through the Green Mountains of Vermont and your GPS data indicates you were in fact cruising along Georgia’s Gold Coast, well, you’re probably going to lose the confidence vote. And two, that sometimes what a bunch of incriminating data seems to be indicating may not be what it’s indicating at all. Mark Twain popularized the phrase: “There are three kinds of lies: lies, damned lies, and statistics” in his Chapters from My Autobiography— we might well add vehicle logs and satellite data to that statement.
Let’s review: on 8 February, the Times published critic John Broder’s unflattering reaction to Tesla’s Model S sedan, based on an overnight test-drive up Interstate 95 along the Eastern Seaboard. Tesla Motors CEO Elon Musk was nonplussed, provocatively tweeting a few days later that the article was “fake” and alleging that the vehicle logs told the “true story”, namely that Broder “didn’t actually charge to max & took a long detour”. Broder batted back those allegations in another Times piece, after which the world waited for Musk to make good on his vehicle logs claim.
And so he did, producing a combative, retaliatory piece on 13 February, comparing the Model S’ GPS and performance logs with Broder’s claims in the review, and outlining what Musk viewed as several glaring contradictions; contradictions that he read as attempts by Broder, whom he alleges has an axe to grind with electric vehicles, to subvert the car. (“When the facts didn’t suit his opinion, he simply changed the facts,” wrote Musk, who sought to drive his point home by self-servingly adding “the Model S was declared to be the best new car in the world by the most discerning authorities in the automotive industry”). At this point, do-or-die electric car buffs had their long knives and sharpening stones out.
By cherry-picking data to make his point, Musk wasn’t being entirely fair to Broder, who, after all, had written a newspaper-style review and not a rigorous work of scholarship. In my experience, while a review isn’t necessarily the gospel truth, most reviewers aren’t trying to mislead or fabricate information. (I can think of a few well-known exceptions, but on balance, I’ve found this to be true.) Most of us read blogs, newspapers, and magazines for their accessibility, not their exhaustive qualifying statements and voluminous footnotes.
So when Musk introduced reams of vehicle data to the fray, he forced a different sort of conversation to occur, a conversation that at first appeared to favor the guy trotting out the detailed charts and illustrations with captions and little accusatory arrows. Hard data versus some writer’s allegedly prejudiced claims? Data wins! Even I was initially concerned about the apparent disparities, writing in my summary of the back-and-forth that “the most alarming point in Musk’s rundown is probably his last, where he surmises Broder intentionally attempted to sabotage the vehicle’s range by driving in circles.” And yet it’s with this very point that it turns out Musk undermines his own underlying data-will-win-out argument.
Tesla’s GPS logs of the trip showed the Model S at one point being driven in circles off a nearly-dead battery in the parking lot of a Supercharger station in Milford, Connecticut. According to Musk: “Instead of plugging in the car, Broder drove in circles for over half a mile in a tiny, hundred-space parking lot.” Clear evidence of attempted sabotage, right? Musk thought as much and cited the logs as proof.
Not so fast: Broder responded to Tesla’s data release in a third piece on 14 February, titled That Tesla Data: What It Says and What It Doesn’t. Here’s his explanation for that half-a-mile jaunt:
I drove around the Milford service plaza in the dark looking for the Supercharger, which is not prominently marked. I was not trying to drain the battery. (It was already on reserve power.) As soon as I found the Supercharger, I plugged the car in.So much for open and shut cases. In his post, Musk didn’t explain the event happened at night, nor do most readers have the faintest idea what it’s like to go hunting for one of Tesla’s new Supercharger hookups in the dark. And Broder’s explanation resonates with me: As a recent first-time diesel owner, I’ve probably driven at least as far around various truck stops looking for proper-sized diesel pumps (the nozzles, that is) when I travel between Michigan and Iowa or Minnesota.
Broder’s claim about driving around in the dark is of course still just that— a claim— but that’s also the point. Tesla’s data on the matter isn’t irrefutable evidence of foul play. Musk isn’t a mind-reader, and on balance, he comes off looking a little foolish, extrapolating “proof” that doesn’t really exist. What’s more, if he’d really been interested in getting to the bottom of these “disparities,” he could have picked up the phone and asked Broder for clarification, instead of dragging the data into the court of public opinion. In short, Musk should have been much more conservative in his analysis of the vehicle logs— he certainly should have curtailed his insinuations about Broder’s integrity.
We’re increasingly surveilled by the technology we use, be it our cars, smartphones, tablets, desktop computers, watches, pens, refrigerators, video game consoles— any so-called “smart” appliance with which we interact and for which there’s detailed evidence of that interaction. Privacy issues notwithstanding, the upside is a kind of elevated accountability– a silent reminder, if you will, that there’s now an electronic trail framing our actions. Sloppy accounting could (and increasingly will) come back to bite us.
But data explained by someone with a megaphone and an agenda — Tesla Motors exists to sell cars and promote itself, after all — is liable to be manipulated to serve that agenda. The cautionary reminder is that data still requires interpretation; it doesn’t speak for itself.
Rico says WHAT
No comments:
Post a Comment