A few days ago, the New York Times published an article criticizing the range and performance of the Tesla
Model S in cold weather. The author, John Broder, claimed that the car performed poorly in cold weather and that he eventually had to have the vehicle towed to its destination on a flatbed.
When Elon Musk strenuously contested Broder's claims of subpar performance, the author doubled down. In a follow-up article for the New York Times, Broder claims that he charged the vehicle properly, reduced his speed to 45 mph, and turned down the heater temperature in an attempt to compensate for the vehicle's fast-draining battery.
Today, Musk released the extensive data logs that Tesla keeps on the cars it sends out for review (the data logs aren't activated for consumer vehicles unless the customer explicitly agrees to it).
All of the images below are courtesy of Tesla Motors and come from Musk's own blog post.
The results are pretty damning. The highlights include:
- The vehicle never stopped running, even at the point when Broder called a flatbed truck for a tow.
- The battery was only charged up to a 32 mile estimated range for the final leg of Broder's trip -- even though the next charger was 61 miles away.
- Broder's claims of traveling 45-54 mph are false. Broder kept the car at 65-81 mph.
- Broder never reduced the cabin temperature. At one point he increased it to 74'F, up from 72F.
- Despite nearly running out of power at several points, Broder charged the vehicle less and less at each stop he made.
- Broder drove the vehicle back and forth in small parking lots at several points in an apparent attempt to drain the battery to the stopping point.
Computer logs aren't perfect. Sometimes they glitch, or report incorrect data. But after comparing Broder's own trip details against the data logs the Model S
kept, it's clear that very little matches up with Broder's story. If this were just a range issue, or a speed issue, or a charge capacity issue, the New York Times' account of these events would have more ground to stand on.
At this point, however, either Tesla's entire logging software is flawed and reports incorrect values across the board for unrelated subsystems, or Broder sabotaged his own coverage. As someone who has reviewed a great many products, it looks increasingly like the latter.
In fact, I think Broder himself has provided
the explanation (even as he denies wrongdoing). Two days ago, he wrote: "This evaluation was intended to demonstrate its practicality as a “normal use, no-compromise car, as Tesla markets it. Now that Tesla
is striving to be a mass-market automaker, it cannot realistically expect all 20,000 buyers a year (the Model S sales goal) to be electric-car acolytes who will plug in at every Walmart stop."
A statement like that comes with its own preconceived set of biases baked right in. I'm willing to bet that Broder charged the car for less time each stop because, with every stop, he was reminded of how long it takes to power up an electric vehicle vs. stopping to pump gas in a tank. It became tiresome. He may have justified his behavior by thinking: "Hey, a normal
family wouldn't put up with this, either. People just want to drive! I'm going to test this car the way people will want to use it!"
That's an important angle when covering an emerging new technology, particularly one that necessitates changes to a person's operating habits. But that's not the story Broder was hired to write
Absent an incredibly compelling explanation or clear proof that Tesla's logging software is flawed, Tesla has won this round. It caught a reviewer throwing a review and lying about the conditions under which the car was tested. I'm willing to believe that Tesla's range predictor might be inaccurate, but there's no way to reconcile the charge time differences and speedometer readings.