[arrrg]

There is nothing wrong with replication, it’s just that it’s wholly unable to clear anything up in this case.

The main important points of contention aren’t even disagreed over here. Those you could figure out by replication.

It’s the interpretation of those results that’s the sticking point here.

Some details that cannot be replicated are also disagreed on (heat, circling in parking lot, …), but replication won’t help you find out which is the correct interpretation.

[stcredzero]

> There is nothing wrong with replication, it’s just that it’s wholly unable to clear anything up in this case.

Sorry, but not only John Broder's intentions are in question here. If a bunch of other journalists replicate the trip, but come out with results more in line with Tesla's view, then this calls into question Broder's accuracy, intent or no.

You know, there's an empirical context underlying everything here.

[slantyyz]

>> If a bunch of other journalists replicate the trip

Is it really a replicable test? They'd need the same weather patterns, traffic conditions and the same car (untweaked or modified since Broder's drive), not a different, "equivalently equipped" model.

In the documentary Revenge of the Electric car, there's a scene where Musk walks into a warehouse full of cars, all having different issues preventing shipment. For all we know, there was a problem isolated to the particular car Broder drove AND, at the same time, Broder was embellishing his story. In other words, the truth is somewhere in the middle.

Of course, we'll never know because we'll never get access to the raw logs.

[stcredzero]

> Is it really a replicable test?

Yes.

> They'd need the same weather patterns, traffic conditions

Okay.

> and the same car

Find an honest 3rd party who already owns the same model.

> not a different, "equivalently equipped" model.

Now you're just being obtuse. Consumer Reports gets to test a car that's the same model. They're not required to come to your house and test your car!

> In the documentary Revenge of the Electric car, there's a scene where Musk walks into a warehouse full of cars, all having different issues preventing shipment.

That's a completely different car! They were taking Lotus bodies and fitting them out with batteries and motors, somewhat by hand. This one is built on a bona-fide assembly line and designed to be electric from the ground-up. Sorry, but this is a huge stretch.

> For all we know, there was a problem isolated to the particular car Broder drove AND, at the same time, Broder was embellishing his story. In other words, the truth is somewhere in the middle.

I'm sorry, but this tactic of thinking of all the confounding things you can, then throwing up your hands and saying "middle!" is the sort of thinking you expect on Fox News or on the playground. If replicability can't settle such debates, then science is in deep doodoo. However, we know from the history of science that it can settle debates 10X more acrimonious than this.

To be fair, you are right that the truth is probably somewhere in the middle. Exactly where is important and something we can get a good locality of through replicating.

[slantyyz]

>> Sorry, but this is a huge stretch.

Why? Even individual cars produced in assembly lines can have problems.

When I worked at Toyota, there was a problem with a Lexus model that was so hard to isolate. Eventually, after a long data collection period, they found that the problem was only existent in cars painted in a certain color.

[stcredzero]

> Why? Even individual cars produced in assembly lines can have problems.

So? If there's some conflicting results, we can then turn around and test the same car as well. If there's a problem, those are measurable.

> When I worked at Toyota, there was a problem with a Lexus model that was so hard to isolate. Eventually, after a long data collection period, they found that the problem was only existent in cars painted in a certain color.

Color and whatever other factors were measurable. Empiricism is possible. Your example supports my point.

[slantyyz]

>> Color and whatever other factors were measurable.

Sure, but it was such a rare problem that they tried to reproduce it using the same model and year and production run, assuming the problem would show up but it didn't. They didn't think colour would have been the identifying indicator of the problem until they had amassed enough data to lead them down that path.

Broder drove a car that has probably been used several times for review and other uses, and may have had more wear and tear and/or maintenance and service than a car sold to anyone else. It could also have been driven under stress conditions that a brand new model may not have been.

All I'm saying is that there's a lot about that car that could be different from other review cars. It could have been a freak occurrence and Tesla was unlucky enough that Broder was reviewing that car when it happened.

In any case, if you go to the Tesla owners forums talking about the NYT article, you'll also see actual owners who live on the east coast were not surprised by what happened to Broder (http://www.teslamotorsclub.com/showthread.php/13633-NYT-arti...).

[arrrg]

Not really, no.

The main questions that could be replicated are not at all disputed. The rest cannot be. Simple as that.

[stcredzero]

The "rest" doesn't matter, and a comparative measure of Mr. Broder's accuracy as a reporter can still be established.

[RyanZAG]

I don't really understand your point.

If someone takes the same car and follows the same route and gets the same results, then it's going to prove the reviewer correct. If repeating the trip by someone else who isn't trying to wreck the car gives different results, then that is very much a valid point here.

[arrrg]

What? I don’t think you quite understand the structure of the controversy here, what is disputed and what is not.