Hacker News Reader

Comment by symfoniq | Hacker News Reader

Moreover, when this sort of thing happens to a car that uses gasoline, the results can be far more catastrophic.

"The freak accident occurred about 10:30 a.m. on Nov. 8, 1994, when the family was traveling to Wisconsin to celebrate birthdays. The van, driven by Scott Willis and carrying his wife and six of their nine children, struck a piece of debris that had fallen off a truck driven by Ricardo Guzman. The debris, a mudflap/taillight assembly, punctured the mini-van's gas tank, and the car burst into flames.

Scott and Janet Willis managed to untangle themselves and escape, but the children were trapped and five of the six perished instantly. The oldest in the car, 12-year-old Ben, died the next day."

http://articles.chicagotribune.com/2000-05-01/news/000501016...

View on HN

This is possible with a gas car, but the gas tank doesn't occupy the whole bottom of the car and even if it is punctured, it won't ignite unless there is some additional spark or heat source.

Given the number of gas cars on the road, anything described as a "freak accident" must have a very low probability. If this is the 13th serious accident of a Tesla, it's unlikely that the probability is very low.

>If this is the 13th serious accident of a Tesla, it's unlikely that the probability is very low.

I'm not sure you've followed the thread here...I don't think anyone's (accurately) saying that this is the 13th spontaneous battery accident. I'm pretty sure the other 12 are standard traffic accidents, of which this is the first to produce a battery fire. All in all, it seems a rather small sample size to make any kind of determination.

Assuming that the normal rate of battery fires after collisions is one per million, the probability of observing such a fire after only 13 collisions is rather low. (Not impossible, just unlikely.) so perhaps the rate is not one per million. Doesn't mean it's one in thirteen, just not one in a million.

If the chance of a catastrophic event happening is one-in-a-million, then it is equally likely to happen on the 13th and on the millionth coin toss.

> Assuming that the normal rate of battery fires after collisions is one per million, the probability of observing such a fire after only 13 collisions is rather low.

What you're arguing is that the probability of it happening on or before the 13th time is less than it happening on or before the millionth time -- Which is self-evidently true of _any_ probability, and is a much weaker argument.

> perhaps the rate is not one per million

This is much closer to the argument you should be making. I.e. "Given that a purported one-in-a-million event occurred after only the 13th test, what is the probability that their estimate of one-in-a-million is accurate?"

(I don't know the math to answer that question, but I hope somebody that does will come along and chime in.)

Time for some fun with binomial probabilities!

Assuming I interpret the actual problem at hand correctly, this is how we formulate it statistically.

We have an event that occurs with probability p = 1/1000000, given an accident (of which there were 13).

The probability of having zero catastrophic events, given that we have observed 13 accidents, is (1-p)^13 = .99987

(For some real hair-raising fun, try this calculation with the success/failure rate of a typical condom.)

Note that this is a purely frequentist interpretation and ignores any Bayesian inference (ie, we assigning a weight of 0 to our Bayesian prior, which is atypical). It's not a very robust way of modeling the situation, for a number of reasons.

I think the right test to apply is the Wilson test.

http://what-if.xkcd.com/65/

Given this XKCD article, I would suggest the probability of a battery fire after an accident is around 1 in 26.

That estimate is a little bold: we can reject the hypothesis that the probability is smaller than 1/260 with p-value 0.05 (good confidence), but with your threshold of 1/26 the p-value becomes poor.

Note that due to the lack of selection bias (battery fire was already among the top security concerns before the accident), we don't actually need a very stringent p-value, so we can be fairly confident that the probability is greater than 1/260.

But my point was that even with probability 1/260, it's likely to be much higher than for gas cars, or you wouldn't have to dig up 20-year-old articles about "freak accidents".

> But my point was that even with probability 1/260, it's likely to be much higher than for gas cars, or you wouldn't have to dig up 20-year-old articles about "freak accidents".

Gas cars have electrical fires all the time. We don't read about it because it's not news.

Admittedly, those cars tend to be poorly maintained, and not nearly new high end expensive vehicles.

Even new cars are not exempt. Electrical fires are often a total loss even if the car looks ok otherwise because of the way the wiring looms are embedded in the chassis.

If 1/13 car crashes resulted in the car bursting into flames, that car would be recalled.

What is the threshold for a recall?

I'm not sure you've followed the thread. The previous poster claimed the sample size is sufficient:

> Statistically, there have been fewer fatalities with Model S than the average for the number of miles driven

Which of course it's not. But the general tone here of "move along nothing to see here" is transparently wishful thinking.

Of course a battery fire after hitting some road debris after only a dozen accidents is something to be concerned about. Maybe we don't know just how much yet but no one should be dismissing it offhandedly.

> Given the number of gas cars on the road

On the other hand, gas power cars have been on the road for a century now, manufacturers have had time to improve safety based, on accidents. Electric car manufacturer can improve from what is right now an excellent record (no explosion).

Let's also not forget that there are more than just Tesla cars out there with the bottom full of batteries: Toyota, Honda and GMC hybrid cars are in the same situation, so the sample isn't as small as you make it sound.

It's worth remembering that airplanes are only the safest way to travel because they put black boxes in the backs of planes to figure out why they were crashing constantly.

Trillions of miles later, we freak out if a plane's landing gear fails and the plane has to skid to an otherwise uneventful landing.

And the fact that pilots spend a lot mote time in training than drivers, and the fact that there are two pilots, and Air Traffic Controllers, and computers constantly checking the pilots input as well as the flight conditions, and ...

That helps prevent pilot errors. It doesn't help identify mechanical/software failures after the plane is in billions of tiny pieces.

Also, plains undergo near-constant diagnostics and preventative maintenance work.

Don't most new cars carry some kind of data recorder? If a failure is due to pilot error or lack of maintenance, we can at least use the data pulled from the melted, twisted wreck to rule out design flaws.

They do, but its built into the airbag system, and only commits the last 15-30 seconds of vehicle data when the airbag is triggered.

Tesla and GM vehicles have more sophisticated telematics than other car makers, and I'm sure Tesla is taking advantage of the data (battery pack performance, etc). You'll also get a call if your battery charge drops to a level that will damage your battery.

Yeah, gasoline is actually surprisingly difficult to accidentally ignite, wheras lithium ion and lithium polymer batteries self-ignite if punctured.

Diesel fuel is even more difficult (drop a lit match into a bucket of diesel and it will likely just go out.) For some reason most people don't choose their vehicle by the safety of the fuel source -- which isn't to say that people won't avoid a Tesla based on news like this.

You completely missed the point. How many people died when the tesla caught fire? This is a prime example of the tesla being safer, and people are using it to pretend it is less safe.

I'm not saying the Tesla is less safe: I'm saying that, contrary to what the parent post implied in his wording, it's almost certainly more likely for the battery to catch fire than for a gas tank.

The parent post did not imply anything of the sort. It was explicitly and clearly comparing the severity of fires, not the frequency.

There are plenty of things to catch on fire in a normal car. I've only own 3 cars and I've been in 4 car fires. (no one was hurt in any of them.) None of them involved the gas tank.

On this note I, during high school, worked at a large multiplex movie theater for about a year. During my shifts we had three car fires break out in parked, unoccupied cars. This was considered totally normal by both my employers and the fire department.

This stuff is clearly happening all the time.

This. A coworker's engine burst into flame for no apparent reason as he was commuting to work a few years back. Nobody died, but the car was a total loss.

I think this is relatively common, too. In my life I must have passed a dozen cars burning on the side of the road.

Gasoline car fires take seconds, not minutes and there are lots of places where things can go wrong, pretty much every junction of the gas line plumbing, the tank and the injection system.

Whereas Teslas are only vulnerable along the entire underside of the car. And don't require an ignition source.

Note that the fire was contained and that it did not enter the car interior.

Containing a gasoline fire after it has had a second or two to run is very hard if not impossible.

That's Tesla's PR spin. You'd probably be in pretty bad shape from the heat if you were trapped in that car. The guy taking the video remarked he could feel it just by driving by.

A key question here is how easy it is to start a fire from a relatively minor and common incident such as hitting road debris. Sure, it's possible in gas cars, and gas fires burn quickly.

But if it's a lot more common in cars with giant vulnerability zones along their entire undersides where shrapnel can start fires through physical damage without even an ignition source...

All I'm saying is this is definitely something we need to monitor closely, not just dismiss offhandedly or spin with clever PR like "the fire didn't enter the cabin (but still could have killed you)".

The battery pack is directly under the passenger seat, and this is typically a non-structural compnent, of very thin sheet metal. One would not want to be riding shotgun.

This is the 13th serious accident of a tesla, but I believe it's the first battery fire.

As the article notes, that is a rather infamous case that is a large part of the reason former Illinois governor George Ryan was convicted, as it uncovered a cash for passing grades on commercial driving tests scheme. It was a horrible tragedy.

Huh. I never knew that - thanks for pointing it out.

Equally interesting is that his successor [0] as governor was Rod Blagojevich, who also went to prison for corruption. That wikipedia article mentions that in the last 40+ years, 4 Illinois governors were charged, convicted, sentenced, and sent to prison for corruption or (in one case) fraud.

Quite a tradition they have there. Pat Quinn must be nervous.

----------

[0] http://en.wikipedia.org/wiki/George_Ryan

Reply on news.ycombinator.com