I agree this isn't as big of a deal as the stockmarket may imply, but this line bothers me:
"That equates to 1 vehicle fire for every 20 million miles driven, compared to 1 fire in over 100 million miles for Tesla. This means you are 5 times more likely to experience a fire in a conventional gasoline car than a Tesla!"
Americans drive an aggregate of 3 trillion miles every year, while Tesla drivers have done 100 million (and they don't cite this number; are they including test drives?). That's well over an order of magnitude difference. Plus, the average Tesla driver is currently probably a superior driver (if for no other reason than they have a brand new expensive car) and has taken better care of their car (since it's within 2-3 years old, tops). In theory, Teslas will eventually become more mainstream over the years -- resold, price drops, lower-end models, etc.
Again, I don't think their conclusion about Teslas being safer overall is wrong. However, their conclusion of the likelihood of a Tesla catching on fire seems off, and the exclamation mark makes this press release seem glib.
> Plus, the average Tesla driver is currently probably a superior driver
I can play the conjecture game too.
The average Tesla driver is probably a worse driver than the average gas-powered car driver. Why? Because they have the disposable income to purchase a Tesla, which means they're probably less careful with their possessions (as they can presumably afford to replace them), plus the Model S is a sports car so they're probably driving faster and more carefree than your average person.
Not that I really believe this, but come on, your argument is completely made up.
100 million miles worth of driving is not a small number. After four years with a ~100 mile/workday commute, I put somewhat less than 100k miles on my car. Multiplying that out, it would take me over 40 years of driving at the same pace to reach 1 million miles and therefore well over 4,000 years of driving to reach the number of miles driven by Tesla owners.
So if I drove that 100 mile/workday commute for 4,000 years, something like this might happen approximately once.
Incidentally, I have seen, during said commute, the remains of an ordinary car that caught on fire. There was nothing left but a burnt skeleton of a car that appeared to be made of ashes. I'd rather be in the Tesla while it was still on fire than the burnt wreck I saw.
I do not understand your comment. This is the risk, as observed, measured with respect to a single person driving heavily. In other words, most people should expect less risk than this. You are correct that a larger population should expect more risk: there have certainly been more accidents than this overall, and a battery fire is not the only thing to worry about.
But risks must be compared to the relevant alternatives. Certainly, those who arrange their life such that they do not drive at all are removing a significant risk, as I can personally attest to having encountered several drivers who were, frankly, suicidal. I speak specifically of the person who pulled into a turning lane to pass (when I was already in it, stopped), and gunned it driving straight at me. By no means should you think that I am diminishing the very real risk we engage in every day while driving.
But most people measure risk by how easily they can imagine a scenario. As such, I offer this for you to contrast with the photos in the article: https://www.google.com/search?q=burnt+out+car&tbm=isch Such things are not uncommon and I have, personally, witnessed such. While measuring risk according to one's memory is a useful cognitive shortcut in many situations, but with the way our news sensationalizes things by highlighting unusual events and under-reporting common ones, it is not a useful metric.
You cant just compare fires in petrol cars driven 3 trillion miles a year with fires in tesla cars driven 100 million miles a year, its just not very representative. You can either take for example BMW and look up how many fires they had for a new 5 series in 100M miles driven and or let Tesla owners drive 3 trillion miles and see how numbers compare.
Just like Elon did it, it sounds nice but doesnt make a lot of sense.
You are correct that we have a higher confidence in the fire rate for gasoline cars. You are not correct in implying that we have little idea of the rate of fires in Tesla cars. We don't have '1 event' worth of data, we have '100 million miles' worth of data. That aside, yes, we should update our estimates based on future data as it is virtually certain that the numbers will fluctuate. Actually, they already have, based on there being more miles driven and no new reported fires. But few people remember to update their confidence based on a lack of events.
Yes and that includes all brands, used cars, old cars etc. Like i said, compare that with similar new cars like the new 5 Series or Benz E Series and i would guess numbers would be much closer.
If you have that data, feel free to share it so that we can compare the risk. I do not feel particularly threatened by a 'once in 4,000 years of heavy driving' type of event on a personal level, though if there were any evidence of systemic risk, I might be more concerned, though I do not own a Tesla. I also assume that there is some non-zero risk of a more catastrophic accident, which will then cause further hand-wringing, but I expect it to occur in a crash that was unlikely to be survivable or for it to involve someone doing something completely moronic. Or quite possibly both things at once and then some.
On an individual level, I am far more worried about identifying and avoiding the disturbingly large number of idiotically suicidal drivers I have met. And I mean that very seriously: as soon as I identify someone driving erratically, I find some way to get away from them, whether it means forcing them to pass me, taking an unplanned exit, or what have you, I will make sure I put distance between us.
But I remain unconcerned about a single report of a small car fire when there are many worse ones every single year (if not every day...)
I agree with you. My disposable income is not enough to buy a Tesla, but it is enough for a MINI Cooper. I drive it like I stole it, to quote my mechanic.
I drive a Tesla. It's a much bigger car and a much more expensive car than my old prius. As such, I am terrified to drive it. I'm neither superior nor inferior of a driver as a result of purchasing and driving a tesla, but I'm certainly more afraid of loss than I used to be -- both the loss of money, and the loss of time I get to spend driving my Tesla should I be in an accident.
It's funny to hear someone say this. I just went from a Toyota Camry (a relatively big car) to Toyota Prius-C, and I have the opposite feelings. Doing a shoulder check is very hard with the Prius, because of the weird back of the car. I now as a result hesitate driving it a lot, instead opting to use my brother's Corolla.
I do use the Prius in the city though, I easily get 75+ mpg when supermiling.
I would counter with, "How dare you assume that all Tesla drivers have disposable income to throw around". I happen to be close to buying one and know a few other owners who are average middleish class people who normally would have bought a porsche or a tesla. I work in tech and as a result am paid pretty decent. Do I have disposable money (or my friends) to buy a Tesla Model S? No! However, given where I live and where I want to drive + the recharging stations, Tesla is a good option for me.
Do not assume just because someone is able to get enough money to buy something expensive,that they have tons of money. By that same logic, anyone who is able to buy a home is clearly flush with cash and able to throw it all around. Completely ridiculous.
The average yearly income in the US is $43,000. The base price of a Model S is $63,000. There is simply no way to afford a Tesla as an average, middle-class laborer without making massive tradeoffs in quality of life elsewhere. If you can afford a car that costs such an absurd amount of money, you are either living in squalor to afford it or you are not middle class. It's tiring to see people who are obviously the beneficiaries of economic privilege pretend as though they are "average." You are not, and to assert that you are demonstrates a fundamental ignorance about the difficulties that "average" people face.
As for the house argument, you and I both know that the two are totally incomparable. Putting a roof on your head is a much bigger priority than buying an electric supercar. People need to pay for shelter. They generally can't afford to pay a lot for luxury cars. If you argument is true, why isn't everyone buying $63,000 cars? (Not to mention the millions of people who can't even afford homes.)
Your posts demonstrates everything that is wrong with class stratification in America. I don't think you or your friends who elected to choose Teslas over Porches know what it means to be "average."
I know a local gardener/landscaper laborer in the SF bay area who makes $25 an hour in cash by working after hours. If a Tesla lease is $600 a month (the website says that, but it's based on some assumptions including where you live), he would have to work about five hours a week extra to afford a Model S.
That means this SF bay area laborer must give up a Saturday or Sunday afternoon every week, but it's hardly impossible "to afford" or a "massive tradeoff in quality of life." On the other hand, it's challenging to haul a few thousand pounds of mulch in a Tesla, I expect.
Ok, so if you commit tax evasion in an all-cash business, you too can work just an extra 1/2 week every month to lease a tesla, after factoring in the tax breaks EXCEPT... from my 5 minutes of research, the $7500 tax credit is not a refundable credit so in his all cash business if he's not at least paying $7500 a year in federal income taxes, he won't get the full benefit of that credit.
Oh, and the $600/mo assumes a $7100 down payment, which needless to say is far more than the average down payment on a car and is really more like what the average american puts down on a HOUSE. In other words: c'mon.
Not to mention the opportunity cost of that money. If you're making $25/hr in SF you do not have a high standard of living, especially when compared to your neighbors which, studies have shown, is exactly how we measure our own financial happiness.
In this case, the SF bay area fellow I know works five days a week for a landscape/gardening company and pays taxes on that income. ($25/hour cash is additional off-the-books income.) And, contrary to your research, California has a $10,000 incentive.
Though you're right about the down payment, so let's assume it will be a used Tesla and therefore cheaper. And you're right about the relative comparison point. But these were rough numbers, remember, and not intended to be a detailed argument as much as a refutation of the impossible "to afford" claim above.
Are you claiming that you have analyzed the Model S and found that due to money saved on gas it is financially cheaper than a conventional car over some timescale? If not, then I truly don't know what you're talking about.
I'm really confused by "average, middleish class people who normally would have bought a porsche or a tesla" - is that a typo? Did you mean "would not normally"? If not, your idea of "average middleish class people" is really oddly out of whack.
Many people by homes because of a calculation that owning a home is cheaper in the long run than paying rent over that same period of time. Many people buy homes not based on such a calculation and simply because they have disposable income and owning a home is nice. Many people also buy homes based on neither financial decision-making nor disposable income, and I seem to recall this having some negative impacts on our economy a few years back, which may make it a poor analogy to why it makes sense to buy a Tesla (or a porsche?!) without disposable income.
I work in technology and would be considered by some (not myself) a Linux expert. I am most certainly not upper class, but I am by no means lower class. I have several friends who also work in tech in similar roles for financial companies (like myself). One of them drives a BMV M5, one of them drives a used porsche 911 turbo. It isn't that a porsche or bmw is more or less expensive. It is something that they save up to buy for themselves. A normal car or a tesla both solve my needs, but a tesla is sure as hell a lot cooler for the sheer amount of tech and engineering. I don't really see where you think I claimed to see any money would be saved by a model S. I never alluded to anything of the sort. But I think this is OT and isn't really productive. Lets focus on tech here at HN and leave the flames for /., where they belong.
This is why politicians use "middle class" to describe who they're fighting for; it's a term that nearly everybody thinks of themselves as belonging to.
You certainly can buy a BMW or a Porsche if you make $50,000 a year, especially if you're single and living in a cheap part of the country. A used BMW 3-series will in many cases be less than a new Honda Odyssey minivan. But you'll be shopping on the used marketplace.
You are far, far better off than you think you are.
You'll probably see the posts saying this to you as condescending, but if you can even think of purchasing a Model S you're in a very exclusive group of people.
When people talk about a Silicon Valley "bubble," part of what they're talking about is that a Tesla driver, BMW M5 driver, and Porsche 911 driver can believe that they are "average middleish class people."
When one tiny sliver of society (web engineers) simultaneously wields so much influence (as web companies do) and is so cut off from the reality of America (as you apparently are), it doesn't bode well.
This is so true. I'm currently a student and am sort of terrified to join the workforce in this industry. I really have no desire to surround myself with people who are so ignorant of and inured to the way real fucking people live. As someone who spent the majority of their childhood in poverty and was homeless for an extended period of time, this sort of hubris is profoundly saddening. It's very upsetting how abundant this form of arrogance is in the tech industry.
Luckily, there's a vast, vast array of options other than going to Silicon Valley. Any city decent sized city around the globe will have options available.
Outside of a couple critical mass areas in the US -- I'm thinking SV, greater Boston, NYC (just because everyone there is in their own little bubble ;) ) -- it's just another job.
And if you're not in the US, I wouldn't even know where this sort of SV halo exists.
If you make >100k, I can assure you, you are not middle class. You are part of the "not rich enough" class. This class has a lot of money and can reasonably do anything they want, but compare themselves to the ultra rich and think "I'm not rich". The problem is they are always looking up and forget to look down.
I grew up in a ~30k household in the 90's. Something I would call lower middle class. Upper middle class is probably ~70k-100k.
Please tell me you're taking into account cost of living here.
$100k does not go very far in San Francisco and Manhattan. Nobody making $100k in those places could reasonably afford a Tesla without severe over-spending (which they're welcome to do if they want to--whatever floats your boat). If you view the middle class as owning detached single family homes with yards, you would be lower middle class at best in those cities at $100k. If your spouse also made $100k, you could afford the crappiest single family home that was still livable.
But it's a really good salary in Denver or Houston; you could definitely find a nice house to buy and afford comfortably, though it might have a bit of a commute. If your spouse also made $100k, you could afford a nice home in any neighborhood with no commute.
And it might nearly qualify you as wealthy in a small rural town of 15k people total, where you would never need to spend more than $150k on any house.
Here in SF, housing costs are significantly higher. Other costs are somewhat higher. But you know what isn't? Mass produced consumer goods for one. An xbox (or tesla) costs the same whether you make $100k in SF or $50k in middle America.
You know what else? Travel. Domestic and international travel is a LOT more affordable when you're making $100k in SF.
And a big one: Retirement. You sock away your SF salary for 40 years and then retire in a cheaper COL area with no state income tax.
The truth is, even here, $100k is a lot of money. It really is. It's not rich by any stretch, but it's certainly not average middle class living. It is frustrating that a couple making over $200k can hardly buy 1000sq ft house while staying inside 30% (of gross income) housing budget, a metric widely seen as "affordable housing." But still, on balance, that couple making $200k is doing very, very well.
And the truth is, there are a lot of engineers here that, with liquidity events (which might just mean vesting your RSU's in an already public company), are earning $200k a year. That literally puts you in the top 1% of wage earners. (But not in THE "1%" of course which is usually a term referring to net worth and not gross income. But still.)
Edit:
The math is different of course if/when you add children to the mix. But I deliberately left them out, because in reality, it's just not a common lifestyle choice in San Francisco.
I wasn't talking about people making $200k, though. I was talking about people making $100k, because that's the number cherry picked by the person I was replying to.
And, yes, of course I'm aware that a Model S is the same price no matter where you buy it, to a first approximation. The point is that housing is so expensive in the Bay Area that you can actually end up having less disposable income left over for your car than someone making $20k less than you on paper in a much cheaper city and in a state with lower (or no) income taxes (Washington, Texas, maybe a few others?). If you don't believe me, you need to just sit down with a calculator that will do taxes and everything, look up comparable rents, and do the math. You sound like you might be surprised.
"You sock away your SF salary for 40 years and then retire in a cheaper COL area with no state income tax."
Small correction -- state income tax in CA is a big deal even if you're saving for retirement. And once you retire, you presumably won't care what the state income tax is.
Retirement savings goes into tax advantaged accounts like 401ks and iras. This is pre tax income. You pay the taxes at the end, when you're drawing from your accounts during retirement.
Not necessarily; you might have chosen a Roth for individual retirement savings. The contribution limit on that type of account is effectively higher than a traditional IRA, since it's expressed in post-tax dollars - very useful for someone trying to stash away as much as possible. Once the money is in the account, it grows tax-free.
But for those people, taxes are immediate rather than deferred.
Honestly, from the sound of it, you just googled retirement plans. :)
Actually, Roth IRAs are limited to $5000 a year per person, and once your income reaches ~$110k you can't deposit into a Roth at all. The vast majority of private retirement dollars are saved in 401k accounts that have have a cap much higher than in a Roth. Three times higher.
And actually, if you do the math (or just go read the math) you'll see that there is very little difference in your total retirement funds whether you use a pre-tax or post-tax account. This is because in a pre tax account you have the advantage of earning capital gains on the IRS's dime. The wisest choice is of course to use both if you can. Though many many people living in this area, due to income restrictions, cannot.
Anyway, I intended my reply to you as just honestly informative. I'm happy that you now are informed, but there was no reason to say "uhh, i mean a roth. yeah. a roth."
You couldn't be further off the mark. I've contributed to a Roth IRA for years, so I know you're incorrect about the individual contribution limit (it's $5,500 now). I live in the Bay Area. I replied to your comment because I've spent a fair bit of time thinking about this exact situation, and you responded in a rude and patronizing manner. That's not why I come to HN.
My reply was in the context of the $100k earner mentioned earlier, who would not be restricted from contributing to a Roth. Our hypothetical young tech worker is going to expect his or her real salary (and tax rate) to go nowhere but up as the years roll on. It's not bizarre to imagine this person might choose Roth rather than traditional for individual savings, independent of any 401k they might be contributing to.
There is only one way to read your comment:
"Small correction -- state income tax in CA is a big deal even if you're saving for retirement. And once you retire, you presumably won't care what the state income tax is."
You didn't qualify that! You added a "small correction" that was anything but. And yes, you got me, I didn't realize that for 2013 (possibly 2012?) the Roth limit was increased to $5500. Still not much stacked against the $24,000 I saved in my tax deferred accounts last year.
Listen man, i'm not one for silly debates on the internet, so go ahead and have the last word. But If you're going to add a "small correction" that is not really a correction, you should expect some disagreement.
Perhaps what you really meant was "That's true, unless you chose a Roth IRA as your only form of retirement savings. In that case, current tax rates matter and future tax rates don't."
That still glosses over the fact that I'm talking generally and you're talking about a specific, somewhat rare scenario. But at least it's accurate. As it was, had I not replied to you, a casual reader would've read my post, then yours, and walked away assuming that retirement savings is taxed at your current tax rate, in your current state of residence, when in reality it's just not so.
Exactly. Also, California is the highest-taxed state in the republic, so adjust that pre-tax salary of $100K downward a bit more in real terms.
Consumer goods like cars and laptops and TVs will cost the same, to a first approximation, anywhere in the continental U.S. But in some coastal areas, you have real estate costs that are 20x that of cheaper areas with salaries/compensation packages that are generally not 20x higher. In central Pennsylvania towns the median household income, for instance, tends to be around $26K and the median house cost is $64K, meaning houses are 2.5X median income.
In Palo Alto, on the other hand, the median house cost is around $2.25M, and the median family income is around $160K, meaning houses are 13.75X median income. Part of this are older PA families who bought when land was cheap many decades ago and are grandfathered in with lower tax rates, but part of this is also people who are stretching their finances dangerously to be able to afford there. (Even though those $2.25M median homes are often smaller than the national average, measured by square footage.)
It's possible to make $100K in the SF bay area and be, in real terms, poorer than someone making half that in central Pennsylvania.
He was referring to the Hacker News demographic, where these places are not such an outlier, and whether or not living in a place like San Francisco puts you in an elite class you still could not reasonably afford both a nice place to live and a Tesla there on $100k, which was my original point.
But I don't really understand why living in an expensive urban area automatically puts one in an elite class. The standard of living is not much better than other places for one thing--what you gain in diversity is lost in housing conditions. Also, it's not like $100k wages there are super rare. You don't need to be a 4.0 Ivy League graduate to score a livable wage in San Francisco. That's the starting salary for a lot of professions in cities like that, not just software developers. Police officers in the SFPD can make as much as a Twitter engineer fresh out of Stanford: http://www.sf-police.org/index.aspx?page=1655
Cost of living is always a factor. I live in the bay area. The salary of a programmer is usually 6 figures here. Granted it's almost impossible to buy a home but if you are frugal with your rent ( for instance getting roommates or living further from your job or in a less hip neighborhood) you have a pretty large amount of disposable income. Teslas don't cost less in South Dakota than northern California. I think cost if living is sometimes overstated because besides housing I don't find other critical staples like food, insurance, utilities to differ appreciably by state. If you want to own THEN the situation changes dramatically.
I'm not really sure what you're getting at. Having to move far out and getting several roommates just to swing the payments on your Model S is not exactly my idea of comfortably affording the car.
I wouldn't buy a Tesla on $100k in North Dakota either. Maybe, if I felt like blowing through money, I'd get a BMW 3 series or Audi A4, both of which are still considerably cheaper than a Model S.
I'm not disagreeing with you re: the Tesla S is way out of reach of someone "only" making $100k.
I was just pointing out that besides housing (especially owning) the cost of almost everything in CA is the same if you move somewhere where the cost of living is lower yet the salaries elsewhere are significantly less. So there is an opportunity to save a lot of money if you are willing to compromise somewhat on your living conditions.
I suppose it depends on whether you think it's important to own your home and/or you mind having roommates. When I first moved to CA I had severe sticker shock re: the rents so I rented a room in nice houses with other young professionals for the first 5 or so years I lived here.
Wha? A low-end Prius is ~$20k. A low-end Tesla is ~$65k. The US median household income is ~$45k.
If you can afford to spend the average household's entire annual income on a car coolness upgrade, then yes, you have disposable income to throw around, and no, you're not middle class.
You're missing the point - he's not actually making that argument, he's showing how easy it is to make any argument if you back it up with anecdotal conjecture rather than data.
Exactly. I rather suspect that Tesla drivers are actually comparable to other drivers, rather than being demonstrably better or worse, but the only way to know would be to actually conduct a study on the subject.
They will be demographically better if they're older, simply due to experience. If you look at insurance, the ridiculously expensive demographic is the male <25-year-old segment. If tesla drivers are underrepresented here - likely, since they're a luxury car at the moment - then yes, it's pretty safe to say that tesla drivers are more experienced, safer drivers.
The page is as much about marketing (maybe more) as it is about safety or the incident. Especially, after the stock dropped, can you blame them?
That said, I'm not really sure why this is such big news. Accident happened due to debris on the road. No one was hurt. The car was severely damaged, but got the owner to safety.
Let's consider this for a second - Tesla Model S had a no-injury accident and made national news. 5 kids burned alive in a Sentra, and this barely stayed on local news for 3 days.
Tesla Model S is a phenomenal vehicle with an excellent, if short-ish, track record. It really annoys me that reporters go out of their way for sensationalist articles, when there are far worse tragedies to cover.
The reason it is such a big deal is of course that we're standing before a major disruption of the whole industry everybody is curious whether or not this new tech is viable or not.
> The page is as much about marketing (maybe more) as it is about safety or the incident.
What does that mean? Taking a common definition of "marketing" to be "expressing the value of a product to potential customers," any article that is about safety or about a recent product incident is also marketing. It sounds like you're using "marketing" as an ill-defined term that basically just means "something bad about the way a company publishes information."
"Bad" is subjective, when I personally think Marketing, I mean putting a spin on things and presenting it in a more favorable light for the company.
This is not a dry facts page, but a public relations communication that is intended to sell you on the fact that Tesla Model S is a safe vehicle. Nothing wrong with that and nothing "bad" as you seemed to imply.
"The nationwide driving statistics make this very clear: there are 150,000 car fires per year according to the National Fire Protection Association, and Americans drive about 3 trillion miles per year according to the Department of Transportation. That equates to 1 vehicle fire for every 20 million miles driven, compared to 1 fire in over 100 million miles for Tesla. This means you are 5 times more likely to experience a fire in a conventional gasoline car than a Tesla!"
Facts stated:
* 150,000 car fires per year according to the National Fire Protection Association.
* Americans drive about 3 trillion miles per year according to the Department of Transportation.
* 1 fire in over 100 million miles for Tesla.
Sales pitch at the end: "This means you are 5 times more likely to experience a fire in a conventional gasoline car than a Tesla!".
The reason I call this a sales pitch or a marketing messaging, is that while the facts technically support this, as pointed out earlier, it's based on a data set of 1 occurrence.
Personally, I buy Musk's argument and if I was in the market for a new car, figuring out how to pay for Tesla would actually be at the top of the list for me. This story only re-enforces my own personal opinion that these are great cars.
"1 occurence" is misleading, as the miles driven is the actual random event. If you throw a coin 100 million times and 1 time it lands on its side - then you can have a rather reasonable estimate on how rare that possibility is even though you have just 1 occurrence. I.e., it may be that the chance is 0.5 per 100m or 2 per 100m; but it'd be very strange if the actual chance is 10 per 100m.
I think this is Elon's one blind spot. For a engineer he's awfully "salesy" at times. The whole lease thing was really distasteful. This less so; but still a bit awkward. Why not just end with "Fires are unfortunately a risk for electric and gasoline cars. Each year in the US there are more than 150,000 fires. As more and more Tesla's enter the world we're going to experience all manner of events" That would be the right way to put this in context. However, I suspect that to overcome the naysayers and truly change a market (electric cars) requires an almost insane love for what you do - which makes you sound a little too defensive when something goes wrong.
I have no idea how Tesla operates internally, but I wouldn't be too surprised if, like at most companies, letters that go out under the CEO's name are vetted by both the legal and PR departments, and possibly written or edited by them. If that's the case, then "Elon Musk" is a sort of composite entity, which includes a human named "Elon Musk", but not solely that human.
I would agree. But there's nothing in here that's cause for concern legally. And PR folks are going to have a tough time overruling Elon based on "style" points. As I mentioned he probably sees nothing wrong with his approach and thus any PR comments would fall on deaf ears.
It's unfortunate that our current media climate requires it, but a straightforward, "accept the inherent risks" approach to the problem would be massacred by news outlets. Were Musk to say something like this, headlines could read "Tesla's Billionaire Owner Disregards Model S Fire Hazzard" and not suffer libel.
Particularly for people such as Musk, who are doing so much to advance the expectations of consumers and raise the bar for what is possible with new technology (and here I'm speaking as an un-abashed Musk supporter), I can't help but support his rigorous defense of his projects. Whether it's the NYT "review" that attempted to slam the Model S mile range or the astonishingly good crash ratings for the same car, I think it serves us all to let Musk use his soapbox to be an advocate for his work at Tesla, SpaceX, etc.
> Americans drive an aggregate of 3 trillion miles, while Tesla drivers have done 100 million. That's well over an order of magnitude difference.
The data on 100 million is probably enough to compare with the 3 trillion miles. The populations don't need to be equal to compare them, just big enough that they are random and distributed enough.
When trying to get popular opinion for, say, an election -- sure.
However, there was one fire over 100 million miles. The problem isn't the 100 million, it's the 1 fire. This wasn't a controlled experiment, either -- they just stopped the clock as soon as the first fire happened, and multiplied. A week ago, they could have said "You have exactly a 0% chance of your Tesla catching on fire" and have been right by this logic.
To think of it another way -- let's say you get lucky and get a hole-in-one your 10th time golfing. Does that mean you'll have 10 hole-in-ones if you golf 100 times? Doubtful.
EDIT: Also, don't forget that Elon is mixing numbers. There's no fire if someone doesn't run over something. All these numbers show is that the average driver is more likely to run over something. Of course Tesla drivers run over fewer things -- there are no 16 year old kids texting while driving a Tesla... yet.
Forgetting about all of the (completely legitimate) concerns with this comparison aside from statistics, we can get a general idea from simple statistics how convincing having the first crash at 100 million miles should be.
For the purpose of our simple modeling, suppose that there is a constant risk per mile of the car catching fire, making an exponential model reasonable. Under this model, observing the first fire at 100 million miles would give a 95% confidence bound on the rate of fires of about one fire every 33 million miles.
If we're comfortable with the stated rate of about one fire every 20 million miles for other cars, then this would give a 95% confidence upper bound on the Tesla's rate of fires at about 60% of a normal car's rate. This isn't the 20% that Elon's statement would imply, but it does suggest a difference (which could just be due to other problems with the comparison).
It probably penalizes them, assuming the no-fires-for-100m-miles was not a fluke. There's no way to know for sure without knowing the true distribution of fires per mile.
By your "hole in one" argument, Tesla is "doubtful" to keep up their pace of one fire per 100 million miles. In other words, their "true" fire rate is actually less than one per 100 million.
I think a Poisson distribution is what you're looking for here. Roughly, if events happen independently and with a fixed probability per time interval, you get a Poisson distribution. Poisson distributions apply to a lot of things, so it's a very useful distribution to know about.
But since the number of cars is increasing, it's not a Poisson distribution; if the chance per car per time is constant, you'd expect the time to the next fire to be shorter.
Presumably it will go up to some degree over time as the average age of tesla's fleet converges with the industry average, as older cars are more likely to experience malfunctions/breakdowns for obvious reasons.
The median car in the US is ~11.6 years old[1], while Tesla's oldest vehicles were released in 2008, and the vast majority of their fleet was sold in the last couple years.
Obviously, the massive differences between electric powered cars and internal combustions engines means that they may never reach parity, but unless Musk has figured out a way to beat entropy, its pretty safe to assume that older cars will break down/suffer leaks/explode more than newer ones.
Did you consider the possibility that it is in fact you who don't understand?
The Law of Large Numbers states that as more miles are traveled, the fires per mile will approach the expected value. It is entirely possible to have 10 fires in the next week.
We won't know what the expected fires/mile is until a much larger sample is collected. It will take years to prove out.
Why do you believe a larger sample is needed? At what sample size would you believe that the Tesla averages less fires than other cars?
Here's a question (for anyone in this thread arguing statistics) that has an actual numerical answer: given the information in the article, what is the probability that Tesla's indeed experience less fires per mile than other cars? If someone doesn't know how to calculate the answer to that question, he shouldn't be arguing here.
Let's see if I remember any of this. If we assume the null hypothesis that Teslas have 1 fire per 20 million miles same as other cars, then P(0 fires in 100 million miles) = 0.67% and P(1 fire in 100 million miles) = 3.4% from Poisson distribution. So the odds that you'd have no more than 1 fire in 100 million miles is 4%. So I reject the null hypothesis with a p value of 0.04. (edit: fixed values)
This seems a bit dodgy since I'm "designing the experiment" after the fact, but I'm not sure how to correct for that. Any Bayesian experts?
You need an exhauseted state space. You cannot empirically infer a legitimate probabliliy, eg n/100m miles) with only a single failure observation, if there are 100 possible ways to fail. At best you have data on (1) of (N) ways to fail, but surely in the case of car accidents N=large.
A total of 2,650 cars were delivered to retail customers in North America during 2012, 4,900 during the first quarter of 2013, and 5,150 during the second quarter of 2013
Assuming 13000 cars on the road, each car would have logged 9k miles to get 110m road miles, as quoted by Tesla. But we know from past industry experience, that road fires are proportionate also with fleet age.
So, if anything we the probability of a road fire is likely to go up as more failure modes are discovered (including by chance), and as the vehicles cycle through a normal working life.
Assume a Tesla vehicle has a constant risk of catching fire per mile. That is, we have a exponential distribution `P(catch fire after t miles | hazard rate) = P(t|a) = a exp(-at)` where `a` is the hazard rate (average fires per mile). Furthermore we'll assume an exponential prior on `a`: `P(a) = w exp(-wa)`. `w` is a parameter that expresses how much prior knowledge we have of `a`. In the limit `w=0` we know nothing at all, except that it's nonnegative.
Our data is the fact that we went 100 million miles before a fire, after which exactly one fire happened, so we want to find the distribution `P(a|t = 100 million)` which tell us everything we want to know about `a`.
Then use Bayes' theorem: `P(a|t) = P(t|a) P(a) / P(t) = aw exp(-a(t+w)) / P(t)`. The normalization factor `P(t)` involves an integral over `P(t|a) P(a) da` from 0 to ∞, which wolfram alpha tells me evaluates as w / (t+w)^2.
So our posterior probability is `P(a|t) = a (w+t)^2 exp(-a(t+w))`, but we can take the limit `w -> 0` at this point for a fully uninformative prior: `P(a|t) = a t^2 exp(-at)`.
So we can just set `t=100e6 miles`, and now calculate things like the expectation of the distribution: `E[a] = 2/t = 2e-8 per mile`. Or the probability that the hazard rate is less than other cars, which is the integral from 0 to 1/(20 million miles): `P(a < b) = 1 - exp(-bt) (bt + 1) = 0.96`.
If I get the math right (assuming a Poisson distribution) then a week ago, based on 0 fires, they could've said with 90% confidence "The chance of your Tesla catching on fire is less than 2.3 cases per 100 million miles". Not 0; but not very high as well.
The relevant metric here seems to be 'miles between fire events'. On this metric, we have exactly 1 data point for Tesla. I would hardly call that "probably enough".
Of course, I'm exaggerating in the other direction. What we really should be calculating is the odds that Teslas burst into flames less often than the average car, given that the average car does so every 20 million miles and the first such event in a Tesla was at the 100 million-mile mark. We're still failing to account for the fact that the average Tesla is newer and probably better-kept than the average car, but it would at least be a reasonable start.
I don't know enough statistics to perform this calculation, but I would like to see how it is done.
An easy starting point would be to get this kind of data on similarly aged cars - cars sold over the last 2-3 years, then see what stats you can get there. I'd guess that you'd find some models that have never caught fire and some that have done so a lot more than tesla's, but that's pure speculation on my part.
We have enough events, because in this case it is appropriate to model the actual random event as 'million miles driven' with a chance of fire happening or not happening. Gasoline cars have a mean of 0.05 fires per million miles, and given the current Tesla data, the mean is 0.01 fires per million miles. I'm not taking out a calculator, but it would come out to an extremely low (0.0001%) the 'true' fire chance is the 0.05 gas car rate or higher; the 95% confidence interval should be 0.01 +- 0.02 or tighter, so still twice better than gas cars.
For an exaggerated example, if Tesla had driven a billion miles and had 0 fires, you shouldn't say that there's not enough data - you definitely would have enough data to say that the chance of fire is below the gas-car rate of 5 fires per 100 million miles.
Why is million miles driven a better way to model it than billion miles driven, for example? If you happen to choose that, there clearly isn't anywhere near enough data.
I'm honestly curious how one models this type of thing statistically, and I am not convinced enough of its obviousness to just accept numbers that someone throws around.
To put it in perspective, the disparity is equivalent to polling 10K americans and extrapolating to all of america (which, for better or for worse, is what most pollsters do).
It's counterintuitive, but the sample size needed for a good measurement doesn't much depend on the size of the overall population. What matters is getting a properly random sample. This is where pollsters fall down, because their "random" sample tends to be heavily biased toward the sort of person who has a landline telephone and doesn't hang up on pollsters.
Polling 10,000 Americans would be vast overkill, in any case.
You say that as if 10k is too small a sample size to extrapolate accurately with, but actually that's a huge sample size that if done properly would be extremely accurate. You don't need to poll anywhere near 10k people to accurately predict all Americans views.
You can sample less than 2000 people and get 99% accuracy with a 3% margin of error for a population of 325 million. Increasing the sample size to 10k simply reduces the margin of error to 1.29%, hardly worth the extra sampling of 8k people.
Do they really? My only interaction with pollsters has been either having them call me or solicit me on the street, and both approaches have a tremendous inherent bias. As far as I know, this is how the big national agencies do things.
There are all kinds of adjustments done afterwards to correct for various factors. For example, you know the age distributions in USA; and if you find out your phone calls are getting twice as many seniors than the proportion should be[1], then you throw away a random parts of them so that they don't skew final 'data' towards the typical opinions of seniors.
[1] Assuming that you're not measuring average age or measuring 'who is at home', but if you want to see, say, the average political opinion of total USA population, which tends to correlate with age.
It still seems impossible to correct for everything. Sure, you could correct for age as you describe, but I imagine that landline phone ownership correlates with political opinions in all sorts of other ways too.
Furthermore, how do you gather the data needed to correct the polling numbers without being able to accurately poll people in the first place? Seems like a complete chicken-and-egg problem.
"how do you gather the data needed to correct the polling numbers" -> you use the census. You need some info about the total population, you get it periodically and it doesn't change that much; you don't need to repeat it for every survey.
> Of course, the refrain I often hear is that as long as you pick the RIGHT 1,000 Americans, it's as good as polling all 319 million.
If it's a proper random sample, then it's far better than sampling all 319 million because it's 95% accurate with about a 3% margin of error and vastly cheaper and actually practical; you can't poll 319 million people.
I'm a big fan of Elon but I think he may be way off with these numbers... :S
If you read the reports from the National Fire Protection Association they do estimate the number of "highway vehicle fires" at around 150 000 per year [1], but they also explicitly say that 'the term “highway vehicle fires” is used to describe the type of vehicle, not the location of the fire' [2].
If you look at fires caused by collision or overturn (which I guess this would be classified as) that's only 4% of the total, or 6000 per year.
I bet you could get the number lower if you only counted fires caused by collision or overturn of a highway vehicle less than 3 years old. Or heck, just the black ones.
Yes, this fire was caused by a collision, but fires caused by other things are relevant. If a Tesla car caught fire in somebody's driveway, people would be concerned about that too, right? So comparing 150,000 highway vehicle fires to 1 Tesla fire, regardless of cause, is pretty reasonable.
In considering whether to buy a Tesla, I would evaluate the safety of the alternatives. Most in the market for a Tesla aren't going to buy the 23 year old Buick LeSabre whose oil leaks caught fire on the Jersey turnpike last week. So I think it's reasonable to compare the safety with a modern luxury sedan. I don't think as many of those catch fire without a collision.
According to some quick math, then, Teslas are 9x more likely to catch on fire from a collision than the average US car (1 / 5K) / (6K / 254M). I put the Tesla fire rate at one in 5000 per year because the average Tesla has been on the road for around half a year, and there are 10,000 of them.
I agree...The kind of driver who can buy a Tesla is likely the kind of driver who has acquired enough education and comfortable wealth to be a competent driver (on the average). Moreover, there's survivor bias here: If you've made it far enough in life to buy a Tesla, you're probably a person with pretty decent habits.
And finally, if you've bought an expensive car like the Tesla, you may be more likely to drive it more carefully than the average person does their Honda Accord.
I don't think Musk is any more dishonest than any other CEO, and he's probably more honest, by far, on average. However, it kind of pains me to see how easy it is for him to sway the hacker crowd with data-interpretations that would be questionable by any standard. If this is among the strongest empirical evidence he can provide, then I think we should maintain some skepticism.
A study by UC Berkeley and the University of Toronto actually showed that luxury car owners were by far the most aggressive drivers, and the least likely to stop for other road users.
For whatever reason BMW stood out most of all as the worst of the worst in this study, but 'expensive car drivers are better drivers' has been proven to not be an assertion you can make from common sense or intuition.
Without reading the study, I'm ready to agree that its conclusions are correct. But I'd argue we're still comparing apples to oranges here:
1. We can't assume that Tesla drivers are the same as traditional luxury car drivers. While cars are a statement of wealth, and many arrogant bastards like making that statement...A Tesla is ostensibly a statement of a few other things, such as concern for the environment, optimism about technology, etc. etc...in the same way, even though many programmers make high-end salaries (low-6 figures), I wouldn't say that we can expect these programmers to golf as much as the average non-programmer low-6-figure-earner.
2. Reading the link you posted, it said the study evaluated based on the frequency of accidents and the size of payout...I couldn't find the study on the website, but this statement from the article was ambiguous to me:
> According to data from IIHS website, the collision insurance losses on BMW 7 Series are more than twice the average for vehicles nationwide, and BMW 3 Series two-doors are more than three times.
Twice the average of what...? Payouts total? Payouts per claim? Payouts per claim normalized by number of total cars of that kind on the road?
3. The NHTS data concerns the number of fires. For the study you cited to be more relevant, we'd have to know what proportion of the insurance claims also involve vehicle fires. If you're thinking, "Well, if a car caught on fire, then that's likely because of an accident...so more accidents mean more fires"...well, a> That's a classic logical fallacy (All dogs are animals. Sara owns a lot of animals. Therefore, Sara owns a lot of dogs) and b> the kinds of accidents that make up a majority of insurance claims (my guess is low-to-medium speed fender benders) may not represent a significant portion of incidents that end up in a flaming wreck.
All in all, I wouldn't say that Musk is clearly wrong here...He could be right, we just don't have enough information either way...so it feels slightly dirty -- a slap in the face of analytical thinking -- for him to cite these statistics as being anywhere near conclusive.
> I agree this isn't as big of a deal as the stockmarket may imply
I firmly believe that what's happening on the stockmarket and the fire are merely a coincidence, there isn't a causal relation between them. Tesla stock has been due to correct back for a while, and a downgrade came in as well (unrelated to the fire).
Edit: first rule of the HN club: only positive things about Tesla, otherwise you are going to be voted down. Neutral opinion is considered negative. :)
I think it's lamentable that we're all looking at the bulk, "brute-force" statistical reasoning instead of considering the details of the case and what that means for the fire safety of the car. The line you quoted was incidental, rather than necessary, to the substance of Musk's argument. Whether it's PR or not is ignoratio elenchi.
If the battery really was punctured by an unfortunately placed piece of metal in a sort of surprising scenario, we ought to stratify by collision type in order to really understand the problem. How do gas cars perform during a gas tank puncture? At similar speeds? And how common are such collisions? The relevance of this consideration is expressed by Simpson's paradox:
I.e. analyses of performance when conditions differ ought to take into account those conditions, at least when they are both a: known (true in this case) and b: not causally connected to the variable under consideration (also true in this case).
A good point. It would be interesting to compare the miles drive on "brand new expensive cars" not just all cars which include obviously older and/or poorly maintained vehicles.
Why would that be relevant? Anyone who buys a Tesla Model S will instantly be in the category of "people who buy brand new expensive cars," so the Model S stats are sufficient.
That's the point. larrys wants to compare the Model S miles driven by "people who buy brand new expensive cars" against other car miles driven by "people who buy brand new expensive cars," instead of comparing the Model S miles [...] against other car miles driven by everyone.
This comparison is a bit off. Most of the cars which catch fire are old. The comparison should be to a similarly priced, brand new luxury car. Very few of those catch fire, either.
I believe your reasoning is incorrect. If we model the number of fires by a poisson distribution (very reasonable), a single sample would correspond to the mode of the distribution. What you would like is an estimate of the mean, i.e. lambda. It so happens that the mode of a poisson distribution is floor(lambda) or ceiling(lambda)-1, and thus it is an underestimate of the mean. In other words, fires are likely to be less frequent than what a single sample suggests.
Mr Market overreacts to bad news, and when a value investor buys.
Agreed, there's insufficient data for meaningful statistics, but his main points were amour-plating, firewalls and that batteries are less fire-prone than "highly flammable liquid".
Apples to apples: how many car fires are started because the car is old and not in good repair? I'll bet the rate is much lower for gasoline cars less than a year old. We'll just have to wait to see how 10-year-old Teslas hold up.
I dislike that statement, too, but for different reasons.
The 30,000x larger sample in the aggregate means the calculated fire rate for non-Tesla cars is probably a more reliable number. Maybe those 100M Tesla miles were "lucky"? Possibly not, though - I don't know enough about cars to guess.
The 3 trillion aggregate miles includes a lot broader sample size of car age than those from a single company that's 10 years old. This, to me, seems important to control for.
Finally - I bet the usage profile of a Tesla is different from the median usage profile of every random car on the street and throughout the country.
"That equates to 1 vehicle fire for every 20 million miles driven, compared to 1 fire in over 100 million miles for Tesla. This means you are 5 times more likely to experience a fire in a conventional gasoline car than a Tesla!"
Americans drive an aggregate of 3 trillion miles every year, while Tesla drivers have done 100 million (and they don't cite this number; are they including test drives?). That's well over an order of magnitude difference. Plus, the average Tesla driver is currently probably a superior driver (if for no other reason than they have a brand new expensive car) and has taken better care of their car (since it's within 2-3 years old, tops). In theory, Teslas will eventually become more mainstream over the years -- resold, price drops, lower-end models, etc.
Again, I don't think their conclusion about Teslas being safer overall is wrong. However, their conclusion of the likelihood of a Tesla catching on fire seems off, and the exclamation mark makes this press release seem glib.