Cap'n Refsmmat

You're not going to be able to download all the data unless the people who built the website specifically added features to do so. Otherwise you have to scrape every web page and extract the data the hard way.

Airbags usually use chemical gas generators instead of pressurized tanks, since you don't want to keep a pressurized air tank in your car all the time. But yes, I feel like a > 200mph impact into the ground probably destroys the structure of the aircraft so much that the airbag can't be much help. Ejection seats would be great, but you wouldn't be able to use a laptop safely (it'd be blasted off your lap and into the guy behind you upon ejection) or move around much, and you'd have to be strapped in much more securely. And have an oxygen mask on at all times. And so on.

Sure, but even an enormous increase in naval and recreational fleets will only contribute another few microns to the total sea level.

Yes, but not by very much. xkcd's What If? has covered the opposite of this question: http://what-if.xkcd.com/33/ So all of the ships in the world contribute very little to the rise in sea level. Read the full article for all the fun details.

Admittedly our software makes the block layout uglier than it has to be. Instead of letting me choose the spacing between paragraphs to be narrower, it merely inserts a blank line between each, so I can't control the gap. If I could I would make it more subtle.

Yep, that's it. 30 posts and you can then give out up to 3 negative points a day (and 25 positive points).

Why insist on indenting, though? A blank space between paragraphs serves the same purpose. It wastes vertical space, but that is not in short supply on the Internet. On the topic of precision: I agree wholly with Ophiolite. Writing exists to convey meaning, and in science the meaning is often complex and counterintuitive. Only with precision and careful planning will your writing succeed. Recently I have been writing a book about the improper use of statistics, and many points center on a subtle misunderstanding of some tricky statistical concept. My editor frequently writes comments like "I didn't quite understand this paragraph," and when I read it again I realize I've said something entirely misleading. ("Power of the coin? What does that even mean?") Readers are endlessly imaginative misinterpreters of statistics, and I have to choose every word with incredible care to convey my meaning correctly. The tragedy is that few of my readers will read as carefully as my editor does, and so most will miss important points.

An interesting but conspiratorial point: even if you do believe that climate change is real, but do not think it is the government's role or duty to address it (or think the remedies are worse than the problem), pretending to be a moron is an incredibly effective obstruction tactic. So if someone believes that action on climate change would kill the American economy by excess regulation, then it's best for them to pretend climate change is a liberal myth.

Senior members (those with 30 posts or more) can give 25 upvotes and 3 downvotes in 24 hours. All others can give up to 10 upvotes and no downvotes in 24 hours. If you're really interested in reputation systems, this is an interesting paper: http://arxiv.org/abs/1405.1429

Yes, it is. Nice try, "Frank."

There are plenty of little devices made for measuring distance with sound. Here's one: https://www.sparkfun.com/products/639

I'm fine with you writing about your book on your blog, since it's a blog. On the forums, you can bring it up when it's relevant to discussion ("I explain this in more detail in my book"), but creating threads to advertise would clearly be over the line. It's a rather fuzzy boundary and I'm not sure where to draw the line. Just how much do you have to bring it up before it's advertising? I'm not sure. Perhaps some other moderators can weigh in. From a practical standpoint -- don't worry, we won't just ban you the instant we see a link. If a moderator does object, they may remove the link and send you a message. (I'm interested in what the mods think for my own reasons too. I'm also writing a book and will find it hard to resist bringing it up. I don't want to abuse my admin position to advertise it freely.)

The page you linked to is merely loading another page inside an <iframe> tag, so you can load that page directly instead: http://www.waynecounty.com/DotNetForms/SHRFInmSearch.aspx

So you're the one leaving organic raisin granola bars in the mod closet instead of resupplying us with Cheese Nips.

Yes! I would not be writing a book now if I had not spent my teenage years trying to imitate Sayonara's incredibly clear debate style.

You want to use a test of proportions: http://stattrek.com/hypothesis-test/proportion.aspx They're a very common need, so you can find them implemented in whatever statistical software you use. You can probably make an Excel formula to do it, even.

If you change your display name, that will change the name that appears on all of your posts. It will not, however, change the name you use to log in. "Current password" is the password you currently use to log in; you're required to enter it when setting a new password as a security measure. I do not know what "Local password" is; I don't see it in my account, since I'm an administrator and see different controls.

You can edit posts (with the Edit link in the lower right) for a few hours after you post them; after that, editing is no longer possible. You can't delete your own threads though.

I'm attempting to read Darrell Huff's unpublished manuscript How to Lie with Smoking Statistics, commissioned by the tobacco industry in the 60s to respond to the claim that smoking causes cancer. (Huff more famously wrote How to Lie with Statistics, probably the most popular statistics book ever written.) I managed to dig up most of the unpublished chapters but I'm trying to figure out why it wasn't published; as far as I can tell, he was ready to sign a contract and then nothing happened.

You need to accelerate protons to incredible energies, which requires running them around in a loop while you pump more energy into them. But to pull them into a loop you need to keep bending their paths, which takes energy. So you don't want to force them to make sharp bends. There's just no way of accelerating a proton to such a high speed in a short distance. You couldn't make an electromagnetic field strong enough to do it in a few inches. One current field of research is laser wakefield acceleration, where lasers are used to induce an incredibly strong electric field to accelerate electrons much faster than you could normally achieve: https://en.wikipedia.org/wiki/Laser-wakefield_acceleration

Sonobuoys are already commonly carried by military antisubmarine warfare aircraft: https://en.wikipedia.org/wiki/Sonobuoy The question is whether they listen on the right frequencies. A P-3 could cover a wide area with sonobuoys faster than a ship could trail a locator device, but perhaps the ship-mounted locator is more sensitive.

Indeed. A great deal of theoretical statistics is about quantifying the errors of different methods and determining which ones perform the best under different circumstances. There's been a lot of work on robustness, for instance, where estimators are checked to see how well they behave when their assumptions are not met.

I'm not sure what your argument is here. I can write the normal distribution in terms of the precision (1/variance), the standard deviation, or any other weird quantity I'd like; the original normal distribution was differently parametrized than the one we have now. So while the sample variance is a good way of estimating the variance parameter, we could equally well calculate the sample precision or sample standard deviation. But I also don't understand AdvRoboticsE529's point.

The nature of science is that statistical arguments are gradually replaced with exact mathematical ones -- when they can be. Kepler's orbital laws, for example, were basically an empirical and statistical argument from data, with no theoretical grounding. Once Isaac Newton formulated a theory of gravity, he was able to replace the empirical argument with a theoretical one and show why Kepler's laws had to be true. Statistics is a way of making progress in science even without the exact theory. Without statistics to give us a rough picture, we will not know what our final theory should look like On the other hand, my current research (I am getting a PhD in statistics) isn't something that can be replaced with exact formulas. I'm trying to map the background radiation levels over a wide area. Background radiation levels are a function of the radioisotopes buried in the ground and in common building materials, such as concrete. There's no a priori way to figure out where the radioisotopes are, unless you can devise math to predict exactly how the Earth would be shaped from four billion years ago to today, including construction and man-made activity. So I have to measure empirically, and to do that I need statistics.

Do you know about monoalphabetic ciphers and solving them via frequency tables? You can tell if a code is monoalphabetic just by looking at the frequency tables; with a quick statistical test you could figure out whether you're producing gibberish or not. But suppose that doesn't work. Let's think of another attack: the known-plaintext attack, where I know that, say, the end of your message is "Signed, Asterisk Propernoun" or whatever. (Military telegrams often ended with the equivalent of "STOP", for instance, or the name of the unit sending them.) Now, "Asterisk Propernoun" repeats some letters--o, e, and r, for instance. So I just have to convert your message into base n, then look at the last digits in that base and see if the 4th and 14th (e) are the same, and the 5th and 11th and 15th (r), and so on. Once I do this, I know your secret base, and I can begin to deduce the code for each character. Known-plaintext attacks were an extraordinarily popular historical method of breaking ciphers. You might be interested in the idea of stream ciphers, which try to approximate one-time pads: https://en.wikipedia.org/wiki/Stream_cipher