This doesn’t quite fit "stuff everyone should know" but these are some basics that won’t take too too long to quickly gloss over.
There are some technical definitions to GNP versus GDP which you should look up some day.
Everyone's expense is someone else's income. So economics is all about linkages. In a closed system, if there were $100 billion of spending then that would imply $100 billion of output. However things are more complicated since "spending" could be repairing depreciated capital (a recurring cost of roughly 10% per year --- source: Piketty) or investing in new projects that might not go anywhere. It is also complicated by residence and nationality.
balance of payments
Some countries (think Germany) export more than they import. Things need to
[[[treasury pic William ... England
A [state-istic] ((link)) made by a mercantile theorist to convince the Bank of England to do something.
Although it might not make sense on the basis of individual transactions at constantly adjusting prices, deals struck for incommensurable goods [^no two houses are alike, for example] on ad-hoc terms, for job purpose of statistics advanced nations try to compute "general price levels". This is a hairy business but that is the only way you can get one summary number out of a zillion transactions ((((how many?)))) that take place every year.
billion prices project
There is also a relatively unique price level at which currencies can be exchanged with each other. For example at abcefg the conversion between USD and JPY is 124¥:1USD. It did this over the last year:
That price level is not the right conversion to use to compare "general standards of living". Again this is not really a well defined concept because banh mi is better in Viet Nam by construction than anywhere else, and French food is better than English by culture. Houses in the South of France are more regulated in what you can and can't do with them. Etc. Nowhere is the same. But statisticians do at least try to make a better inter-country comparison than just using the spot rate. They compile Purchasing Power Parity numbers every few years which are meant to be used to make better cross-country comparisons. Read about it some time.
(The alternative is a hamburger index)
What are the three most important things to know from statistics? They're related.
Sums of flat random draws (a.k.a. "truly random") are not flat. They are mound-shaped. (the middle is more likely than the tails)
Check out John D Cook's post on various central limit theorems. Varying the assumptions you often get a similar story. In the stupidest guise the CLT is known as the "80/20 rule", which is a content-free religious statement you can see everywhere you look if you believe in it, or nowhere if you don't believe in it. Ignore 80/20.
The sum of equiprobable draws needs to be uncorrelated with each other. If two dice are taped together, then each of their rolls considered individually will be flat (even or equiprobable). But their sums would not be mound-shaped, because one die will not be cancelling the other. The dice need to be rolled separately and independently of each other.
(If you need to deal with "dependent random variables", just call a statistician (like me). link to R list. I've written about this elsewhere but let's just not get into it.) An example of correlated RV's could be home price defaults. If one homeowner is in trouble in a company town, it could be because the company is laying people off. Their neighbour's experience is not independent of theirs if they work for that same company: they are at risk of losing their job and defaulting as well.
Numbers in the real world [^not in mathematics] usually come with a ± error term. You can think of these like splotches.
[[gaussian pic]]
The number "isn't" exactly 92347. It is around 92347. Is it really close to 92347 or could it actually be quite different?
Here is a good example of (in my opinion) academic shenanigans from a prominent thinker/speaker using people's lack of understanding of splotches.
The books you will find about this will tell you about p-values and H₀ and α and β and Bayesian maximum-likelihood lalalalalala..... Just think of a splotch without a precise shape. It is like a range except the stated value is more likely the ends. Maybe the likelihood is tilted north or south; whatever. Don't worry about moments which are just a way of quantifying the tilt, spread, warp, of where the numbers is "likely" to be. Just know that it doesn't have to be an exact Gaussian shape but it might be roughly mound-shaped.
∑ f^k = 1/(1−f), for 0<fraction<1.
You can derive this yourself and should. Write out the terms and then draw a big box around them, call the box F, and then do fF−F.
(my blog post on it)
- DCF analysis videos
Don't spend too much time on the CAPM. It's contentious theory [Buffett link].
Sums of independent dice are Gaussian distributed. (My blog post on it.) So if the price evolution were totally random it would still be a bit mean centric.
Look at differences in stock prices to see how the daily change is usually bounded. Look at [[this data example]] to see the "discussion" over price of some security over a trading day. (At tick level.) [[histo]]
This tick data looks much more to me like I would imagine a bazaar in Tashkent, except electronic so everybody "goes to the market" with computers from their offices. Meaning more people can be there at once, which seems like a good thing.
[^Gaussian probability is roughly proportional ½ to the distance from the average. So 1/2, 1/4, 1/8. We measure in square-root of distance so these numbers become 0σ ∝ 1, 1σ ∝ 1/4, 2σ≈1/16 ]
require(quantmod); getSymbols('SLB') #Schlumberger
require(magrittr)
SLB %>% stem
SLB %>% log %>% stem
watch out also when people start talking too much about black scholes.
- heading toward the blue
- quant.se
- NP: wrong, wrong, wrong
Like I said there is a lot, lot, lot of F.U.D. in finance. People at least think they can become important by saying stuff that's over your head and you can't critique it. Just don't bother. All you need to know is that BSM is an approximation. In maths terms I would just say it provides a better coördinate system (IV's) for traders, than dollars or percents. IV is a better scale for options traders to have their discussion in; things make better sense to both sides, it essentially cuts out a calculation that both sides would have to do anyway. The idea that traders worship the BSM model the way academics do is … what do you call it when you accuse someone else of committing the sin you commit? Freudian Transference?
Options on the other hand are not so complex. Buying a call north of the spot means you are bullish on a stock and expect it to end north of where it is today. (Calls have a fixed expiration date.) You need to pay premium to the seller for them to sell you the right to call them in the future.
You can also sell options
[m0ar theta ken griffin boat] [scammy "retirement plan" guy]
I thought about paying my rent this way when I was poor ("Wait ... I can just have money? Hellz yea!"), but they don't let poor people insure stock market players. You need to have tens of thousands of dollars at least (more like tens of millions if you want to trade for a living), and the exchange asks you for margin every time a position moves against you. In other words the exchange is going to make sure you are good for the money in case it expires ITM before they let you insure someone. This is one of the things the exchange does to make sure people from both sides want to trade there. (I.e., not the boys in the schoolyard punching each other because they aren't actually good for their word.)
table of http://www.nasdaq.com/symbol/slb/option-chain
Let's look at the volumes. Where are people able to come to a disagreement?
ok, getting really far afield here. But software (esp web) is hot right now.
Libraries are more important than languages. A library is code that someone else has already written that you can just call. For example it takes a lot of code to TCP/IP/UDP/serve/client/parse a HTML GET request [link] that’s sent over the web. But a library like curl does it for you. curl is xyz lines long, it’s been used by a lot of people, it’s been edited by a lot of knowledgeable hackers. So just use that with a "scripting language" (a.k.a. an easy language) and in a few
lines you can do what you want to do. For example look at bastards book of ruby.