30 July 2014

Bombs Away!!

Completely silly of course, but this has to be read. I think I know some of those committed!

Bayes Rum and Coke

Markus Gesmann provides the most lucid, yet succinct, presentation of Bayes I can recall seeing. Doesn't convince me that Bayes has any justification in data analysis or data science or stats. But that's just me I guess. The data are, and that's the end of it. By the bye: his example is far more concrete than is typically used.

While I prattle on that policy trumps data all the time, Bayes provides a mechanism of injecting policy sub-rosa. Not my cup of tea. Look to incentives/policy if you want to predict the future. It was policy, not data, which generated The Great Recession, and stats folks (well, 99.44% of them anyway) who blindly believed that "yesterday looked pretty much like the day before, today looks like yesterday, so tomorrow will look like today". The inherent folly of MCMC. In Nature, where God imposes immutable rules of behavior, that's reasonable. In human Nature, where some humans get to change the rules to suit themselves whenever they feel like it, the resulting (historical) data is always suspect. That data stream, after all, wiggles around in response not to just perturbations of Mother Nature, but to the whims of human rule makers. There's not the slightest reason to assert that this combinations of influence will continue as-is in the future.

By coincidence, the Gesman piece references a Charpentier presentation, and I'm considering musing on a recent posting by him. One that's not overtly Bayesian, as it happens. We'll see.


28 July 2014

How to Properly Set The Table

A continuing sub-theme of these endeavors is the growing innumeracy of both the general population, and more significantly, the developer community. Two recent essays pick up the theme.

First, we have a PG developer, who was once a pure coder by his own admission, advocating against loops. Froot loops? Too much sugar. Now, whether one needs to use a host language (python, here) rather than the engine's data loading command (syntax peculiar to each engine, since it's not SQL) is another issue. But looping? Nah.

Second, of much greater importance, is a NYT Magazine piece on "new math". I read the dead trees version yesterday, which was titled, " (New Math) - (New Teaching) = Failure ", while the web version is titled, "Why Do Americans Stink at Math?". The dead trees title is more descriptive of the text.

As mentioned a few times in the course of these endeavors, I was a guinea pig for at least two previous "new math" initiatives, back in the dark ages of the 1960s: SMSG (which we alliterated, Some Math Some Garbage, but was officially School Mathematics Study Group) which was an implementation of New Math, per se and later (for me) the TutorText approach.

If you follow the link to the New Math essay, you'll see complaints from that era which are nearly verbatim to current complaints voiced by opponents to Common Core today. If I didn't know better (and I might not), I'd conclude that the opposition derives from religious refusal. "The Bible doesn't teach math, so why should my kids have to know it?" The opposition is strongest in Red Religious States where the earth is 6,000 years old, after all.
One 1965 Peanuts cartoon depicts the young blond-haired Sally struggling to understand her new-math assignment: "Sets . . . one to one matching . . . equivalent sets . . . sets of one . . . sets of two . . . renaming two. . . ." After persisting for three valiant frames, she throws back her head and bursts into tears: "All I want to know is, how much is two and two?"

The issue, as it was in the 1960s, is two fold. First, teaching arithmetic to ten year and younger juveniles is fraught with danger. Adults, whether curriculum developers, mathmeticians, text book authors, or teachers, simply don't remember why learning math-y stuff was tough at that age. We're still only guessing. Sally is right. But, Sally is also wrong. The issue is how soon to move from rote memorization (2 + 2 = 4, and don't forget it) to understanding how addition works.
And yet, once again, the reforms have arrived without any good system for helping teachers learn to teach them. Responding to a recent survey by Education Week, teachers said they had typically spent fewer than four days in Common Core training, and that included training for the language-arts standards as well as the math.

"Why Johnny Can't Add: The Failure of the New Math". That is from 1974, toward the tail end of the first New Math excursion. Especially, read the comments. There aren't many, and some are awfully cynical.
But our innumeracy isn't inevitable. In the 1970s and the 1980s, cognitive scientists studied a population known as the unschooled, people with little or no formal education. Observing workers at a Baltimore dairy factory in the '80s, the psychologist Sylvia Scribner noted that even basic tasks required an extensive amount of math. For instance, many of the workers charged with loading quarts and gallons of milk into crates had no more than a sixth-grade education. But they were able to do math, in order to assemble their loads efficiently, that was "equivalent to shifting between different base systems of numbers." Throughout these mental calculations, errors were "virtually nonexistent." And yet when these workers were out sick and the dairy's better-educated office workers filled in for them, productivity declined.
(Emphasis mine.)

Remind any of you, if only distantly, of sub-prime mortgages and London Whales? It should.

I'll close with the punch line:
One of the most vivid arithmetic failings displayed by Americans occurred in the early 1980s, when the A&W restaurant chain released a new hamburger to rival the McDonald's Quarter Pounder. With a third-pound of beef, the A&W burger had more meat than the Quarter Pounder; in taste tests, customers preferred A&W's burger. And it was less expensive. A lavish A&W television and radio marketing campaign cited these benefits. Yet instead of leaping at the great value, customers snubbed it.

Only when the company held customer focus groups did it become clear why. The Third Pounder presented the American public with a test in fractions. And we failed. Misunderstanding the value of one-third, customers believed they were being overcharged. Why, they asked the researchers, should they pay the same amount for a third of a pound of meat as they did for a quarter-pound of meat at McDonald's. The "4" in "¼" larger than the "3" in "⅓" led them astray.
(Again, emphasis mine.)

22 July 2014

I Still Hate Neil Irwin

The morning, during the week, begins with a reading of briefing.com, via Yahoo! Finance, then a toddle off to the newstand for The Times and thence to the coffee place to inhale both. Most often, these toddles involve creating the rough draft of one of these missives. Since I don't put one up every day (it just seems like, at times), most never get past that stage.

Today's puzzlement: why do the mainstream pundits remain puzzled that interest rates remain low? It's as if, and likely so, that none of them ever read up any physics or chemistry. They view returns on investment in purely fiduciary terms, a fact which goes a long way to explain why few saw The Great Recession coming. That Giant Pool of Money was sitting out there in Y2K, and Masters of the World in the C-suites had no use for it. They simply couldn't figure out profitable, enough, ways to turn moolah into machines. So, the banksters ginned up fiduciary instruments in sufficient quantity to sop up the Pool. The Pool hasn't gone away, and the MotW still don't have any idea what to do. Share buybacks are the current new idea. That and sub-prime car loans. Read up on the latter, you'll get a chuckle.

The essay was nearly complete, in noggin, when I got Irwinned (a term, alas, that I'll be using a lot I suspect). OK, go read it. Just remember, you've read it all, save the quotes of course, here before. I don't run in the same circles, alas.

I do find him naive`, on a few points, though:
If companies increased their spending enough to close that gap, it would mean an extra $220 billion in annual economic activity and perhaps a couple of million more jobs. But there may be even more important and lasting consequences for this lack of spending by businesses.

Capital spending improves worker productivity. And worker productivity improves living standards.

The implication: as productivity rises, so do wages. In the Adam Smith notion of free market economics, yes. But the data are overwhelmingly contradictory over the last 4 decades. Little to none of productivity increases have ended up in wages. Employment doesn't, overall, go up. Just as the Foxconn guy intends to replace cheap Chinese hands with robots, so too do American capitalists. When 5 workers can build a robot which replaces 100 workers in use, you see that the arithmetic favors the MotW. And so it has happened. The Great Migrations, from farm to factory, in the 19th and 20th centuries, were based on unskilled labor moving to another unskilled labor. Given the leverage involved, with the current notion of income distribution, increasing STEM graduates will only create a larger pool of disenchanted and unemployed smart people. The MotW have shown little inclination to buy more STEM folks than they need, and they don't need all that many, especially when they can get them nearly for free from India. Guess who those Eastern European cybercriminals are? Yup.
"As manufacturing approaches full capacity, wage pressures will build and many businesses will be forced to redirect their profits away from stock buybacks and toward actual physical investment, which will help bolster U.S. GDP," [Scott Anderson, chief economist of Bank of the West] writes in a research note.

I gotta get me a pair of them rose colored glasses.

The real impetus for physical investment is better productivity, making more widgets cheaper; we're long past the time when ditches are really dug by hand. The MotW will only make such physical investments if they have, or foresee, unmet demand. But with the 99%'s share of income at best stagnant, there is no macro- unmet demand. Remember all that talk about 450mm silicon wafers? Hasn't come to be, has it? Productivity gains, from the perspective of MotW, depend on a) unmet demand with current capital and b) freezing out labor from the gains. The problem is that the MotW have gotten b) in spades but, as Your Good Mother said, "what happens if everybody behaves like you?" As the MotW squeeze out labor, and we don't get a new theory of income distribution, demand sags.

As demand for consumer widgets sags, so too does demand for more efficient capital. I think that's called a death spiral. Whee!!!!

They're Still Freedom Fries

The missives in these endeavors generally aren't reactionary to other posts out in the blogosphere; that behavior is mostly directed to news organs (I'm talking to you, Irwin). But today brings a reference post to a post, on the subject of income inequality. In this case, Tour de France winnings. What's amusing, and futile to my way of thinking, is the effort expended in attempting to find a quant model to fit these data. Ain't gonna happen.

Winnings, in any sport outside of table stakes poker (and you can name others, I suppose), are determined by the policy of the sponsors. First gets 100 units, second gets 10 units, and so on down the line. At one time, though I'd have to look it up, memory tells me that professional tennis players forced the purse distribution to be *more* unequal (here's a report on the current method). That's it. That the "winner take all" (yup, the girls) approach (and, yes, I'm talking to you, LeBron) in sport has infected normal enterprise goes a long way to explaining the global economic malaise. Why is an extra two points per game actually worth an extra $4 million/year; or whatever? That's the distribution function; set by the sponsors based on whim. In what rational world are boys and girls who play games awarded vast amounts of moolah, over folks who actually produce? What value do they provide? None, of course. Well, a bit of reflected shine on the owners.

The money comes from sponsors, TV, gate receipts (none of those, that I know of, for the Tour), and whatever. That money comes from the revenue of the sponsors, which is to say Ma and Pa Kettle who buy the sponsors' widgets. The widget buyers don't get to decide how much of the price of a widget should go to LeBron or Lance. (Aside: in today's NYT Business section is a piece on corporate boards, which asserts that boards have decided that boards shouldn't truck with shareholders, much less the public. I think that's a working definition of oligarchy.)

Policy trumps data, when they disagree. Good luck finding a general quant model of income distribution in sports. Just look up the formula for the sport/event you're curious about. Then go have a micro-brew.

17 July 2014

And Now For the News

Today has to be the sort of day that makes your average Wall Street quant want to jump out of a 30 story window. Separatists/Russians/Satan shoot down a 777, and Mr. Market goes all bananas. Not just the nowhere near ripe greenish kind of nanner, but the fully brown squishy ones that melt in your hand. Not in your mouth.

To make matters worser, at 12:27pm, Briefing.com had this tidbit:
No sooner did we post a Floor Talk comment at 11:14 a.m. ET noting the otherwise resilient nature of the equity market in the face of worrisome-sounding headlines about new sanctions being levied against Russia that selling interest picked up in noticeable fashion.

Here's a bit from 11:14 am:
The overnight buzz revolved around the new sanctions the US Treasury Department imposed on entities in Russia's financial, energy, and defense sectors. That was a step up from prior sanctions that targeted individuals and ratcheted up the diplomatic standoff between the US and Russia over Ukraine.

Data is driven by policy. And the errant grey swan. Capitalists are Patriots only when they figure they can make more money that way. "Sanctions on Vlad!!! But, but, but... he's our good buddy with all that petro!! We can do business with him, just like Dubya said he could! Look deep into his soul; he's our kind of guy. Leave our good buddy alone!!" Daddy Warbucks, indeed.

16 July 2014

The Answer is Yellen in the Wind [updated]

By now y'all have heard about the Yellen remark that small cap biotech is over-valued:
Valuation metrics in some sectors do appear substantially stretched -- particularly those for smaller firms in the social media and biotechnology industries, despite a notable downturn in equity prices for such firms early in the year.

As it happens, that wasn't in the speech, but in this paper with the speech.

What she said in the speech got little notice (the Times guy did, though):
"Accordingly, we are closely monitoring developments in the leveraged loan market and are working to enhance the effectiveness of our supervisory guidance."

In other words, we're keeping an eye on the banksters. Doesn't get much play, alas.

The thing is, small (esp. micro-) cap biotech is heavily a retail playland, where plungers are hoping that their ship will come in; Foobaz Oncology really does have The Cure and the smarter-than-the-pros plungers will get their 10-bagger. Fact is, Idera did that (go multi-bagger, not cure cancer, to be clear), and more over the last year and a bit. Went from $.19 to about $6.50; fallen back to Earth since. So, yes there are multi-baggers to be found in *-tech. GT Advanced Tech when Saturn lift vehicle when it was announced that Apple had paid it to make sapphire. Turns out, Apple didn't quite to that, and GT has too fallen back to Earth.

But, the report is quite right: because *-tech is in the business of doing new things, the FOTM respawns quickly, and the plungers go off and chase a new one with some frequency. The result is ludicrous valuations, using standard metrics. And, of course, standard metrics don't mean much in most small cap *-tech since they're not yet producing anything. One is betting, more explicitly than with other sectors, on a compound or widget that might never exist or, if it does come to exist, never work. How's your Apple Newton doin'.

[update]
Seems one can't win for losin'. One Mark Schoenbaum takes issue with "Yellen" and is quoted thus, from CNBC:
"When I look at P/E ratios in biotech, I'm really asking the question, 'Are we in a bubble?' 'cause I think what the Fed perhaps - and I'm not a Fed observer - but I'm told what the Fed is looking for perhaps is signs of bubble."

I don't watch CNBC, so I'll take the report as accurate. As explained before, Yellen didn't say what Schoenbaum says she did. She didn't say it at all. The statement was in a staff report. Moreover, as Schoenbaum conveniently ignores, the report didn't talk about biotech globally, but small-cap biotech. For an analyst, he's some combination of sloppy, stupid, and lying.

To see that the Fed might just be right, go to Yahoo! finance (or whatever) and run a two year compare chart with PBE and the S&P 500. As the signs say in the NYC subways, "Mind the gap!!"

15 July 2014

Apple of My I

Here's what's interesting about the Apple and IBM deal: if they're really building new apps for the devices, and if the server platform is built by IBM, will IBM go for server side (aka, RM) control, or keep doing the COBOL/VSAM gig via java/etc.? This may well be the last, best chance for the RM/RDBMS centric model. Or will the Kiddie Koders finally win it all? Time will tell.

14 July 2014

Fool Me Once

Buried in a Motley Fool piece on IBM's death (OK, it doesn't say death, just implies it) is this tidbit:
To buttress my point, Microsoft inked a deal with Violin Memory in April that will introduce Violin Memory's Windows Flash Array to servers in a bid to speed up popular Microsoft applications. Violin's Flash Array is a 64TB storage array, modified to run Windows Storage Server 2012 R2, and includes RDMA, or Remote Direct Memory Access, and SMB, or Server Message Block, direct on the flash array.

Beta tests carried out on the flash array have revealed performance gains for Hyper-V, SQL Server, and Microsoft VDI. According to Enterprise Strategy Group analyst Brian Garrett, what the two companies have accomplished could have far-reaching implications for the future of flash storage, since it might do away with the need for app servers.

Note the last eight words of the quote. If we can rid ourselves of app servers, we can rid ourselves of ORMs. And if we can rid ourselves of ORMs, we can implement client-agnostic Organic Normal Form™ databases. Yum.

10 July 2014

What A Buzz Kill

Buzz Aldrin got reported in Daily Tech from Reddit. I don't do Reddit (a major character flaw, I know), so I'll content myself with the DT quotes. That's all I need, anyway.
OUR resources should be directed to outward, beyond-the-moon, to establishing habitation and laboratories on the surface of Mars that can be built, assembled, from the close-by moons of Mars. With very little time delay - a second or less. Much better than controlling things on the Moon from the Earth. So when NASA funding comes up for review, please call your lawmakers to support it.

Buzz, according to the Wiki is a West Point grad in mechanical engineering. May be that's why he doesn't get it.

The reason NASA, and everybody else, has lost interest in space "exploration" after getting to the moon is simple physics and chemistry: the moon is as far as one can move humans in tin cans using chemical rockets. Fueling humans to Mars simply isn't possible with known chemistry. We know how much throw weight we can move past the moon, and it ain't much more than a Beetle. And the Area 51 aliens haven't yet divulged the secrets of Warp Drive. Even if we could get to Mars, there's no there, there. Once Again, Gertrude nails it. We'd have to ferry all life support from... somewhere? The Area 51 aliens haven't shown us how to make that Replicator, either.

We know where "earth like" planets are, and they ain't accessible with chemical rockets. And what, exactly, was the benefit of going to the moon? We just beat the Russians. BFD.

Buzz, how about we clean up our act here? Recognize that we really live in one of these, and make the best of it. We all live "Under the Dome".

07 July 2014

My Relationships Have All Dried Up

I've been reading up Hadley's new book in progress, "Advanced R", and the first section on functional programming, and right at the beginning of that section, has this gem:
To prevent bugs and to make more flexible code, adopt the "do not repeat yourself", or DRY, principle. Popularised by the "pragmatic programmers", Dave Thomas and Andy Hunt, this principle states: "every piece of knowledge must have a single, unambiguous, authoritative representation within a system". FP tools are valuable because they provide tools to reduce duplication.

"Oh, Ariel, what irony!!" OK, I made up a quote from "The Tempest", but those two words seem to strike a chord when I read that snippet. DRY amounts to saying, "for each of the entity types, code and data, within each type build orthogonally". Yet, by building DRI into a SQL database (as close to the RM as we can currently get), one gets as much DRY as it is possible. Eliminates a tonne of code, too. Oops!! I suppose that's why coders will genuflect to DRY (even if they're non-practicing parishoners), but bring out the knives at any mention of DRI data. Dreadful partisanship.

04 July 2014

A Tasty Appletini

Once again, Apple steals from the best (well, they hope so anyway). And, once again (remember the Burberry Girl?), they've got it wrong.

Apple's level of self-absorbtion measures about a gallon, while self-awareness is, may be, a teaspoon. None of the devices that Apple makes, or is assumed to make (iWatch, iWhatever), is susceptible to being owned multiply. Nor will either this guy or the Burberry Girl be able to convince those with more money than sense to take that leap. For one thing, it would mean making distinctly different versions (not just color, 5C) for distinctly different milieus. Normal watches, such as TAG Heuer makes, fit that bill. Some are dressy, some are utilitarian, some are sporty, and so forth. Those will lots of excess cash can be persuaded to buy a different one for each different occasion. Apple has shown no temperment for such sort of product differentiation; to the contrary, they limit models in order to drive down BOM cost and drive up margin.

All the objective evidence says that wealth continues to concentrate, which makes Apple's TAM for the iWhatever no bigger than what it is for the iPhone. This is not a good thing for Apple, which, along with its zealots, has puffed up with pride at selling only to the upper 20% (or whatever%). The iWhatever is not likely to move the needle much, as I've asserted before. Unless Apple has exclusive access to some blingy hardware. Do they? Only The Shadow knows.

For now, as is said of that cocktail (or, in this case, the appletini),
One martini is all right. Two are too many, and three are not enough.
-- James Thurber

Apple's customers have to get past one for Apple to score big on the iWhatever. Place your bets.

Channeling Yoda

Today Krugman, finally, says it (likely, written before the Brazil World Cup road collapse, for just a touch of international irony). The Jedi tinged title, of course.
Since 2008, however, our economy has been awash in unemployed workers (especially construction workers) and capital with no place to go (which is why government borrowing costs are at historic lows). Putting those idle resources to work building useful stuff should have been a no-brainer.

From the point of these endeavors, and being Vulcan Mind Melded, the phrase "capital with no place to go (which is why government borrowing costs are at historic lows)" is what matters. There is a reason that all that moolah, whether Chinese, German, or USofA, decided to assault the American housing market: all those multi-million dollar CEOs just couldn't (and can't) find productive use of fiduciary capital. They borrow, at near zero cost, to buy back stock, to buy competitors, and dole out much of it to themselves. In other words, they re-cycle all that moolah amongst themselves. Is it any wonder that we're not much better off, as a country? The United States of Mississippi.

All those quants, and especially the amateur wannabes, just can't get it through their thick heads that, if you want to understand the future, looking at the past is of limited value. For natural processes, where God makes the rules (and, thus, the rules don't change) the past very much informs the future. In human endeavors, especially finance and banking, where Men make the rules (and, thus the rules change, most often opaquely to those not in on the gag) looking to the past to inform the future leads to, well... The Great Recession. "House prices have been rising exponentially, so they must always rise exponentially! Give us more sub-prime ARMs!!" OK, not a direct quote that I can cite, but that does explicate the behavior.

In finance and banking, one might say, "follow the money". And that's a reasonable approximation. A better one is to follow the incentives, since the money, as water, finds its own level. Money flows according to the plane of incentive. Right now, the incentive is purely fiduciary. If that continues long enough, what Stein said of Oakland will be true of the entire United States of Mississippi, "There is no there, there."

In other words, the problem of capital allocation and increasing productivity aren't a post Great Recession manifestation. They *caused* The Great Recession. The holders of fiduciary capital, in the prelude to The Great Recession, wouldn't or couldn't find productive use (in the sense of physical production) for the money. So, it all flowed into fiduciary instruments. The lowest risk, historically, was USofA housing. Spanish beach condos as well, but that's another episode. A Ponzi scheme, self-generated, in the sense that there was no single Mr. Ponzi, short of the guy running Countrywide (Angie Mozilo). Or, as one participant observed, a game of musical chairs; hoping that one's company always finds a seat when the music stops.

01 July 2014

Kiss My Avatar

"Avatar" was the first 3D movie I recall seeing. I know I'd seen at least one of the 'spears thrown at your face' farces from kidhood, but no memory. "Avatar", on the other hand, was revelatory, if only because it displayed in polarized 3D as opposed to red/green silliness.

For some time, there's been talk of 3D NAND in SSDs. Well, here 'tis. Courtesy of Samsung, naturally. Now, they've got a repurpose for all that big node fab. Neat.
By stacking transistors (i.e. cells when speaking about NAND) vertically, Samsung is able to relax the process node back to a much more convenient 40nm. When there are 32 cells on top of each other, it is obvious that there is no need for a 10nm-class node because the stacking increases the density, allowing production costs to scale lower. As we have seen with the history of NAND die shrinks, a higher process node provides more endurance and higher performance, which is what the 850 Pro and V-NAND is all about.

Holy Smokes!!!!
... I was told that the warranty is not automatically denied if 150TB is reached under a client workload. In fact, Samsung said that they have a 128GB 850 Pro in their internal testing with over eight petabytes (that is 8,000TB) of writes and the drive still keeps going, so I tip my hat to the person who is able to wear out an 850 Pro in a client environment during my lifetime.

I wonder what this does to "Enterprise SSD" outfits like Violin Memory? Nothing good, I expect.