|
|
Musings Report 2017:44 11-4-17 Fun with Numbers: False Precision
You are receiving this email because you are one of the 500+ subscribers/major contributors to www.oftwominds.com.
For those who are new to the Musings reports: they are basically a glimpse into my notebook, the unfiltered swamp where I organize future themes, sort through the dozens of stories and links submitted by readers, refine my own research and start connecting dots which appear later in the blog or in my books. As always, I hope the Musings spark new appraisals and insights. Thank you for supporting the site and for inviting me into your circle of correspondents.
Fun with Numbers: False Precision
Sometimes numbers (data) surprise us because they don't align with the public perception that's been constructed with the numbers.
We're trained to demand data to support any position or claim, and so those who wish to shape our conclusions about what's fact, or at least supported by data (i.e. public relations, advertisers, political parties, etc.), focus on massaging the data to magnify or diminish the desired effect on the audience.
But numbers can surprise us because they're new to us, and much bigger or smaller than we expected.
For example, I ran across this article from the MIT Technology Review and was surprised that the often-vaunted IT-tech sector in India has fewer than 4 million workers. Given the population of India is 1.3 billion, and that 12 million new workers enter India's workforce annually, this number is underwhelming.
Not having seen this statistic before, I would have guessed the Indian IT workforce employed at least 20 million. I was off by a factor of five.
India Warily Eyes AI: Technology outsourcing has been India’s only reliable job creator in the past 30 years. Now artificial intelligence threatens to wipe out those gains.
Another recent example can be found in the controversy over the influence Russia-originated advertisements may have had on Facebook users. Facebook claimed that 100+ million users might have seen the offending posts--a big number to be sure--but as a percentage of the vast sea of FB posts, these offending posts turned out to be a tiny number, less than 1/10th of 1%.
100 million sounds like a big number, but it's suspect because we have no idea how many of FB's users were actually influenced by the offending posts. Attempting to measure this is fraught with difficulties, as any survey would be self-selected and self-measured.
Less than 1% sounds like a small number, but it doesn't really tell us much about the end result, i.e. how much influence did the offending posts actually wield.
In other words, these numbers offer a false sense of precision, when the reality is no one has any idea how many FB users were actually influenced and to what degree they were influenced. Combined with the modest $100,000 spent on the offending posts out of the billions of dollars in Facebook advert revenue, the data offers us very little insight into the questions, "how much influence did these posts have on the American electorate, and did they influence the outcome of the 2016 presidential election?"
Part of my job is to critically review the flood of statistics the federal government issues on the economy and society. I've found that much of the precision in such numbers as GDP and the unemployment rate is false, meaning that the inputs are intrinsically imprecise and later adjustments make a mockery of the initial number.
Sometimes this imprecision is noted--"within a range of plus or minus 5"--but the mainstream media presents such headline numbers as fact.
Some of the imprecision is engineered: why estimate "owners equivalent rent" when we can measure actual rents paid? Why guesstimate the number of small businesses arising and closing when we can use tax filing data from small businesses courtesy of the IRS, which already tabulates all income by type and source?
It's often difficult to place all the data we're given in the proper context, that is, a context that helps us correctly interpret the relative size and importance of the data point.
Numbers that aren't contextualized are easy to sensationalize, exaggerate or discount. Is 100,000 a big number? It depends on the context. It might be very large or it might be modest, depending on the context.
For my part, I try to provide some context whenever I post charts, graphs or data. Ultimately, it's not just the validity of the data that we must consider, but the adequacy of the context. In so many cases, the difference between factual reporting and quasi-propaganda is not the data but the the context that the data is presented in.
For example, I like to provide this context when discussing cryptocurrencies:
All of the World’s Money and Markets in One Visualization.
Summary of the Blog This Past Week
Let's Clear Up One Confusion About Bitcoin 11/3/17
What's Driving Social Discord: Russian Social Media Meddling or Soaring Wealth/Power Inequality? 11/2/17
What the Kennedy Assassination Records Reveal: Uncontrollable Incompetence 11/1/17
Why Is Bitcoin a Big Deal? 10/31/17
Japan Just Killed the "Bitcoin Will Be Banned" Meme 10/30/17
Best Thing That Happened To Me This Week
I discovered a volunteer papaya tree is a "hermy," a hermaphrodite that will pollinates itself. There is no way for an amateur to know what type of papaya you have--male, female or hermaphrodite--until it flowers, which takes about 10-12 months from germination. Luck was with me on this volunteer.
Market Musings: The Bubble in Bitcoin
With bitcoin up almost 10-fold this year (recently topping $7,000), and more than double its mid-September low around $3,000, I reckon it's time to review the "bitcoin is a bubble" case.
I agree with the author of this essay, who says nobody knows much of anything about cryptocurrencies:
A Letter to Jamie Dimon: And anyone else still struggling to understand cryptocurrencies
If we don't understand the value proposition, how can be value the asset (or whatever it is)?
Many pundits claim bitcoin (BTC) is nothing but a ponzi scheme, a scam to avoid capital controls and taxes, etc.
Others claim it isn't an asset class like stocks or bonds, it's a currency, and so everyone trying to compare it to stock bubbles is barking up the wrong tree.
If we're not sure what BTC even is, why is demand pushing the price up 10-fold in less than a year?
Maybe the demand is a madness-of-crowds speculative frenzy. That's what most pundits assume.
I've wondered (in my blog posts) if it isn't the super-wealthy seeking a hedge against the erosion of the current high valuations of various assets and the entire fiat currency system.
It seems to me that if BTC is nothing but a fad or speculative frenzy, it should pop rather definitively, just as the dot-com NASDAQ declined by 80% and remained beaten down for years.
If BTC doesn't crash and stay crushed, then it's not following the pattern of a speculative frenzy.
I have stated that I don't think we'll discover the value of bitcoin until we experience another Global Financial Crisis similar to 2008-09. If it acts as a hedge, the price will rise. If it fails as a hedge, the price will fall along with all the other over-valued assets.
It may seem like it should be easy to predict what will happen to BTC in a global financial crisis, but the blockchain is a new technology with no "killer app" in everyday use. Right now it is a tiny slice of the global economy, basically signal noise: all cryptocurrencies are worth $200 billion in a global asset market of $500 trillion.(see the above link All of the World’s Money and Markets in One Visualization)
We'll just have to see how it plays out.
From Left Field
Why Personal Tech Is Depressing -- we rarely saw articles like this 5 years ago...
The Coming Software Apocalypse: A small group of programmers wants to change how we code—before catastrophe strikes. -- you might want to have a printed copy of the emergency repair manual of critical technology on hand...
An Unusually Candid VC Explains Why VCs Are a Bad Idea -- VC = venture capital...
Digital capitalism produces few winners: Apple, Amazon, Facebook and Google might post huge profits, but many of their staff see little financial benefit. are we surprised?
THE WEB BEGAN DYING IN 2014, HERE'S HOW-- worth reading.
Amazon unpacked: The online giant is creating thousands of UK jobs, so why are some employees less than happy?
What Killed the Democratic Party?
Democrats once represented the working class. Not any more -- Democrat stalwart Robert Reich flees the sinking ship...
Gold Bugs Embrace Bitcoin, Upending Retail Sellers--not sure if bitcoin is to blame...
What is your biggest regret? Here are people's devastatingly honest answers-- When I posed this question on Twitter, the stories poured out and patterns emerged. Real regrets are about bad choices in love, learning and loss, being held back by fear – and self-blame
I Left Vancouver Because Vancouver Left Me -- sounds like San Francisco... all but the top 5% being ground down by high costs, overcrowding, long commutes....
F-Stop Models How to Transform Lawn Into Urban Farm in Austin (via Chad D.)
The world's biggest grave robbery: Asia’s disappearing WWII shipwrecks -- weird how pre-1945 steel and lead have become valuable....
"As technology advances, it reverses the characteristics of every situation again and again. The age of automation is going to be the age of 'do it yourself.'" Marshall McLuhan
Thanks for reading--
charles |
|
|
|
|
|
