Taleb and the language of risk

Last night, I read about Nicholas Nassim Taleb on English Language and Usage StackExchange (EL&U). Professor Taleb wants to introduce a new word to the vocabulary of global financial collapse, antifragility:

So let us coin the appellation “antifragile” for anything that, on average, (i.e. in expectation) benefits from variability.

Consensus on EL&U was that this was a creative but unnecessary neologism. I echo the concerns of other EL&U users: Antifragility might cause confusion (maybe it is “anti-fragility”). There are many adequate, extant words that Taleb could use, however, antifragility is a term that will be uniquely associated with him.

I am not convinced that there are many entities that actually thrive due to uncertainty. A delta hedge that is long volatility is the only construct that I can think of off-hand. Perhaps that was what Taleb had in mind.

The original Black Swan

The Black Swan by Thomas Mann; 1954 UK First Edition

There was a slightly less contemporary black swan, the novella written by Nobel-prize winner Thomas Mann toward the end of his long and distinguished literary career.

The plot of that short fiction work also pertained to an anomalous event, one that could be considered a statistical outlier.

Economic Models for Turbulent Times Part 2

A new research study has already received unusual attention. The Network of Global Corporate Control [PDF] discovers a relatively small group of multinational companies with disproportionate influence over the global economy. The authors, a trio of complex systems theorists at the Swiss Federal Institute of Technology in Zürich, are supposedly the first to empirically identify such a network.

The problem is approached using mathematical models designed for capturing behavior of complex natural systems. The study applies this methodology to a large data set of corporate information, to map ownership among the world’s transnational corporations (TNCs). Previous studies reported that a few TNCs drive much of the global economy. However, they analyzed fewer companies. Due to limited data availability and computing resources, past studies did not consider the effect of indirect ownership.

Methodology

Orbis 2007, a repository of over 30 million private and public companies, published by Bureau van Dijk, was the data source. The study sample used the 43,060 largest TNCs, and derived the associated ownership linkages. The network structure was based on the relationships between shareholding interests, then weighted by each company’s operating revenue. This yielded a directional map of global economic power.

Quantitative results

The model revealed a core group with networked ownership, see image below. These 1,318 companies:

• all had ties to at least two other companies in the core group
• on average, were each connected to 20 other core group companies
• represented 20% of all global operating revenues,
• collectively owned the majority of the world’s largest blue chip and manufacturing firms:
• in total, generated 60% of all global revenues.

Risk perception and reality

This is an excerpt, selected by Moi, from the article Risk perception, a recent post that appeared on the Soapbox Science Blog, Nature Publishing Group.

Universal symbol of radiation and fear. Image via Wikipedia

Sometimes, no matter how right our perceptions feel, we get risk wrong. We worry about some things more than the evidence warrants (vaccines, nuclear radiation, genetically modified food), and less about some threats than the evidence warns (climate change, obesity, using mobile phones when we drive). That produces a Perception Gap, the gap between our fears and the facts.

The Perception Gap produces dangerous personal choices that hurt us and those around us (declining vaccination rates are fueling the resurgence of nearly eradicated diseases). It causes the harm to health of chronic stress (for those who worry more than necessary). And it produces social policies that protect us more from what we’re afraid of than from what in fact threatens us the most (we spend more to protect ourselves from terrorism than heart disease)… which in effect raises our overall risk.

We do have to fear fear itself…too much or too little. So we need to understand how our subjective risk perception works, in order to recognize and avoid its pitfalls.

Here was the take-away for me: Societal risk management has to recognize the risk of risk misperception–  recognizing the risk that arises when our fears don’t match the evidence. This is truly the risk of The Perception Gap. It has always been relevant, and becomes so once again in light of the recent E-coli outbreak in northern Europe. The Guardian UK used that as a starting point for a well-written and up-to-date article about the hazards of risk misperception and the consequences of irrational behavior.

Kahneman and Tversky did extensive research on this topic. I am not concerned whether articles like the one referenced above are derivative, in the sense of revisiting past work. Possibly it is an application in the context of current events. Or it may be entirely original new work. My concern is solely that there is an awareness of the reality, and that it be acted upon.

Radiation levels in Japan and the U.S.A.

Radiation levels in Japan post-Fukushima

The source for this chart is Ryugo Hayano, Ph.D. Professor Hayano is the Physics Department chair at The University of Tokyo. Click on the chart to view a larger version, with higher resolution. It used to link directly to the Hayano account on one, then another image sharing site, but both are out of business now. (Lack of persistent URLs is a problem everywhere.)

Radiation levels in Japan March 15 to April 10, 2011

I offer my thanks to @hayano and Daniel Garcia. Daniel R. Garcia Ph.D. is a nuclear scientist from France, doing a postdoc at TEPCO, in Fukushima. He was there prior to the earthquake and tsunami. Daniel frequently sends updates as @daniel_garcia_r. He works at the reactor site every day, takes photos, and makes them available via Twitter.

Control board of Fukushima 1 nuclear power plant when all was well

Both Daniel and Professor Hayano are reliable, because they never confuse Becquerel with Sievert with Roentgen. They know radio-isotopes and their half-lives better than nearly anyone. Daniel was needed to assist the press a few weeks ago, when there was confusion between Cesium 137 versus Iodine 137 and again between Iodine 131 versus Uranium 137.

Other locales, other radiation levels

The Radiation Network is an excellent resource for radiation information in the U.S.A. and other parts of the world. It is a network of civilian volunteers using a protocol to report radiation readings, 24 hours a day, 7 days a week. Sensor stations are located throughout the contiguous 48 states, Hawaii,  Alaska and Norway. There was one in Northern Japan. Sadly, that sensor went off-line last month.

The Radiation Network is non-profit, all volunteer and headquartered in Arizona. Tim is the public face of the Radiation Network. Using software developed for this purpose, Tim collects and aggregates the real-time data from the sensor stations, then updates the map online with the readings at one-minute intervals. The Radiation Network went online nearly a decade ago, and offers reliable baseline measurements for comparison. This facilitates detection of any incident. The criteria for elevated radiation levels include:

• Rule-out protocol for false positives, e.g. spikes due to sensors  malfunctioning;
• Level of radiation that is significant: Higher than the threshold AND sustained, and how long “sustained” is;
• Exogenous causes such as geography. Readings in Colorado are always higher due to the higher elevation,

The website is basic but functional. There are Radiation Network maps of Europe, Japan, and the US (broken out for Alaska and Hawaii), and a message. The message is a running log of updates.

Political risk exposure and short URLs

URL shortening was rarely seen anywhere other than micro-blogging platforms such as Twitter and Status Net’s identi.ca. Shortened URLs are not prudent from an information security point-of-view, as one takes a leap of faith by clicking on a link that is not descriptive. Descriptive links are also preferable for economic reasons, as they are reputed to figure positively in the mysterious world of search engine optimization (SEO) for page rank.

Yet shortened URLs are gaining acceptance. They are very convenient.

Twitter introduced its own shortening service in September. Facebook did too. Google provided URL shortening with its goo.gl product in December 2009. Google expanded the range of goo.gl for use on any domain, as it was restricted for use with Google product pages before October. However, there is a new and surprising consideration when making a case for, or against, URL shortening: Political risk exposure.

Top-level domains (TLDs) are assigned by ICANN. Generally speaking, each sovereign nation has its own TLD. For example, websites registered in Australia use the .au suffix, German sites are .de , while Japanese sites are .jp . The Libyan government is the official registrar, as designated by ICANN in 2005, for all .ly sites, which are also the domain-of-choice for leading URL shortening services bit.ly , ow.ly and vb.ly .  What will be the consequences of Libya’s domain seizure of vb.ly on October 6, reported by When All Your Short Links Belong to the Libyan Government, on these .ly URLs?

RowFeeder is a social media oriented web analytics service. It stands out from the glut of other Twitter-verse services by delivering reports directly to a spreadsheet. Founder Damon Cortesi described a new feature for RowFeeder customers: availability of URL shortener bit.ly.

RowFeeder Offers URL Shortening with bit.ly

You can now put a bit.ly link in the tracking field, and have a new column in your downloads with bit.ly click counts at the time of each post… [storing] the click data along with the Tweets and Facebook posts about a specific piece of content.

In light of the recent disruption in the .ly domain space, I enjoyed the closing lines of the announcement:

Please note: This feature has not been approved by the Libyan government, so count clicks at your own risk. Our vb.ly integration is on hold pending recent news.

*Emphasis is NOT mine.

Evolution Robotics

Bias is bad. My prior post could be misconstrued as pejorative commentary, unfairly targeted at Tweetup, an innocuous, and free-of-charge, client application for users of the Twitter micro-blogging service.  Twitter certainly is responsible for the silly avian-themed jargon that is steadily seeping into common vernacular, undermining my ability to sound impressive when pretentiously blathering away, as I’m doing right now. However, Idealab, the owners of the Tweetup social influence metric software did not create the Cult of the 140 Character communication standard.

Tweetup is actually one of many applications, services and bona fide products offered by holding entity Idealab.  This Pasadena, California based company has a refreshingly original and diverse assortment of creative product offerings. In fact, Idealab is no parvenu, having existed as a going concern since the 1990’s, all the while fostering an impressive  range of mostly feasible technical innovations.

Supermarket Loss Control by Evolution Robotics

I was quickly riveted by the Idealab site’s video demo of their TunnelHawk system. TunnelHawk’s design imperative is to mitigate retail store losses due to customer and employeee theft.  To make clear that I am not predisposed against Idealab (in the aftermath of my earlier Tweetup monetization tirade), I am featuring a product video of IdeaLab’s clever response to supermarket petty theft,  the robotically driven TunnelHawk supermarket checker.

Idealab: Evolution Robotics Retail TunnelHawk Video.