« Recent Positive Developments | Main | Shaving the Violence Part Deux »

Doomsday Clock Speculist Challenge

[We're moving this entry back to the top to give those who haven't had a chance yet to tell us where you think the minute had should be on the Doomsday Clock. Come on, we know you've been thinking about it... For those who have been following the development of this post, please be aware that two Updates have been added to Kathy's original.]

The Doomsday Clock, based at the University of Chicago, has been ticking off the metaphorical minutes until apocalyptic midnight since the beginning of the cold war between the U.S. and former Soviet Union--circa 1947. In those days, the threat of the U.S.S.R. launching nuclear weapons kept school children hunkered under their desks, practicing bomb drills as naively as they did tornado and fire drills.

The demise of the Soviet Union didn't stop the Clock, however. Its keepers, the Bulletin of Atomic Scientists (also Chicao-based) keep it calibrated to the changing face of the treats to global survival. Since 2002, for example, the clock has been set at seven minutes to midnight.

Some purists might argue that the Bulletin is straying too far from its traditional message on nuclear issues. On Jan 17, 2007, the Doomsday Clock was set to five minutes to midnight, and the Bulletin issued this statement:

"We stand at the brink of a second nuclear age. Not since the first atomic bombs were dropped on Hiroshima and Nagasaki has the world faced such perilous choices. North Korea’s recent test of a nuclear weapon, Iran’s nuclear ambitions, a renewed U.S. emphasis on the military utility of nuclear weapons, the failure to adequately secure nuclear materials, and the continued presence of some 26,000 nuclear weapons in the United States and Russia are symptomatic of a larger failure to solve the problems posed by the most destructive technology on Earth.

As in past deliberations, we have examined other human-made threats to civilization. We have concluded that the dangers posed by climate change are nearly as dire as those posed by nuclear weapons. The effects may be less dramatic in the short term than the destruction that could be wrought by nuclear explosions, but over the next three to four decades climate change could cause drastic harm to the habitats upon which human societies depend for survival.

This deteriorating state of global affairs leads the Board of Directors of the Bulletin of the Atomic Scientists--in consultation with a Board of Sponsors that includes 18 Nobel laureates--to move the minute hand of the “Doomsday Clock” from seven to five minutes to midnight. "

Could there be mitigating factors the Bulletin scientists didn't include in their calibrations? In the spirit of Stephen and Phil's lyrical response to the Doomsday argument, I am hereby issuing a challenge: the formation of an ad hoc Bulletin of Speculists to present an alternative setting for the minute hand of the Doomsday Clock.


Of course the Doomsday clock is not a measurement of the actual risk of the world coming to an end. It's always been a measurement of the nervousness of atomic scientists that a major invention of their field will end the world. I suspect politics has played its part. Looking over this Doomsday Clock graph I see a rough correlation between the setting of the clock and the party holding the U.S. Presidency:


click for a larger image

Here's the official announcement of this latest move:

The Bulletin of Atomic Scientists (BAS) will move the minute hand of the "Doomsday Clock" on January 17, 2007 [to 5 minutes to midnight]...the first such change to the Clock since February 2002. The major new step reflects growing concerns about a "Second Nuclear Age" marked by grave threats, including: nuclear ambitions in Iran and North Korea, unsecured nuclear materials in Russia and elsewhere, the continuing "launch-ready" status of 2,000 of the 25,000 nuclear weapons held by the U.S. and Russia, escalating terrorism, and new pressure from climate change for expanded civilian nuclear power that could increase proliferation risks.

One problem is that the BAS is trying to set the clock for two different things – actual Doomsday (which I would take to be the end of the world – at least for humans), and nuclear war.

Though it isn't politically correct to say so, a regional nuclear war would not be the end of humanity. It doesn't even mean the rise of some weird post-apocalyptic Mad Max world. Our civilization would plod on after a nuclear war between India and Pakistan or between Israel and Iran. If the U.S. got nuked by a terrorist group it would probably be a single bomb destroying a single city – probably a major city. Our economy would be devastated, but our civilization would limp away and then eventually charge back. Ditto on a nuclear attack on U.S. interests from North Korea.

The risk of regional nuclear wars (which includes the risk of a nuclear terrorist attack) has risen as the risk of a big nuclear exchange (the kind that would endanger the human race as a whole) has diminished.

We had some close calls during the Cold War. Everybody knows about the Cuban missile crisis (although you couldn't tell it from the setting of the Doomsday Clock at the time). Few know how close we came to going out in 1983.

But we made it through somehow.

So, Kathy, I think mitigating factors that those BAS guys aren't considering include the following:

  1. Regional nuclear war doesn't = doomsday.

  2. Mutually assured destruction persists as a deterrent for a bigger nuclear war.

  3. The global economy has continued to grow as a factor that decreases tensions between the big nuclear players. China may not like us politically, but they do like having a market for their products.

  4. It's been assumed that mutually assured destruction wouldn't deter regional nuclear war. I'm not so sure. The nuts that run Iran and North Korea might actually value their lives. Hard to say.

  5. Another politically incorrect opinion – The War on Terror is working. The end of this war isn't in sight, but the terrorists have been too busy with the full-time job of surviving us to attack us directly.

doomsday-clock.jpgAs for where I set the clock, I think that we should first agree on where the lowest and highest risk should be set. There's no rule against setting the Doomsday clock outside of 15:00 to midnight (in 1991 the clock was set to 17 minutes to midnight). But the Doomsday clock graphic seems to suggest that as long as there is an existential risk, the clock should be set somewhere between 11:45 (lowest risk) and midnight (Doomsday).

But we need more clocks. One "nuclear war" clock could reflect the risk of nuclear war in whatever form. That, I'd argue, is what the BAS Doomsday Clock has become. It's not a human extinction clock, it's a nuclear-war-of-all-kinds clock. That being the case, I'd say the current setting of 5 minutes to midnight sounds about right. The risk of regional nuclear war is high.

The risk of human extinction from nuclear war is much less. If I was setting a clock for that I'd put it at 11:47 - just a little higher than the minimum risk. Of course if we actually do have a regional nuclear war, this nuclear extinction clock would shoot close to midnight. There is just no telling how this country would react if New York or Washington were wiped out. Let me put it this way – the War on Terror is also a war to save the rest of the world from our response to such an attack.

A third clock could reflect all existential risks. Risks like these.

I'd set that clock at 7 minutes to midnight.


Some of the following (which originated in a 'backchannel' email exchange among the bloggers here at the Speculist), recapitulates facts and concepts already touched on by Kathy and Stephen. To the extent that our readers find any such repetition distasteful, I offer an apology in advance. However, I think that if these relevant facts are not laid out in these particular terms, it might damage the foundation of the ideas I'd like to discuss.

Begin email extract...

The greatest value to the "Doomsday Clock" (besides as a marketing campaign by the Bulletin of the Atomic Scientists, and a fairly successful one) is as a fairly visible metaphor for, and estimate of, one sort of existential risk. When the Clock was initially conceived and presented, that existential risk was fairly clearly defined: Global Thermonuclear War between the only two social constructs capable of engaging in that activity. First between the United States of America and he Union of Soviet Socialist Republics, then among blocs of nation-states either capable of independently developing nuclear weapons or granted them as clients of such independently-capable states.

With first the departure of France's independently-acquired nuclear arsenal (The "Force de Frappe" lit. 'Strike Force') from the direct control of either the North Atlantic Treaty Organization or the Warsaw Pact and then the subsequent advent of nuclear arsenals, admitted or supposed, under the control of nation-states either unaligned with or more-or-less loosely associated with the original adversaries, the definition of the risk being characterized by the Clock expanded and became much more complex. While the fundamental risk remained Global Thermonuclear War prosecuted by one or both of the only nation-states capable of accomplishing such a civilization-threatening feat (single-handedly or 'cooperatively'), the contributing risks represented by escalation and alliances opened a larger number of paths from the status quo to the unthinkable outcome and some of those paths had distinctly lower thresholds standing between origin and outcome. A number of writers in the period made believable projections of these paths their stock-in-trade.

As the definition of the chain of risk evaluated by the Clock evolved and expanded, its utility as a widely-accepted summary estimate of the probability of the fundamental existential risk became diluted.

After the reductions in overall capacity by the relevant 'players' in the 1980's, and the dissolution of the Soviet Union in the '90's, the risk evaluated by the Clock, and thus its overall relevance, was diminished considerably. Other risks, global and potentially existential (catastrophic meteorite impact, pandemic disease, climate changes, collapse of the global socioeconomic infrastructure due to insufficient forward planning "Y2K") or "merely" regional (local famine, 'brushfire' conflicts formerly closely confined by their implications for the balance between superpowers) increased in their relative importance and/or attention regardless of any change in their independent

In order to capitalize on the significant stock of intellectual and moral authority at their disposal, and not least to continue selling their publication in the market in the wake of such decreased attention, the Educational Foundation for Nuclear Science (parent organization of the Bulletin) extended the factors considered by the Doomsday Clock. The recently-publicized addition of 'global warning' as a sufficiently-significant influence as to notably move the measure in the direction of greater certainty is only the most proximate and public acknowledgement of this evolution.

HOWEVER: The fundamental concept of the Clock was, and remains, to place an easily-understood and widely agreed upon value on the probability of one or more large-scale risks to society. As the exact risk being evaluated becomes less-closely defined the ease of understanding, wide acceptability, and thus the utility of the value is diminished.

In the intervening years since the establishment of the Clock, and particularly since the turn of the millennium, the understanding of risk and means by which it can be rationally estimated, even without a great concentration of highly-accurate information in the hands of a few analysts, have been developed. One such means of understanding and evaluation, leveraging recent improvements in communications and computing technologies, is the concept of Prediction Markets. (Addressed early on in the Speculist: ) Evolved out of insurance underwriting, actuarial science, and futures markets, and attempting to capture information held by, but not necessarily shared among, a wide sample of observers and enlist economic self-interest to drive what might otherwise be a largely- or completely-altruistic exposure of limited information and expertise, such markets operate best (make the most accurate, and profitable, predictions) when the risks being evaluated are clearly and concisely defined in time, space, and scope. In formulating such a risk prediction, or 'contract', one would frame the risk to be evaluated in specific and measurable terms. The Bulletin's Clock, in attempting to remain relevant in light of changing patterns of risk, has sacrificed the specificity necessary to be a reliable estimate of those risks even if its value is established by a large, knowledgeable, subset of the population.

While there is considerable room for debate as to whether it is more or less ethical to frame such risk predictions in specific human terms versus less emotionally-charged, but less specific, purely-economic terms, the more specific the prediction, the more precisely the risks attending to the event(s) or outcome(s) predicted may be estimated.

Perhaps an example would be appropriate:

"One million or more people dwelling in Sub-Saharan Africa will die of starvation and malnutrition by (on or before) 12 'o' clock midnight, December 31st, 2007 as verified by official population and mortality statistics and estimates published by the World Health Organization not later than December 31st, 2008"

is a far more specific statement of risk than a commodities futures contract that says, in effect:

"COB 12/31/07 CBOT Sorghum >= US$10.00 / cwt."

[Translation: At market close, on the Chicago Board of Trade, on December 31st, 2007, Sorghum (also known as millet, considered livestock feed in North America and Europe, but a fairly popular grain for human consumption in parts of Asia and Africa) will be sold for $10.00 per hundredweight, roughly $4.20 per bushel. Implied in this price are a number of factors including changes in demand among all users, worldwide crop sizes, transportation costs between successful producers and desirous consumers, and, conceivably, the unavailability of appealing foodstuffs in places, like sub-Saharan Africa, where people eat millet.]

Tragedy, too, can be evaluated in futures markets. To a certain extent, it already is. When property insurance is underwritten, in Boise, Bangkok, or Baghdad, some consideration must be made of the threats to that property posed by both natural and man-made causes. However, the specific contribution to the estimated risk posed by causes that others, including ourselves, might be interested in, is not transparent in the insurance contract itself.

While it may seem ghoulish and morally questionable to market a futures contract stating a specific cause of risk to specific property or individuals over a specific timeframe, there is a great deal of value in knowing what others might think about the probability of such a threat. Whether those individuals have vested interests in the proposition under consideration than the possibility of making a certain, limited, amount of money for guessing or estimating that risk correctly in light of the passage of events or otherwise, the society at large gains a great deal of useful information regarding that particular threat and possibly regarding risks contingent upon human action in general.

Finally, as Thomas Sowell points out in his Basic Economics, insurance and futures contracts can serve as means by which risk itself can be transferred from those with fewer resources to absorb and deal with it to those who have greater resources. The farmer who must wait until harvest not only to know how large his crop is (i.e. how many units he can bring to market), but also the size and quality of all other competing farmers' crops (which set the market price per unit) bears a substantial amount of risk. By contracting with a speculator to sell his crop, however large, at a price per unit established in advance of the harvest, the farmer limits his risk to the factors most closely under his or her own control (the productivity of his or her crop) and transfers the risk that others might out-produce the farmer to the speculator. The speculator, in turn, can hedge his or her bets by investing only a part of his or her capital in any one commodity or market, perhaps reducing the possibility of 'making a killing' by paying the farmer pennies on the eventual dollar value of the farmer's crop in particularly lean years, but offsetting the overall loss should bumper crops reduce the value of the farmer's output below the price agreed upon in advance.

To make a long two-cents worth short: The Doomsday Clock has outlived its original value in light of the dilution of its prediction and the expansion of understanding of the nature of risk and the possibilities for more successfully predicting and hedging risks developed since the Clock's inception. If the Bulletin really wished to remain in the vanguard of risk-awareness, the Directors would establish the Bulletin as the definitive window on and sponsor of an openly-traded, cash-based, highly-specific Predictions Market.

...End email extract

AFTERTHOUGHTS: One objection likely to be raised to my call for formulating marketable predictions in a specific, quantifiable, and mutually verifiable manner is that, particularly as predictions become more specific geographically and socially, the market for them also becomes more easily suceptible to manipulation by less-and-less potent actors. The worldwide wheat crop probably couldn't be materially manipulated by anything less than a nation-state or other actor of similar scope and capability. On the other hand, the continued health and welfare of a single individual can easily be altered (negatively or positively) by deliberate action on the part of another single individual (or the same individual, if the incentives were right). Political assassination is a canonical example here. Deliberate manipulation of the outcome of a prediction market contract could, potentially, deliver a profit to someone who was willing to influence the outcome of the prediction. Such manipulation for gain should, however, be fairly transparent as the trading value of the relevant contracts swung substantially away from previous values just prior to the perpetrators' intervention. There is considerable evidence of just this sort of manipulation having taken place in the days leading up to the terror attacks on September 11th, 2001 as the downturn in the valuation of airline stocks resulting from the attacks was played to advantage by investors with advance knowledge of the attacks. The fact that this manipulated investment was not only detected (eventually) but that it served as a link that tied these investors back to the organization that committed the attacks (thereby ultimately causing more harm than gain to the manipulators), as well as existing laws regarding securities fraud, insurance fraud, insider trading, and, more directly, criminal and civil laws covering physical and economic harm, intentional or otherwise, committed between persons or corporate bodies, would, I believe, serve to sufficiently dis-incentivize rational attempts at manipulation.

Finally, as my direct answer to the challenge posed in the original posting, I believe the clock should be set at:

Nuclear Weapons Use, >US$2x1013 GDP decline, 93 years, = $0.01 ^ $0.005

[Translation: The odds of a nuclear exchange causing the planetary GDP to be cut in half from today's value, (a serious effect, but not quite the civilization-ending catastrophe I grew up expecting) occuring in the remaining years of the 21st century are about 1 in 100 and took a half a chance in 100 uptick recently.]

UPDATE: (January 23, 2007 10:28PM MST)

It struck me that the kinds of 'existential threats' under consideration here, and the probabilities assigned to them by whatever chosen representation, are also contributory factors of the factor L (representing the average lifetime of intelligent civilizations) in the Drake Equation (q.v.) first discussed in this blog in July of 2004 (see item #6), again in 2005, and as recently as last week. (Actually the sum of the probabilities of all existential risks, expressed as 'years of lost life-expectancy'*, would be the reciprocal of L, if I recall my algebra correctly at this hour.)

*See Chapter 8 - Understanding Risk in Bernard L. Cohen's "The Nuclear Energy Option" for a discussion of formulating risk probabilities in this fashion

BTW - Thanks, Kathy, for providing such an interesting topic for discussion!


I was but a young child when the "Doomsday Clock" began its ticking. Now, as an old man who has seen so much happen, whew! , It would do well for all to have any important matter concerning their family of families taken care of by the Date December 21, 2012.

Great idea. Personally, I think the folks who track the Doomsday Clock are way too concerned with nuclear war. As anyone who has taken a hard look at the issue of existential risks for humanity can tell you, there are many other ways we might wipe ourselves out. I think the clock should be set at about 11:30 -- recognizing that we're closer to destruction than we should be comfortable with, but also acknowledging that the end is by no means inevitable.

In the next 100 years, we could all die pretty easily from a number of things.

I'm satisfied that in the next 100 years we will solve the problems needed to avoid the significant threats after that period. We'll be in space, so there will be no most estinction level events besides a major solar flair or the sun exploding (before we leave the solar system).

So 100, with probably a 0-10% chance of something really, really bad happening.

Let's go with 10%. 1000 years out of the maybe 2 million years of human-like life.

That's the final 0.05% of the day.

Of 24*60*60=86.4K seconds in the day, 0.05% is 43.2 seconds.

We are less than 1 minute to midnight.

But, if you consider human life as the last 5000 years of partial modernity, 1000 is 20%, meaning we're at 8pm.

hmmm - I'd say the clock should be set at about 30 years till doomsday. Perhaps something like a calendar is more appropriate. I'm quite serious. Why 30 years? Murphy's Law of Mad Science...the IQ/resources required to destroy the world goes down ~1 IQ point per cycle of Moore's Law, Metcalf's Law and the other Kurzweilian Power Laws. So, if today it requires an IQ of 160 or above to destroy the world, by 2037, millions of people will be able to do it by then.

Leave the minute hand at global/existential risk, add a second hand for regional nuclear war. Regional nuclear war is at least as likely at the top of each minute of global risk as anywhen else - and at midnight, the "region" is increased slightly to include the whole planet. (Leaving the calendar pages for future technology capable of devastating the solar system or a N-parsec region of the galaxy)

Great comments, suggestions and scenarios. This is way out of my league, but it's sure fun to get Speculists riled up!

Post a comment