International Network of Engineers and Scientists Against Proliferation

Evaluating and Reducing Nuclear Weapons Risk Over the Next Two Decades

Ike Jeanes

The next two decades are apt to produce vast economic changes and may also include a reduction of, but not an elimination of, nuclear weapons.

If so, how much will limits on the number of warheads, short of abolition, reduce risk? There are ways to evaluate this risk and to minimize the public health threat that is produced prior to abolition. This paper will illustrate the risk of death when weapon limits from 10,700 to 10 warheads are applied in nuclear weapon nations.

Reducing Risk of Death

It is helpful to review the basics. There are only two ways to reduce risk of death caused by nuclear weapons: 1) to reduce the probability of occurrence and/or 2) to reduce the severity of occurrence. All nuclear weapons risk reduction hinges on the ability to achieve desirable outcomes in these areas.1 The nuclear weapons problem poses such high risks of death that it requires that both be employed.

Nuclear deterrence presents an imbalance that is directly at odds with the principle of reducing severity. It effectively is the intent of nuclear deterrence to exchange a "large number of deaths" for a very "low probability" of their occurrence. By this technique, nuclear deterrence increases severity in the hopes of reducing probability. This technique can only be successful in reducing risk of war deaths if the reduction in probability is proportionally greater than the increase in severity. It is doubtful that nuclear deterrence meets this criterion.

A nuclear use results either from willful or accidental causes. Nuclear deterrence only purports to beneficially influence the probability of willful causes. It does not reduce accidental causes. Indeed, for the most part, it is the core source of accidental use.

Thus, in the post-Cold War era nations are now left relying on nuclear deterrence, although:

The combined influence of the two components of risk, 1) probability and 2) severity, can be seen by using mathematical expectation. Mathematical expectation serves to measure risk and results from multiplying the probability of occurrence times the severity of the event, e.g., a 1/100 annual chance of a nuclear use (once in one hundred years) times 125 million deaths would produce an average 1.25 million deaths per year.2

Expected deaths per year is the same as average deaths per year. By applying mathematical expectation for all options, one can evaluate the public health risk that the nuclear weapon system presents.

To Tame Chance

Simon Jackman has remarked appropriately that, "probability is the workhorse of statistics. . . . Probability provides a rigorous mathematical language for communicating uncertainty, for managing uncertainty, or, in Hacking's phrase for `taming chance.'" The nuclear weapons problem is highly amenable to modeling probability with the objective of "taming chance."

I am the author of the computer program Nukefix,3 which does such modeling. While the numeric conclusions in the following will, of course, vary depending upon the data entered, they reflect outcomes from assumptions that, in my considered judgment, are reasonably optimistic. The data input, however, has not yet stood the test of extensive analysis or commentary by others, and thus outcomes from it should be regarded as an expression of my assessment alone.

On the other hand, the methods used in the program fall squarely within the mainstream of the laws of probability. Thus, in my view, they are on very solid ground and are suitable for evaluating the influences of probability and severity.

Risk of Death

Any future nuclear use is apt to "turn the world on its head," and forever alter how we humans view the world. Notwithstanding, in order to effectively control the system, it is necessary to make distinctions in terms of severity, probability, and annual risk of death.

One of the chief characteristics of proliferation is that it makes a "small" (emphasis on the quotation marks) attack somewhere in the world more probable. The data that I used suggested that with nine nuclear weapon nations there would tend to be a 50% chance of a nuclear use within 22.7 years.4

In 50% of cases in Fig. 1, the attacks tended to produce 12.9 million deaths or fewer, as indicated. Only a few warheads are required to cause a vast number of deaths. For example, India has three regions in which population densities exceed 40,000 per square mile (19.2 million people). Only 16 to 33 primitive nuclear warheads (25 kilotons TNT explosive yield) could utterly destroy these regions.6 That a single attack with so few weapons can cause as many deaths as were experienced in all of WWI is the chief reason that the median attack size is large and highly resistant to reduction, unless number of warheads are reduced to very low levels.

At the extreme, nuclear powers as large as Britain can produce attacks causing 120 million deaths, and superpowers can produce attacks up to a billion deaths [even if their likelihood is minuscule].7

In Fig. 1, having the remote possibility of such massive attacks results in 18% of attacks being larger than 120 million. It also results in the average deaths caused by the aggressor being in the neighborhood of 78 million. Deaths in excess of 78 million would tend to occur in approximately 22% of the cases in Fig. 1. This suggests that nations may not be as far from grisly Cold War scenarios as they may wish to think.

If one is skeptical about the graph, one has the option of exploring yet more optimistic conditions, which would produce a more favorable outcomes. Indeed, it is advisable to do so, because, by doing so, one can see what levels of reliability nations must achieve in order to sustain "nuclear peace" for the long term. Such an exercise also provides an opportunity to engineer a system that presents a low annual risk of death.

Fig.1: Probability that a nuclear attack would be of the indicated size or smaller.5

Annual Risk of Death, a Major Public Health Epidemic

The Nukefix program suggests that deaths caused by the nuclear aggressor alone, over the long term, would average approximately 2.4 million deaths per year under conditions indicated by the relatively optimistic data at startup. Such a death rate, if continuous, would be sufficient to cause as many deaths within 3.1 months as the United States military suffered in all wars during the 20th century - World War I, World War II, Korea, Vietnam, and the Gulf War combined.

And, such an annual death rate would also be sufficient to cause far more deaths than the United States would experience in a typical year for any of the following public health risks: heart disease, cancer, deaths attributed to cigarette smoking, AIDS, lung diseases, all car accidents, or all deaths from guns in the United States (including all homicides, suicides and accidents).

By contrast, for all conventional wars in the 20th century to date, 95 nations delivered approximately 1.18 million deaths per year worldwide.8 The indicated average nuclear weapons risk caused by the aggressor alone (not including retaliations, fallout, etc.) of 2.4 million deaths per year would be produced by (and in the main be experienced by) 9 nations.

Expected deaths per year caused by the aggressor can be determined using fundamental actuarial principles, i.e., mathematical expectation.9 By using probability principles, which serve to explain deaths over long periods of time, one can develop instructive epidemiological data for nuclear weapon deaths. The goal is to reduce these deaths by preferentially altering probability and severity.

While expected deaths per year, if the initial data input was correct, would correctly reflect outcome over very long terms, we can expect that there will be considerable variability over terms of 100 years or less. Fig. 2, for example, indicates that in 50% of cases it would be reasonable to expect 1.16 million or more deaths per year10 and 32.4% of the cases would tend to exceed the average 2.4 million deaths per year.

Largely because vast attacks of up to a billion deaths are possible (albeit extremely unlikely), and because there is low probability of smaller attacks occurring in comparatively rapid succession, 10% of cases in Fig. 2 show that the nuclear weapon system would tend to produce approximately 13 million or more deaths per year from the aggressor alone. In sum, this graph suggests a very dangerous system, which, in some instances, still has the capability of going wildly out of control.

Fig.2: Probability that expected deaths per year would be of the indicated size or larger.11

Table 1, the preliminaries

Table 1 illustrates the degree of risk reduction, other things being equal, that would be observed in Nukefix when going from 1996 upper limits of 10,700 warheads in the superpowers down to limits of 10 warheads per nuclear weapon nation. The principle feature is that reducing the number of nuclear warheads to one third of 10,700 (i.e., 3,500) does not reduce risk of death to one third of its former value. This occurs, in part, because, with fewer weapons, targeting is applied to areas that, on average, are more densely populated, and because arsenals of non-superpowers often are undiminished.

Table 1 indicates death per year at the median and at the average. In general the median is the more conservative figure and would tend to more closely approximate the outcome over a relatively short term. The average, on the other hand, best reflects the true state of the system, most particularly over the long term. It gives greater weight to the probability of large scale attacks.

 
Table 1. The Influence of Warhead Reductions on Risk of Death.10
#weap %redm DPYm %reda DPYa %RU R&U Russia US %Onat Onat popR

popUS

10,700

100% 1,055,529 100% 2,378,619 74% 1,749,752 1,192,614 557,139 26% 628,866 883 1168

7,500

98% 1,038,792 97% 2,304,349 73% 1,675,483 1,192,614 482,869 27% 628,866 1158 1442

5,000

93% 980,185 86% 2,052,671 70% 1,423,806 1,024,835 398,972 31% 628,866 1474 1788

3,500

88% 924,981 77% 1,828,997 66% 1,200,132 864,751 335,380 34% 628,866 1777 2148

3,000

85% 899,528 73% 1,730,292 64% 1,101,426 794,369 307,058 36% 628,866 1904 2292

2,500

83% 874,301 69% 1,635,350 62% 1,006,484 734,733 271,751 38% 628,866 2114 2437

2,000

80% 841,799 64% 1,516,861 59% 887,995 645,639 242,357 41% 628,866 2323 2716

1,500

76% 801,037 58% 1,374,738 54% 745,872 538,933 206,939 46% 628,866 2587 3094

1,000

71% 754,056 51% 1,220,416 48% 591,550 435,183 156,367 52% 628,866 3134 3513

500

65% 689,753 43% 1,025,063 39% 396,196 288,922 107,274 61% 628,866 4171 4821

250

60% 632,946 37% 871,263 33% 287,199 209,105 78,094 67% 584,064 6051 7044

125

50% 531,207 27% 645,335 29% 185,316 131,978 53,338 71% 460,019 7659 9662

50

33% 350,257 17% 412,455 28% 115,406 85,290 30,116 72% 297,049 12474 13666

25

26% 274,987 12% 283,183 23% 64,825 48,661 16,164 77% 218,358 14405 15001

10

20% 217,120 8% 196,183 20% 38,861 28,352 10,508 80% 157,323 21507 24713

Legend: #weap (upper limit on number of warheads allowed in each nation), %redm (reduced level of deaths per year at the median), DPYm(deaths per year at the median), %reda (reduced level of deaths per year at the average), DPYa(deaths per year at the average, i.e., expected deaths per year), %RU (percent of deaths per year caused by Russia and the United States), Russia (deaths per year caused by Russia), US (deaths per year caused by US), %Onat (percent of deaths per year caused by nuclear nations other than the US or Russia), Onat (deaths per year caused by nations other than the US or Russia), popR (population density per square mile targeted by Russia), popUS (population density per square mile targeted by the US).

 

The Influence of Warhead Reductions on Risk of Death

It takes only 354 to 800 Russian warheads to destroy every city and town in the United States that has a population density in excess of 1,000 people per square mile (36.3% of the U.S. population, 94 million people).12 Thus, START II treaty levels of 3,000 to 3,500 deliverable warheads by year 2007 (roughly a decade from now) offer little protection by way of reducing the severity of attack for nations like the U.S. or Russia. Only 10% of these numbers of warheads hitting cities in either nation would produce unprecedentedly appalling destruction.

In Table 1, by reducing the upper limit on the number of warheads in the superpowers from 10,700 to 3,500, the expected deaths per year caused by the aggressor were reduced from approximately 2.4 to 1.8 million deaths per year. This benefit would accrue largely to China and India, because of their large populations.

Yet the reduction of risk of death at a START II 3,500 warhead limit is moderate. Risk of death would still be fully 77% to 88% of its 1996 level.

Table 1 suggests that warheads would have to be cut to 1,000 or fewer warheads just to reduce risk of death to approximately one half of what it was at the 10,700 level in 1996.

The reason that "deaths per year" is instructive is that it is evenhanded measure of risk. It allows us to explore what preventive actions are most effective. With it, one can also explore the separate influences of superpowers and of proliferation.

For example, in Table 1 at the 3,500 warhead level, 66% of the risk of death comes from the superpowers and 34% from other nations. Thus, arms reductions at this level still leave the superpowers as the primary cause of death. Russia and the United States together would produce a risk of approximately 1.2 million deaths per year. If the two superpowers, Russia and the United States are removed from the system, deaths per year drop to approximately 629,000.

Table 1 primarily illustrates the influence of arms reductions (severity). Added benefit can be achieved by additional controls designed to reduce probability. The consequences of changes in probability concurrent with arms reductions can be seen in the Nukefix "T" command.

Very Low Levels of Warheads

The distinguishing feature of nuclear weapons is that, even at very low levels, they present the prospect of the aggressor alone producing more deaths than were produced by all sides in past World Wars. Compared to the 40 million deaths produced by all sides in WWII, only fifty Russian warheads can produce 20 to 72 million deaths when striking population densities of 12,500 per square mile.13

And, Table 1 suggests that limiting all nations to 50 nuclear warheads would present a scale of death in the neighborhood of 412,000 deaths per year caused by the aggressor.

Proponents of nuclear weapons have widely argued that nuclear deterrence offers sufficient protection that diminishing its effect by reducing warheads to "low" levels, such as 50 per nation, would increase risk. In Nukefix, even when the effectiveness of deterrence is reduced in the two superpowers to fully one quarter of the deterrence starting values,14 annual death rates only increase 28% to an average 528,000. This, however, is still lower than the death rate that would tend to occur at even the 125 warhead level.

This suggests that low levels of warheads would indeed tend to deliver far fewer expected deaths per year than systems predicated on highly effective nuclear deterrence. The small influence that deterrence has is attributable to the large death-producing roles that proliferation and accident play.

Further, a reduced effectiveness of deterrence, even if it were to occur, can easily be offset by improving accident reliabilities and increasing civility, as can be readily modeled in Nukefix.

Russian Economic and Nuclear Realities, the next 17 years

The Wall Street Journal, January 28, 1998 indicated that the Russian GDP had fallen 43% since 1989, a contraction worse that the Great Depression in the 1930s.15 To recoup from a 43% reduction will require many years of impressive economic growth just to bring Russia back to its 1989 GDP.

For Russia to recover from a 43% decline and regain the level of prosperity it experienced in 1989 would require APPROXIMATELY 17 YEARS AT A 3.36% GROWTH RATE.16

By comparison, the average growth rate in the US economy for the past century was approximately 3.3%. Thus, Russia would have to sustain a faster growth rate than one of the most capable economies in the world, just to recoup its 1989 position. And, to regain its relative economic position vis-…-vis the United States, Russia would need to achieve an economic growth rate fully 3.36% faster than the rate United States will experience over the next 17 years.

For these reasons, it is unlikely that Russia will recoup its former economic strength in the near term. Such circumstances raise the uncomfortable prospect of an economically deprived nuclear superpower for many years to come.

In The Limits of Safety: Organizations, Accidents, and Nuclear Weapons17 Scott Sagan repeatedly pointed out that systems for maintaining safety at high reliability levels are sufficiently complex, expensive, and difficult to maintain that the maxim, "Richer is safer", dominates.18

This brings into sharp focus the point that less affluent, or frankly impoverished, nuclear powers are apt to be much more likely to cause an accidental/inadvertent nuclear use than during the Cold War. Nations that face, or could face, adverse economic conditions include: Russia, North Korea, Pakistan, and India.

Accidental/inadvertent use poses a great threat, even in a competently managed system. In the post-Cold War era, the threat is exacerbated by the "Richer is Safer" phenomenon.

Accident and Miscalculation, Probability Basics Meet Post-Cold War Reality

Nuclear weapon nations are in the peculiarly unsatisfactory position of primarily relying on deterrence, when the preponderance of nuclear dangers worldwide come from increased chance of accidental use. The probability of accidental use can easily be two to seven times the chance of a willful use.19

The Nukefix program shows that the protective influence of nuclear deterrence is NOT the distinguishing feature of the nuclear weapons system. The combined influence of accident and miscalculation, typically, has far greater influence.

In Nukefix by analyzing reliability levels needed to prevent accidents and reliability levels needed in deterrence decision-making, one can approximate the percentage of cases in which accident or miscalculation would tend to occur. For example, under reasonably optimistic conditions, as shown in base-case condition at startup, the program suggests that there would be a 89.2% probability [89.2% = 66.6% + 22.6%] that a nuclear first use would result from miscalculation or accident. Probability for all causes of a nuclear first use can be subdivided as follow:

l In 66.6% of cases, a nuclear first use would tend to result from an accidental use [This occurs when an accidental attack is twice as likely as a willful attack.]

l In 22.6% of cases, a nuclear first use would tend to result from a decision-making miscalculation in a willfully initiated attack, i.e., the aggressor lost the war. [This value can be derived from the probability of Decision Error and the probability of a psychological/psychiatric Irrationality.]

l In 10.8% of cases, a nuclear first use would tend to result from a willfully initiated attack in which the attacking nation achieved its military objective, i.e., the aggressor won the war. [In this instance only a small proportion of willfully initiated wars are won by the aggressor. The rest are accidents or miscalculations.

While there are obviously all sorts of variants in the kind of conditions that one can develop in a computer program by changing the data that is entered, the central trait of the nuclear weapons system is that it is difficult, with a proliferated world of nine nuclear weapon nations, to construct plausible scenarios of long-term nuclear peace where the influence of miscalculation and accident is acceptably low, when there is a predominant reliance on nuclear deterrence.

China, the next two decades

In contrast to the Cold War where the Soviet economy was often roughly 35% the size of the US economy,20 China's economy is already roughly 57% the size of the US's.

With each passing day the United States is now facing in China a nuclear weapons nation with more economic power than it ever faced before. If the United States' economy sustains a 3.5% growth rate, and if China sustains a 8.8% growth rate, the two will be economic equals in 11.5 years, the year 2009. And, at these rates, within the two decade period that this paper addresses, China's economy might well be 55% larger than the US economy.21

Such a relationship was already being considered in military terms by the Pentagon's Office of Net Assessment in 1995, as indicated in Scientific American:

"One reason for a reassessment is that, within a few decades, the threat to the U.S. may come not from a small rogue regional power but instead from what has come to be known as a `peer competitor': in essence, a new superpower, such as China, a resurgent Russia or perhaps even India."22

Unlearning the Lessons of Nuclear Deterrence

For purposes of nuclear deterrence, in the 1960s, Secretary of Defense Robert McNamara defined unacceptable damage as being conservatively from 20 to 25 percent destruction of the populace and 50 percent of industrial capacity.23

Using McNamara's assumption of attacking roughly 25% of the populace, note that it takes 1,025 to 2,189 U.S. warheads to destroy the area in which 24.8% of the Chinese reside, and it would take 709 to 1,600 Russian warheads (which are larger) to destroy the same area.24

In sharp contrast, it takes only 174 to 394 Russian warheads to destroy the area in which 25.5% of the U.S. population reside. This suggests a lopsided condition, where it takes only a few hundred warheads to destroy a large proportion of the US population, but four times that many to have roughly the same relative impact on China.

Thus, if Russia and the US are of a mind to stick with their Cold War deterrence policies of both counterforce (predominantly "military") and countervalue (civilian/city) targeting, they are apt to be highly resistant to reducing deliverable warheads to fewer than 1,200 and 1,525 respectively, if they allow for 500 counterforce warheads. And, if they argue for a more conservative assessment with 1,000 counterforce warheads, they respectively may be inclined to resist going beneath 2,600 to 3,200 warheads.25

For these reasons, START II levels of 3,000 to 3,500 warheads were not strongly objected to, because nations were substantially making a virtue out of a necessity. The arsenals were bloated, and they could be easily reduced without breaching conventional nuclear deterrence thinking.

However, the need to reduce nuclear weapons beneath the 1,000 warhead plateau at 1.2 million deaths per year will require breaking with long-held deterrence rationales. This brings us squarely to an event that occurred almost one hundred years ago.

In the midst of peace, fifteen years before WWI, Tsar Nicholas II of Russia proposed the Hague Conference, in 1899, as Paul Walters related, for the purpose "above all, of putting an end to the progressive developments of the present armaments," the outcome being: "When the Conference met, a committee of military and naval experts was appointed, but the only result was to produce a number of reasons to prove that any agreed limitation of armaments was impossible [emphasis added]."

In 1907 the second Peace Conference at the Hague was held, again called by the Tsar: "A British resolution, to the effect that military expenditures had much increased since 1899 and that governments should seriously examine the question, was adopted unanimously, all agreeing to do nothing at all. [emphasis added]"26

Similar to the "deterrence" era today, the period was characterized as the "Armed Peace".27 WWI began unimpeded seven years later, with no vital interests at stake.28

In Conclusion

Public health is caught in an historical crosscurrent. The ascendancy of China and doubts of military intentions of adversaries wed nations to nuclear deterrence. And nuclear deterrence is also supported within mainstream academia, as summed up by the political scientist Scott Sagan: "Among both political scientists and historians who study this issue, a near-consensus seems to exist: a long and distinguished list of scholars argue that nuclear weapons have been a moderating force in international relations."29

Nuclear deterrence theory, however, is being wrongly applied because it fails to account for the influence of accident. Few scenarios, even at the outer limits of plausible optimism, suggest that accident and miscalculation will account for significantly fewer than 50% of nuclear incidents.30 And, degradation in safety and increased probability of accidental use in Russia and some minor powers can be anticipated over the next two decades because of the "Richer is Safer" phenomenon.

This suggests that in the post-Cold War era a "laser beam" focus should immediately be directed worldwide towards preventing accident and miscalculation, NOT towards enhancing nuclear deterrence.

Moreover, nuclear deterrence is being wrongly applied, because its theoreticians have failed to adequately account for the annual risk of death that it produces. Conservative actuarial methods used here suggest that, even competent nuclear weapons system can present an annual risk of death that can exceed the number of deaths produced in a single year by major public health risks in the United States.

The reduction of warheads to 2,000 in each of the superpowers, as has been contemplated for START III, is not apt to reduce this risk of death by much more than 36% as shown in Table 1. Significantly deeper reductions in risk of death will require abandoning a nuclear deterrence rationale that is predicated on countervalue targeting and on attacking 20 to 25% of an adversary's population. So, absent an ideological sea change in military and governmental thinking, arms reductions may have a proclivity to stall at high levels of deaths per year - approximately 1.4 to 1.5 million per year and 1,500 to 2,000 warheads per superpower, as in Table 1.

Vanguard attempts to control weapons approximately two decades before WWI and WWII were quashed by military rationale. Reducing nuclear warheads to levels as low as 50 will require a major shift in mainstream governmental and military thinking. The path will have to be accurately and quite persuasively presented, or it, too, is apt to be rejected. The chief difference this time around is that the risk of death is far higher.

This paper illustrates how severity and probability can be evaluated for the purpose of reducing those deaths. 31

1 A dynamic model of security and risk in nuclear arms race is given in: J. Scheffran, Strategic Defense Disarmament and Stability - Modelling Arms Race Phenomena with Security and Risk under Political and Technical Uncertainties, Doctoral Thesis, Department of Physics, University of Marburg, FR Germany, 1989.

2 Accuracy in mathematical expectation depends on a correct estimates of probability and severity. The particular examples that follow assume that, on average, there would only be one chance in 270 years that Russia would initiate a nuclear use of any size, and only one chance in 3,300 years that it would initiate an attack of maximum size. It was also assumed, on average, the US would initiate a nuclear attack of any size only once in 580 years, and it would initiate a maximum size attack only once in 2,900 years. All intermediate size attacks were also specified for each nuclear nation. To the extent that these probability assumptions would be insufficiently optimistic, the particular data presented in the following would tend to be unduly pessimistic. Contrariwise, if the estimates are excessively optimistic the presented data, too, would tend to be excessively optimistic. Similarly, faulty estimates in severity caused by errors in estimates of population density in the involved area, errors in estimates in the number of deliverable warheads, or errors in estimates of the size of warheads can produce misleading results. The population data used here was carefully derived, as was the area of death-dealing impact from blast and thermal effects. And, the Nukefix user can vary probability and severity data via the "U" (Ultimate) command. In the main, making reasonable estimates is not an insurmountable problem. By making upper-limit, optimistic estimates of probability and conservative estimates of severity, one can have increased confidence that the case is not stated in an unduly pessimistic manner. Since the following estimates do not include estimates of deaths from radioactive fallout or nuclear winter/autumn, or from retaliatory attacks, it probably errs on the conservative side for estimates of total severity. Notwithstanding, it is necessary that considerable additional study be given to accurately estimating the components of probability and severity. 3 Nukefix runs on IBM compatibles and presently can be downloaded at http://www.nukefix.org . 4 Ike Jeanes, Forecast and Solution: grappling with the nuclear (Blacksburg: Pocahontas Press, 1996), pp. 28-34, 103, 137-143, 148-154, 239-256, 261-263 shows various methods for approximating the average annual probability per nation (composite RL). Where RL is the inverse of the annual probability, the 22.7 years is a consequence of a 295 RL. Nukefix also illustrates methods for approximating RL, as in its "W" (Early Warning Command).

5 Fig. 1 appears in Nukefix when the "Q" (Quick) command is entered and varies according to the data the user enters. The data shown here reflects the base-line starting condition developed from the range of probabilities and sizes of attacks under consideration. The bumps along the curve reflect the sum of the influences of individual nations. The curve is not the result of a single formula, but reflects the consequences of applying mathematical expectation to approximately 2,600 potential outcomes.

6 Nukefix, "M" command and the "Targeting the World" page.

7 Worst case would be an attack on most densely populated regions of China. Based on the "Targeting the World" page in Nukefix. Additional discussion is provided in the Nukefix "M" command.

8 Jeanes, Forecast and Solution, pp. 134, 653.

9 As can be seen when pressing the "DPY Method" button in the Nukefix "U" command.

10 Note that coincidentally this is close to the previously referred to 1.18 million per year conventional war deaths.

11 The graph in Fig. 2 appears when the Nukefix "Q" (Quick) command is entered.

12 Nukefix, the "Targeting the World" page.

13 Nukefix, "T" command and "Targeting the World" page.

14 As can be seen on the "Deterrence" page in Nukefix.

15 WSJ, p. A1.

16 1.75 = 100/(100-43); 1.75 = 1.033617

17 Scott D. Sagan, Limits of Safety: Organizations, Accidents, and Nuclear Weapons, (Princeton: Princeton University Press, 1993), p. 18.

18 Jeanes, Forecast and Solution, p. 652.

19 The Nukefix start up screen suggests two times greater. The Nukefix "W" (Early Warning) command suggests that for nations operating under launch-before-detonation policies that it can often be seven times greater.

20 Jeanes, Forecast and Solution, p. 439.

21 Nukefix, "Y" (EconomY) command. Any 5.3% spread favoring China produces substantially the same outcome.

22 Gary Stix, "Fighting Future Wars," Scientific American, Vol 273, No. 6 (December 1995), p. 93; Jeanes, Forecast and Solution, p. 440.

23 Jeanes, Forecast and Solution, p. 300.

24 Nukefix, the "Targeting the World" page.

25 1,209 = 709 + 500; 1,502 = 1,025 + 500 and 2,609 = 1,600 + 1,000; 3,189 = 2,189 + 1,000.

26 Jeanes, Forecast and Solution, p. 561.

27 Jeanes, Forecast and Solution, p. 549.

28 WWII also began because effective inhibitors were rejected many years (14 to 19 yrs.) in advance. Jeanes, Forecast and Solution, pp. 571-598.

29 Jeanes, Forecast and Solution, p. 645. 30 Nukefix, the "Deterrence" page . 31

Ike Jeanes is an independent researcher and the author of the book Forecast and Solution: grappling with the nuclear (Blacksburg: Pocahontas Press, 1996). He also is the author of the computer program Nukefix. Some of the text here is used in that program. He developed and manages the www.nukefix.org website.