Devlin's Angle

January 2003

Dear President Bush

In early December, I received the following email message.



This just in from Washington...

From: "George W. Bush" [address restricted]

To: "John Brockman" [address restricted]

Subject: Science Advisor

Date: Sat, 7 December 2002

Dear John,

I appreciate your taking the time to recommend the appointment of Keith Devlin to be my next science advisor and I am pleased to hear of his interest in the position.

I am impressed with the resume of Dr. Devlin which you sent earlier. Could you please ask him to prepare a memo which answers the following question:

"What are the pressing scientific issues for the nation and the world, and what is your advice on how I can begin to deal with them?"

In addition to obvious issues that have dominated the headlines during my first two years in office, I would hope to hear about less obvious scientific issues as well.

I need the memo by the end of December.

Thank you for your help.



I smelled a rat long before I got to the end. Sure enough, the message continued:

I wish the above was really an email from President Bush. It is not. It's the set-up for this year's Edge Annual Question - 2003, and because this event receives wide attention from the scientific community and the global press, the responses it evokes just might have the same effect as a memo to the President....that is, if you stick to science and to those scientific areas where you have expertise.

I am asking members of the Edge community to take this project seriously as a public service, to work together to create a document that can be widely disseminated to begin a public discussion about the important scientific issues before us.

In case you haven't come across it, The Edge is an on-line scientists discussion group organized by writer and literary agent John Brockman. Membership of the Edge Community is largely restricted to practicing scientists who have published successful science and science-related books for a general audience. The intention is to provide an ongoing debate of the current leading edge of science in a form that a lay audience can follow. By making writing ability a criterion for contributors, Brockman has clearly excluded from this particular forum a great many leading scientists. But the gain is that The Edge is consistently one of the best places to go to find accessible, stimulating discussions about the latest results and trends in science and technology.

Starting three years ago, The Edge has kicked off each new year with an open-to-all-members discussion about a specific topic -- "the Edge Question". The spoof email I and all other Edge members received gave us this year's assignment.

So what would you have written? If you are, like me, a mathematician, you almost certainly have a good science general education under your belt, and you probably read Scientific American and one or two other science magazines fairly regularly, but that hardly qualifies us as Presidential Science Advisors. Since I always take The Edge Question seriously, I wondered for a while if I should perhaps simply opt out this year. But then it struck me. Now more than at any time in history, in addition to a Science Advisor, the President needs a "Quantitative Data Advisor". Here is what I wrote:

Mr President,

I am pleased to learn that I am being considered as your next Science Advisor. Unfortunately, as a mathematician, I do not feel sufficiently well qualified for that position. I do, however, feel there is a clear and demonstrated need for someone on your team to offer advice on interpreting quantitative data, particularly when it comes to risk assessment. I would like to suggest that you create such a position, and I would be pleased to be considered for it.

For well understood evolutionary reasons, we humans are notoriously poor at assessing risks in a modern society. A single dramatic incident or one frightening picture in a newspaper can create a totally unrealistic impression. Let me give you one example I know to be dear to your heart. The tragic criminal acts of September 11, 2001, have left none of us unchanged. We are, I am sure, all agreed that we should do all we can to prevent a repetition.

Strengthening cockpit doors so that no one can force an entrance, as you have done, will surely prevent any more planes being flown into buildings. (El Al has had such doors for many years, and no unauthorized person has ever gained access to the cockpit.)

Thus, the remaining risk is of a plane being blown up either by suicide terrorists on board, by a bomb smuggled into luggage, or by sabotage prior to take-off. In any such case, the likelihood of significant loss of life to people on the ground is extremely low. So low that we can ignore it. The pilot of a plane that has been damaged while in the air will almost certainly be able to direct the plane away from any urban areas, and the odds that any wreckage from a plane that explodes catastrophically in mid-air are overwhelmingly that it will not land on a populated region.

I know that what I say might sound cavalier or foolhardy or uncaring. The hard facts the numbers present often fly in the face of our emotional responses and our fears. But the fact is, we have limited resources, and we need to decide where best to deploy them. This is why you need someone to help you assess risk.

That leaves the threat to the plane and the people on board. Let me try to put that risk into some perspective. For a single individual faced with a choice of driving a car or flying, how do the dangers of the two kinds of transport compare in the post September 11 world? We know the answer, thanks to a calculation carried out recently at the University of Michigan Transportation Research Institute. In order for commercial air travel to be as risky (in terms of loss of life) as driving a car on a major road, there would have to be a September 11 style incident roughly once every month, throughout the year.

Let me stress that this figure is not based on comparing apples and oranges, as some previous airline safety studies have done. By being based on the lengths of journeys, those previous studies made airline travel appear safer than it really is. The figure I have given you is based on the computed risk to a single individual. It compares the risks we face, for the journey we are about to take, when any one of us decides whether to board a plane or step into our car. In other words, "How likely am I to die on this trip?"

The answer, as the figures show -- and let me stress that the calculation takes full account of the September 11, 2001 attack -- is that there would have to be such an attack once every month before air travel offers the same kind of risk as car travel.

In short, Mr. President, most of the current effort being put into increasing airline safety is a waste of valuable resources. In a world where fanatical individuals are willing to give their own lives to achieve their goals, we can never be 100% safe. What we should do, is direct our resources in the most efficient manner possible.

In that connection, if you have not already done so, I recommend you see the movie The Sum of All Fears, where terrorists smuggle an atomic bomb into the United States in a shipping crate and detonate it in downtown Baltimore. Leaving aside the details of the plot, the risk portrayed in that film is real, and one where we would be advised (and I would so advise you) to put the resources we are currently squandering on airline security.

That is why you need expert assistance when it comes to interpreting the masses of numerical data that surround us, and putting those numbers into simple forms that ordinary human beings, including Presidents, can appreciate.

Numerically yours,

Keith Devlin

I began my "letter" with the mindset Brockman asked for, namely suspending disbelief and pretending I really was writing a letter to President Bush. By the time I had finished, I wished I really was. Not because I see myself hanging around the West Wing -- although I'm sure it would be fun. ("No, Mr. Jennings, I have no current plans in that direction; however, if my country ... ") No, what happened was that as I wrote, I realized just how much the President really does need such advice.

Sure, he is not short of advice, coming from many sources. But that is just the problem. There's too much of the stuff. Even summaries are at best of limited use. Carefully crafted summary reports, with columns of figures, spreadsheets, and graphs appended at the end, take time to digest and appreciate. Time the President, or the CEO of any company, simply does not have. Instead, he has to rely on a small group of trusted colleagues to interpret that data and put it in a form he can appreciate and make use of. And that means it has to be presented in simple everyday terms that can be understood and appreciated on a human level, without any special effort.

Mathematics and statistics, like the natural sciences, draw their strength from being abstract. They are powerful tools, and no President or CEO should ignore them. But if I had the fate of even one person in my hands, let alone a large part of the world, I would want more than percentages. I would want those numbers put into human terms. One question I would want to know the answer to before making any decision is "If I were in that position -- if I were that airline traveler, policeman, army sergeant, school-teacher, or whatever -- what would it mean to me?" In other words, I'd want to see the individual human face of the issue.

That means that among those who advise the President, and whom he trusts, there should be someone able to take quantitative data, perhaps masses of the stuff, and present it to him in simple human terms -- terms that mean something to a human being. Many of his decisions so far make me suspect he does not currently receive such assistance.

To see all the contributions to this year's Edge Question, direct your browser to take you to The Edge.

The full report on airline safety that I refer to is Flying and Driving After the September 11 Attacks, by Michael Sivak and Michael J. Flannagan, American Scientist, Vol 91, No. 1, January-February 2003, pp.6-8.

Devlin's Angle is updated at the beginning of each month.

Mathematician Keith Devlin ( is the Executive Director of the Center for the Study of Language and Information at Stanford University and "The Math Guy" on NPR's Weekend Edition. His book The Millennium Problems: The Seven Greatest Unsolved Mathematical Puzzles of Our Time was published last fall by Basic Books.