News Centre

Cambridge appoints Professor of Risk

12 January 2007

A new Professorship designed to help improve people's understanding of the mathematics of risk is being established at the University of Cambridge.

The new Winton Professor of the Public Understanding of Risk will seek to help individuals, institutions and government refine their decisions in risky situations.

Risk is a factor in all human activity and different people react to risks in very different ways. Questions requiring a scientific ability to assess the chances of something happening - or not happening - arise all the time. Here are some examples:

• Following the poisoning of the Russian ex-spy Alexander Litvinenko, traces of polonium-210 were found at various locations in London that he had visited. Statistically, how probable is it that someone who visited the same locations at a later stage would contract radiation poisoning?

• A recent study of transfusion patients given blood contaminated with the human form of mad cow disease has indicated that the 24 still alive are at "substantial" risk of contracting vCJD. What are the risks of contracting vCJD via a blood transfusion? How do they compare to the risks of getting the same condition by eating meat?

• An apparently healthy woman is judged to be at risk of breast cancer and is advised to undergo mastectomy. Should she do so?

• A person has to cross a main road to reach the shops. Should (s)he walk straight across the road, or use an available footbridge instead?

• How sensible would it be for me to invest in the stock market today? Might delaying improve my prospects greatly?

• A 29-year-old man decides to marry his girlfriend of three years. What is the chance that he will meet a more suitable partner at a later stage?

As these examples show, risks need to be considered in both the most ordinary of situations, and in high-pressure environments. Risk assessment is often based on analysis of data, but there is always a danger that statistics can be abused. Expert evidence in courts has been subjected to close scrutiny in recent high-profile cases such as the prosecution in the USA of OJ Simpson, and the cases of SIDS (sudden infant death syndrome) in the UK.

"The way to confront risk is via mathematics and statistics," Professor Geoffrey Grimmett, head of the Department of Pure Mathematics and Mathematical Statistics, said. "This new Professorship will enable Cambridge to play an important role in clarifying the understanding of risk in many fields of human endeavour. It will strengthen our ability to reach out beyond Cambridge to government and the public alike."

The new Winton Professorship has been created in perpetuity in the Statistical Laboratory of Cambridge University, thanks to a £3.3 million donation from The Winton Charitable Foundation. David Harding, a Cambridge alumnus and Managing Director of Winton Capital Management, a London-based hedge fund, is a Trustee of the Foundation.

"My time at Cambridge studying Natural Sciences showed me the importance of accuracy in empirical information and its interpretation," he said. "This has been a key factor in my career in finance and is highly relevant to our awareness of the risks that affect us in our everyday lives. I am delighted to make a contribution to public debate and policy by helping create this new position at Cambridge."

Professor Alison Richard, Vice-Chancellor of the University of Cambridge, said: "The Winton Charitable Foundation has been wonderfully generous in its support of the Cambridge 800th Anniversary Campaign. Endowed positions such as the Winton Professorship are of very special value to the University and I look forward to appointing the inaugural holder of this major new post."

Notes for Editors:

Risky situations

Below are three "real-life" examples where a better understanding of the science of risk might be of use:

1.) Mastectomy, or not?: Breast cancer is a common problem for women. Its incidence among the population can depend on factors such as country of residence, age and genetic make-up, and the true risk to an individual cannot be found without analysing a complex set of data. The progress of the disease can be slowed by treatment and the choice of treatment must depend on this analysis. As yet, there is no certainty in this process and each possible action involves danger and risk. There are apparently healthy women who are judged to be at risk of breast cancer and some of these may be advised to undergo mastectomy. It is to be hoped that these women, in facing such an important decision, have been helped to take into account the great cost of suffering from cancer, the probability of the disease, and the relative certainty, yet distress, of mastectomy.

2.) Variant Creutzfeldt-Jacob disease. The incidence in humans of vCJD may be linked to exposure to a disease in cattle called BSE. There are various possible routes for such transmission, including exposure to infected animal tissue, and to blood transfused from infected donors. It has been a major issue of public health in recent years to identify the risks to humans of vCJD, with substantial consequences for the UK ranging, for example, over farming methods, medical strategy, and international relations.

A lot of progress has been made: data has been gathered systematically over a long time-scale, there have been intensive studies of such diseases in humans and animals, and the evidence has been analysed using techniques of mathematical modelling and statistics.

3.) The marriage example: In a mathematical idealisation of the marriage question called the 'secretary problem', we imagine that a large number of candidates for a post are to be interviewed in sequence. The interviewing panel is required to operate the following rule: after each interview, the panel must decide whether or not to offer the job to that person, and there is to be no recall of previous candidates. What strategy should be adopted by the panel in order to make as large as possible the chance of appointing the strongest candidate?

The answer is to interview a certain proportion of the candidates without making an offer, and then to appoint the first person (if any) who is better than all those seen so far.

Some obvious complications arise in trying to adapt this to the marriage problem, such the unknown number of suitable spouses, the long time-scale involved, and the difficulty of judging the candidates given incomplete information.

For more information, contact:

• Tom Kirk, Office of Communications, University of Cambridge, Tel: 01223 332300, Email: tdk25@admin.cam.ac.uk

  Interviews with Professor Geoffrey Grimmett can be arranged via email: G.R.Grimmett@statslab.cam.ac.uk