ENBIS 2007: Dortmund

Box Medal winner 2007: Poul Thyregod

Mr President, Fellow statisticians.

I feel grateful and honored by receiving this prestigious award. I am sorry that my current health condition prohibits me from being with you in this moment.

I believe that all who are present here share the interest in using our statistical tools and insight for solving real life problems. For some of us, it is more than just an interest, it is a dedication. This dedication is nurtured by the feeling of joy and satisfaction we experience when our analysis provides value-adding insight into sources of variation, and the transfer of variation in a real problem.

For me it all began in 1959 when I started studying mathematical statistics. I recall that I chose this topic because I had heard about a Danish retail company that designated the location of their factories such that the travelling time to their shops was minimized. Input was the location of their shops, the frequency and length of periods of red light in the various traffic signals, and various other traffic data. It was all fed into a computer, and out came the optimal location. Later, I learned that this problem usually is considered a problem in OR rather than statistics, and also, I wonder how elaborate a model that could be formulated and solved by computers in 1959, but nevertheless, this early example of evidence based management lured me into statistics.

In 1959 computers were just entering the scene, and very quickly the new technology was adopted by many statistical practitioners as a powerful tool. In the early days of computers their capacity was rather limited, 1 kilobyte, and they were mainly used for running standard programs for the analysis by classical linear models. In many ways that was satisfactory for the kind of data that were analysed at that point. In these early days data were moderate in number, and low in dimensionality.

Personally, I am still influenced by the good statistical practice of those days, namely plot your data, and prefer models that reflect the physical reality in a meaningful way.. However, as time passed, more powerful computers were developed, and along with that, our concept of data was extended. Now, a data point may be a spectrum, an image, a DNA-string etc. and the number of data points presented for analysis is
sometimes counted in millions. Similarly, we have experienced that the statistical community has faced the challenges and developed tools for solving problems involving such data. The problems may relate to recognizing patterns, like in character recognition, to problems of detecting deviations from patterns, like in fraud detection. I feel fortunate that I have had the privilege to follow and be part of this development.

It is characteristic for the field of statistics that theories and concepts have often been developed in fruitful interaction with the application areas. Some tools and concepts have simply been adopted from the application field, an example is the concept of kriging in spatial statistics. Another example is force of mortality and hazard rate in reliability theory and life sciences. I think that it has been important for the success of our statistical methods that they are problem driven, and that they are developed in this interaction and cross-fertilization with application areas.

Also, statistical practitioners often have a background and basic education in another area than statistics. The diversity of statistical practitioners in ENBIS is our strength. What matters is the dedication to using statistics as a tool for providing insight and adding value to enterprises. ENBIS was created at a point in time where enterprises all over Europe were open to use evidence based management, often in the form of six sigma initiatives. In the new Europe national borders were fading away, and enterprises and production were becoming multinational. Although national statistical societies with more or less emphasis on statistical practice already existed in Europe, time was ripe for creating this forum where statistical practitioners could interact and exchange ideas across national borders. The successful exchange of ideas being as well within industrial sectors as across sectors.

The programme for this conference also reflects a diversity of tools and problem areas that is characteristic for applied statistics. I wish you all a successful conference, learning from the experience of other practitioners, and widening your professional network.

Personally, I find that ENBIS has widened my statistical horizon to industries that I had not been aware of, to European countries where I was ignorant above the activities for statistical practitioners. Also, ENBIS has given me many new friends.

Thank you for this, and once again, thank you for the honorable prize.

Poul Thyregod


Problem driven

It was only a few decades after Fisher in 1921 had published the notable memoir “On the Mathematical Foundation of Theoretical Statistics” emphasizing the exact solution of sampling problems. The ideas and concepts in this memoir were rapidly adopted as basis for statistics curricula all over the world.
Personally, I later learned also to appreciate Fisher for his insightful solutions to a variety of problems from statistical practice. His solutions were characterised by a modelling of variation and transfer of variation that reflected the physical realities in a meaningful way, even if this  resulted in models that were non-linear.

In Memoriam

Unfortunately, Poul Thyregod past away after a long illness on 11th July 2008, aged 69, just one year after he received the George Box Medal. As Andrea Ahlemeyer-Stubbe put it he is our role-model ENBIS8 in Athens was dedicated to his memory. Since there was no chance to actually add a session or similar to the conference just two months before the actual event, a short memorial speech was added to the program and two posters were presented. Below you find the two posters as PDF.

About Poul Thyregod

Document title Document summary
Poul Thyregod -- his last paper
This poster displays Pouls last paper, a great example of his way of thinking and how he has (and will) always influenced the way statistic was(is) presented.
Poul Thyregod - Memorabilia
This poster shows a number of highly interesting information about Poul. This includes also the letter you find above.