Exclusive Interview with Bill Benter: Extended Version

I became aware of the work going on at the University of Southampton Oh perhaps 10 years ago through publications by some of the faculty here and particularly meeting faculty members at conferences where a number of the faculty here actively participate in the international conferences in this particular area which are held from time to time and I made some friendships and became increasingly interested in what was going on here do you work with any particular research projects I do I do I've been supporting a number of research projects here at the University of Southampton over the years and currently we have a project going involving the modeling of complex systems modeling uncertainty in complex systems which is ongoing and being conducted by a couple or overseen by two of the professors here and I'm very very pleased with the progress of this research what are the secrets of success in handling these very unpredictable and speculative markets I believe I mean I think there is there have been a great many of successful approaches and I can hardly speak for all of them but I believe that a combination of sound and theoretical analysis combined with data-driven empirical research I think one has to be willing to roll up their sleeves and assemble the the large rich datasets that are necessary to really test and refine theoretical models on real-world data I think the marriage of those two abilities is that is the secret to some of the more successful efforts that have been made in a way the more we know the easier it is to predict these markets indeed indeed and I think testing one's theoretical ideas and adapting and refining theoretical results to real-world data data is the flavor of the month isn't it we hear so much about large datasets becoming available more and more very much very much you know advances in computer technology and the Internet have allowed the assemblage of data sets that were heretofore unthinkable back when data sets had to be manually assembled data was a very precious commodity you know with the early researchers on horse racing involved datasets of only 200 races now 200 thousand is a common size for a data set so it's a it's a whole new world in data driven research but the the effort can still be considerable to assemble these big data sets and that sometimes where I think researchers can get lazy is in not spending you know being happy to work on the theoretical side of the problem but not willing to undertake the sleeves rolled up efforts to assemble these large rich data sets and I think that's essential a lot of work but would you say it's an interesting enjoyable and as well as rewarding thing to do oh absolutely I find you know there's nothing like the pleasure of seeing theoretical ideas tested on real-world data and seeing that the world behaving the way it theory says it should this is quite a quite a thrill and and something that you know I look forward to you know hard work with you know then pen and paper on theory and then the but the the culmination of that effort and the real reward comes when one is able to test on a large data set the results of one's theoretical calculations and and find it vindicated it seems hard to believe that something that sounds so unpredictable like the results of a horse race can be predicted if you build enough data yes indeed in fact you know that's a misconception that certain events which are popularly thought to be random or hard to predict such as horse racing is in fact one of the most predictable you know phenomenon that there is in the sporting world that the results of a horse race can be predicted with more accuracy and a greater ability to discern winners from losers than in almost any other sport but it's evidently not easy everyone will be doing it no it's not easy I wouldn't say that and one has to do it accurately enough to overcome and to to sort of outperform the ability of the the general betting public in the horse-racing markets who are effectively doing the same thing through the through the betting markets and so you have to make a serious effort at it but if you do that it's it's remarkably predictable it's one of the surprising successes and and this was really unknown at the time we first undertook this research how predictable it really would be and we were pleased to find that it was more predictable than anyone thought although that said the booking shouldn't be worried them too soon no no I think the bookies are quite capable of protecting themselves and and you know continuing to profit in their business kind of research projects do you think PhD students can get involved with here I there's a very rich set of things to explore I mean in one area is public betting behavior it's it's still you know very much an unknown why it is that the public chooses to value certain types of information more than others and what errors or patterns or anomalies there may be in public betting behavior I think also the application of advanced more advanced modeling techniques I mean statistics with the current availability of advanced computing power almost unlimited computing power available as it is today at universities like Southampton that you can undertake the statistical techniques and modeling approaches that would have been prohibitively computationally intensive in former times and so there there's methods and and you know approaches that can be used that are really quite exciting and it's an area that students are interested in and enthused by indeed yes so we I think my colleagues here and I would agree that you know the real-world applications for things such as horse racing really excite the students in a way that some of the the dryer problems in statistics don't quite have that capacity and I think as it has a as an introduction and a way to get students excited about statistics and about the possibilities of advanced modeling technologies I think horse racing is a great introduction but it's it's more than that it's a it's a very interesting and robust area of research on its own and I think it's helping that work being done on things like horse racing is actually helping to advance the state of the art in modeling techniques hopefully without applicability to a far wider range of real-world problems

Channel: Southampton Business School

Views: 34,290

Keywords: Southampton management school, university of southampton, Bill Benter, Centre for risk research, Interview, Full

Id: Ptg2XFaEtLY

Length: 7min 37sec (457 seconds)

Published: Tue Mar 25 2014