Computational Methodology for Modelling the Dynamics of Stat Arb
A Computational Methodology for Modelling the Dynamics of Statistical Arbitrage
Andrew Neil Burgess
Decision Technology Centre Department of Decision Sciences
A thesis submitted to the University of London for the degree of Doctor of Philosophy
UNIVERSITY OF LONDON LONDON BUSINESS SCHOOL
1
October 1999
To my parents, Arnold and Carol.
© A. N. Burgess, 1999
2
3
Acknowledgements
Thanks to my supervisor, Paul Refenes, for bringing me to LBS, keeping me in bread and water, helping me with ideas and even reading the bloody thing! Thanks to past and present colleagues at the Neuroforecasting Unit/Decision Technology Centre. Especially Yves Bentz, Peter Bolland, Jerry Connor, Stefania Pandelidaki, Neville Towers and Peter Schreiner; for discussions, hard work (not all the time), company and six good years. Also all the ex-CRLers who showed that even a real job can still be fun. Thanks to the visitors to the LBS group, Fernando and Paul, for good times and hard work in London, Helsinki and Melbourne. Thanks for the people who helped keep it real, especially to Pratap Sondhi for ideas and support when he was at Citibank at the beginning; the other sponsors of the Neuroforecasting Club and then the Decision Technology Centre; Botha and Derick down in S.A. for trusting me to build a trading system; and David and Andrew for all their efforts on behalf of New Sciences. Also to whoever first decided to hold NIPS workshops at ski resorts; and to the regulars at NNCM/CF conferences: John Moody, Andreas Weigend, Hal White, Yaser Abu-Mostafa, Andrew Lo and Blake LeBaron in particular for their enthusiasm and inspiration. Finally, my love and thanks to Deborah, who had to put up with me whilst I was writing up and whose photographs of chromosomes had to compete for computer time with my equity curves.
4
Abstract
Recent years have seen the emergence of a multi-disciplinary research area known as “Computational Finance”. In many cases the
Andrew Neil Burgess
Decision Technology Centre Department of Decision Sciences
A thesis submitted to the University of London for the degree of Doctor of Philosophy
UNIVERSITY OF LONDON LONDON BUSINESS SCHOOL
1
October 1999
To my parents, Arnold and Carol.
© A. N. Burgess, 1999
2
3
Acknowledgements
Thanks to my supervisor, Paul Refenes, for bringing me to LBS, keeping me in bread and water, helping me with ideas and even reading the bloody thing! Thanks to past and present colleagues at the Neuroforecasting Unit/Decision Technology Centre. Especially Yves Bentz, Peter Bolland, Jerry Connor, Stefania Pandelidaki, Neville Towers and Peter Schreiner; for discussions, hard work (not all the time), company and six good years. Also all the ex-CRLers who showed that even a real job can still be fun. Thanks to the visitors to the LBS group, Fernando and Paul, for good times and hard work in London, Helsinki and Melbourne. Thanks for the people who helped keep it real, especially to Pratap Sondhi for ideas and support when he was at Citibank at the beginning; the other sponsors of the Neuroforecasting Club and then the Decision Technology Centre; Botha and Derick down in S.A. for trusting me to build a trading system; and David and Andrew for all their efforts on behalf of New Sciences. Also to whoever first decided to hold NIPS workshops at ski resorts; and to the regulars at NNCM/CF conferences: John Moody, Andreas Weigend, Hal White, Yaser Abu-Mostafa, Andrew Lo and Blake LeBaron in particular for their enthusiasm and inspiration. Finally, my love and thanks to Deborah, who had to put up with me whilst I was writing up and whose photographs of chromosomes had to compete for computer time with my equity curves.
4
Abstract
Recent years have seen the emergence of a multi-disciplinary research area known as “Computational Finance”. In many cases the
Bibliography: 355 Recognition in Forecasting Financial Markets, February 15, 1991, reprinted in (Trippi and Turban, 1993) Box, G 357 Davis, L., (Ed.), 1987, Genetic algorithms and simulated annealing, Pitman, London 361 Juels, A., and Wattenberg, M., 1996, Stochastic Hillclimbing as a Baseline Method for Evaluating Genetic Algorithms, in Touretzky, D 364 Newbold, P., 1974, The exact likelihood function for a mixed autoregressive moving average process, Biometrika, 61, 423-426 366 Schwarz, G., 1978, Estimating the Dimension of a Model, Annals of Statistics, 6, 461-464 367 Thierens, D., 1995, Analysis and design of genetic algorithms, Unpublished doctoral dissertation, Catholic University of Leuven, Leuven