Stat 890 - Special Topics: Statistics in Sport

Welcome to what I think is the only full credit graduate course in Statistics in Sport! I thank the Department of Statistics and Actuarial Science at Simon Fraser University for allowing me to pursue my special interests.

Lectures: Mondays and Wednesdays 12:30-14:20 in TASC2 8070

Office Hours: Drop by anytime.

Brief Description: This course involves data analysis, modelling and problem solving related to statistics in sport. The course will be structured into well defined units where the units do not depend greatly on one another. In this way, if something catches your interest, you can show up for a lecture or two without enrolling in the course. The course requires a certain amount of statistical maturity but will not be as technical as most of our graduate courses in statistics. It would be good to have an enthusiasm for sport.

Why teach a course on statistics in sport? The short answer is that I like sport. But there is more to it than that. I think back to all of the statistical genetics talks that I have attended, and I still know very little about genetics. The relative simplicity of sport allows us to quickly get at the heart of problems, and this enables sensible modelling. My hope is that we can use sport to learn how to think about problems, to gain familiarity with a number of technical tools and to put statistics into practice.

Why do research on statistics in sport? Again, the answer is that I like sport. However, I believe that there are opportunities to make an impact in this broad area. We will see examples where statistics have been used badly and there are opportunities for improvement. The proliferation of sports data also offers opportunities. I also think that the time is ripe for contributions from statisticians on sporting problems. Observe the popularity of ``Moneyball: The Art of Winning an Unfair Game'', the 2003 bestseller by Michael Lewis that chronicles the success of the Oakland A's and their use of quantitative methods.

My observation is that statisticians who do the best applied research tend to work closely with subject matter experts. For example, many biostatisticians have connections with medical facilities. It is through these ties that meaningful research problems arise. In statistics in sport, the beauty is that many of us are subject matter experts. I have been watching sports on television my entire life. Our intuition about sports combined with statistical training leads to opportunities.

What will be learned in this course? In addition to the broad objectives previously stated, students will be introduced to a number of technical topics that may be of value in other domains. For example, we will gain some familiarity with sequential importance sampling, rejection sampling, the Gauss-Seidel algorithm, simulated annealing, estimation using order statistics, hierarchical modelling, Markov chain Monte Carlo, regression trees, agglomerative clustering algorithms, Brownian motion, etc.

Textbook: We will not have a textbook for the course although ``Statistics in Sport (1998)'', edited by Jay Bennett contains chapters on a number of sports and is an excellent reference. We may occasionally use the Bennett volume although we will be primarily reading papers that have caught my fancy. Two other books ``Statistical Thinking in Sports'' (edited by Albert and Koning 2008) and "Anthology of Statistics in Sports" (edited by Albert, Bennett and Cochran 2005) consist of collections of papers relevant to our course.

Grading:

• Assignments - 20%
• Presentation - 30%
• Final Exam - 50% (10:00am-1:00pm Monday Dec 8/2008, K9509)

You will not be punished with excessive assignments in this course. There will be roughly four short assignments that involve computing. I want you to have time to carefully read the literature provided. One of the goals of the course is to provide you with experience and confidence in reading technical papers, and we will read a terrific number of papers. You should read the papers in advance of coming to class. I also want you to have time to prepare an excellent lecture. There is an underlying premise that you are all keen and will not abuse the freedom from midterms and excessive assignments.

Regarding the presentation, ideally, I want you to initiate a research project of your own on statistics in sport. Typically this begins with a question on sport that has been in the back of your mind. You will likely need to collect data relevant to the question. You may work in teams and you may need to provide your fellow students with reading material in advance of your presentation. I am hoping that with feedback from the class, you may be on your way to some publishable work. Clearly, this is a serious assignment, and you should start thinking about it from Week 1. I encourage you to browse through the online journal ``Journal of Quantitative Analysis in Sports''; it may give you some ideas concerning your presentation.

If you are unable to come up with a line of research, and I hope this is not the case, then you may present a paper on statistics in sport. I may be able to direct you to a suitable topic. My grading of the presentation will be more favourable towards those trying to do something novel over those presenting existing papers.

The final exam is three hours long and is closed book.

Course Outline :

• Week 1 (beginning Sept 1/08) - There are no classes the first week. Ironically, I will be at the RSS conference in Nottingham giving an invited talk on ``Using sports and the Olympics for statistics teaching, learning and research''.

• Week 2 (read papers 1, 2, 3, 4, 5, 6 below)
  • introduction to the course
  • balanced draws in highland dance

• Week 3 (read papers 7, 8 below)
  • sports gambling
    • lines, odds: American and European
    • parlays, teasers, futures, etc
    • vigorish, middling, scalping (i.e. arbitrage)
    • testing for profitable systems and the issue of multiple comparisons
    • money management: fixed wagers, fixed percentage wagers, the Kelly system
  • introduction to one-day cricket (for future lectures)

• Week 4 (read papers 9, 10, 11, 12, 13, 14 below)
  • the hot hand: does it exist?

• Week 5 (read papers 15, 16, 17, 18 below)
  • one day cricket
    • the Duckworth/Lewis method
    • quantifying the margin of victory in matches
    • performance measures for batting and bowling
    • optimal batting orders

• Week 6 (one lecture this week due to holiday on Thanksgiving Monday Oct 13, no preparatory reading)
  • modelling the outcomes of major league baseball games

• Week 7 (read papers 19, 20 below)
  • handicapping in golf

• Week 8 (read papers 21, 22, 23, 24 below)
  • questions of interest from "A Statistician Reads the Sports Pages"
    • where to aim in darts?
    • under what circumstances should you kick on fourth down in American football?
    • when do you pull the goalie in hockey?
    • how many majors will Tiger Woods win in his career?

• Week 9 (read papers 25, 26, 27, 28 below)
  • comparing performances from different eras
  • comebacks in team sports

• Week 10
  • TBA

• Week 11
  • TBA

• Week 12
  • student presentations

• Week 13
  • student presentations

Reading List:

I am going to provide you with all of the papers in a convenient binder. The papers are listed below:

1. Verducci, T. (2004). Welcome to the new age of information. Sports Illustrated, April 5, 50-62.
2. McKeon, J. (2004). This is the ultimate? Bull! Sports Illustrated, April 5, 67.
3. Blount, R. Jr. (2004). As so often happens. Sports Illustrated, April 5, 68-73.
4. Swartz, T.B. (2007). A graduate course in statistics in sport. Invited paper in the Proceedings of the 56th Session of the International Statistical Institute, Lisboa.
5. Maher, M.J. (1982). Modelling association football scores. Statistica Neerlandica, 36 (3), 109-118.
6. Swartz, T.B. (2007). Improved draws for highland dance. Journal of Quantitative Analysis in Sports, 3 (1), Article 2.
7. Insley, R., Mok, L. and Swartz, T.B. (2004). Practical results related to sports gambling. The Australian and New Zealand Journal of Statistics, 46, 219-232.
8. Beaudoin, D. (2003). The best batsmen and bowlers in one-day cricket. MSc project, Simon Fraser University.
9. Tversky, A. and Gilovich, T. (1989). The cold facts about the ``hot hand'' in basketball. Chance, New Directions for Statistics and Computers, 2 (1), 16-21.
10. Larkey, D., Smith, R.A. and Kadane, J.B. (1989). It's okay to believe in the hot hand. Chance, New Directions for Statistics and Computers, 2 (4), 22-30.
11. Tversky, A. and Gilovich, T. (1989). The ``hot hand'': Statistical reality or cognitive illusion? Chance, New Directions for Statistics and Computers, 2 (4), 31-34.
12. Hooke, R. (1989). Basketball, baseball and the null hypothesis. Chance, New Directions for Statistics and Computers, 2 (4), 35-37.
13. Wardrop, R.L. (1995). Simpson's paradox and the hot hand in basketball. The American Statistician, 49 (1), 24-28.
14. Dorsey-Palmateer and Smith, G. (2004). Bowlers' hot hands. The American Statistician, 58 (1), 38-45.
15. Duckworth, F.C. and Lewis, A.J. (1998). A fair method for resetting the target in interrupted one-day cricket matches. Journal of the Operational Research Society, 49, 220-227.
16. de Silva, B.M., Pond, G.R. and Swartz, T.B. (2001). Estimation of the magnitude of victory in one-day cricket. Australian and New Zealand Journal of Statistics, 43 (3), 259-268.
17. Beaudoin, D. and Swartz, T.B. (2003). The best batsmen and bowlers in one-day cricket. South African Statistical Journal, 37, 203-222.
18. Swartz, T.B., Gill, P.S., Beaudoin, D. and de Silva, B.M. (2004). Optimal batting orders in one-day cricket. Computers and Operations Research, to appear.
19. Bingham, D.R. and Swartz, T.B. (2000). Equitable handicapping in golf. The American Statistician, 54 (3), 170-177.
20. Swartz, T.B. (2007). Third report on a proposed handicapping system for the Royal Canadian Golf Association.
21. Stern, H.S. and Wilcox, W. (1997). Shooting darts. In the column, A Statistician Reads the Sports Pages, Chance, 10 (3), 16-19.
22. Stern, H.S. (1998). Football strategy: Go for it! In the column, A Statistician Reads the Sports Pages, Chance, 11 (3), 20-24.
23. Berry, S.M. (2000). My triple crown. In the column, A Statistician Reads the Sports Pages, Chance, 13 (3), 56-61.
24. Berry, S.M. (2001). How Ferocious is Tiger? In the column, A Statistician Reads the Sports Pages, Chance, 14 (3), 51-56.
25. Simon, G.A. and Simonoff, J.S. (2002). Were the 1996-2000 Yankees the best baseball team ever? Chance, 15 (1), 23-29.
26. Berry, S.M., Reese, C.S. and Larkey, P.D. (1999). Bridging different eras in sports (with discussion). Journal of the American Statistical Association, 94, 661-686.
27. Gill, P.S. (2000). Late game reversals in professional basketball, football and hockey. The American Statistician, 54, 94-99.
28. Stern, H.S. (1994). A brownian motion model for the progress of sports scores. Journal of the American Statistical Association, 89, 1128-1134.