Transcript IPM 45: Studying Variability through Sports Phenomena

IPM 45: Studying Variability
through Sports Phenomena
Discussion of papers by
•Steve Clarke - Swinburne, Australia
•Tim Swartz - Simon Fraser, Canada
•Phil Everson - Swarthmore, U.S.A.
•Yamaguchi, Sakaori, Watanabe - Rikko, Chuo,
Toyo, Japan
Discussant: Larry Weldon, Simon Fraser, Canada1
Clarke: Studying Variability in Statistics
via
Performance Measures in Sport
• Variability - why is it important?
– Reproducibility of Apparent Effects
• Estimation, Hypothesis Testing (not Description)
– Inherently of Interest to Context
• Golf: 12th hole at Augusta Masters - par 3
key hole because of variability
• Triathlon: Equal Weighted Components? SD
• Sports Interest Motivates
– Interest in Variability
– Interest in Descriptive Stats
2
Clarke: Simulation Example
• The Karrie Webb golf example:
estimating chance of record being broken,
unprecedented event …
• Data + Simulation is an underutilized method
for data analysis - we should teach it in early
stat courses.
3
Clarke: Studying Variability in Statistics
via
Performance Measures in Sport
Overall …
Sports Examples for Interest in
Variability
Importance of Descriptive Statistics
Power of Simulation for Data Analaysis
4
Swartz: A graduate course in
Statistics in Sport
• Advanced Statistical Methods
Overall …
– But context is fairly straightforward (in sport)
• Unfamiliar
some)
Contexts
Sports
Contexts (to
make
casesSports
interesting
to many students
– Like typical consulting experience
Case
StudyStudy
CourseApproach
is particularly suited to graduate courses
• Case
– “process” course
Complexity can be designed to suit student level
• Discussion/Communication Emphasis
– Seminar discussion and student presentations
• Active Student Involvement
– Best practice for useful learning
5
Everson: Teaching Regression
Using American Football Scores
• Motivation of Real Data (768 games)
• Effects of Randomness
Overall … in Sport
– “lucky winner”
Real
Data Set
• Regression for
Prediction
Demonstrating
Utility of Regression
– Not
just “curve fitting”
Illustrating Subjective Probability
• Subjective Probabilities Justified
In Widely-Followed Sports Context
– Bookie’s probability guesses accurate
• Nice Use of Graphics (Spread N(0,13))
6
Yamaguchi, Sakaori, Watanabe:
"A Trial of Statistical Education using Sports Data in Japan"
• Social Science students math-phobic
Overall …
• Use interest in baseball to motivate
•Use
Show
Audience
pitchInterest
types incan
Sport
betocounted,
Motivate and
Statsbe
Ed.
studied numerically (fast ball, slider,..)
• Distributions and Mixtures of Distr’ns
• Causality Lesson: Home Run rate and Strike
out rate correlated, obviously not causal
7
Session Summary
Sports provide examples of
1. Focus on variability & simulation
2. Case studies for graduate education
3. Gamblers need regression
4. Distributions exist without math
8
Summary Summary
Use Sports Examples to Motivate!
Thank you
9