Customer Reviews

Mathletics: How Gamblers, Managers, and Sports Enthusiasts Use Mathematics in Baseball, Basketball, and Football

5 star:		(5)
4 star:		(3)
3 star:		(1)
2 star:		(1)
1 star:		(0)

 

 

Wayne Winston addresses a myriad of topics in baseball, basketball and football via a statistics-heavy approach. There are 50 different "bites" spread out over 350 pages. There are many familiar topics for quantitative sports fans - Pythagorean theorem, platoon effects, player evaluations in different sports, and power rankings to name but a few.

The entire book is moderately math heavy - over half of it is devoted to quantitative solutions using algebra, statistics and Excel worksheets (which you can find online via included addresses). If you do not enjoy the mathematical side of sports, you'll find most of the book unreadable. If you do enjoy math, stats or using quantitative approaches to gambling, this book is a nice review of most of the interesting approaches out there. The bibliography of cited books reads like a "who's who" of credible quantitative sports texts.

A vast majority of the "bites" are already discussed extensively in other sources. The advantage of this book for most readers is that you can get such a diverse taste of different topics under one cover. If you are a sports modeler, the wide array of topics and approaches could help stir your own creativity. On more than one topic, I found myself saying "this assumption isn't valid!" But my making these assumptions and challenging them yourself, his approach opens up many unintended doors for the reader. For example, one bite addresses and argues that teams should pass more and run less than they do. To support this hypothesis, the book looks at a payoff chart for the yardage gained from a pass attempt versus a run attempt. The payoff chart does not consider volatility (rushing for 3 yards EVERY play is better than passing for 20 yards 1/4th of the time). It also doesn't look at the "disaster side" of passing - interceptions, quarterback fumble distribution, or greater offensive issues in a 2d and 10. Each article does make you think, which is its own payoff.

My only criticism is that the writing style seemed a little clunky. If you are not fluent in math, the combination of writing style and the amount of non-math quantity may turn you off.

The book talks about various aspects of using statistics and probability theory in professional sports. It is divided to four parts: baseball (MLB), American football (NFL) and basketball (NBA), and the fourth section talks about some sport gambling and general comments that are not a good fit to any of the other sections. The author of the book is a professor for operations and decision technologies and was also a statistics consultant for several professional teams such as the NBA's Dallas Mavericks (season 2006-2007).

Generally, the topics discussed in the book are interesting (to me both as a sports fan and with an interest and background in Math') and include topics like how to evaluate players, is there a correlation between teams wealth and the probability to win and how to compare players from different years.

However, the book itself is not an interesting read mainly because each topic is discussed in a very shallow level. The basic flow of each topic is to introduce the motivation of what statistical insights we are now checking, give the required math formula (usually without enough explanations or examples to understand it thoroughly), and than a single conclusion of the analysis is presented before continuing to the next topic. This results in the reader being left without any interesting findings or insights to learn about the topic in question with respect to different years, teams or players. For each given topic I could easily come up with several other questions that every die-hard NBA fan would like to see treated.

Basically, the book looks like a cooking book, that present an idea, gives you the formula (often discusses Excel implementation) and leaves all the hard work to the reader. Now, as the author consulted the Dallas Mavericks, most of these conclusions (in the NBA section) refer to that team in that relevant year (2006-2007). I would expect to see additional interesting results on each topic and without that I think the book will disappoint most readers.

Let me exemplify what I mean: a topic named "Are college basketball games fixed?" sounds like a very deep topic which should have profound consequences. However. this topic is exactly two pages long (57 lines to be exact), that obviously results in a very shallow treatment without any important conclusion. Another example, "Analyzing Team and Individual Matchups" topic (again, in the NBA section) is two and a half page long of text (and another 2 pages table) which only deals with the Spurs-Mavericks 2006 western conference semifinal. I guess most readers would want deeper analysis of more interesting encounters and see some data manipulation on other series.

As I wanted to get more information about these topics a quick look on various Internet website (which the book does give as good references points) I could find various articles that were more interesting than the book itself.

Overall, the book gave me a feeling that it is mostly a quick and naive compilation of a series of articles that were already posted somewhere else. Each topic (or an article in a blog\newspaper in a previous life) is treated without any deep thought or any desire to show deeper observations. If you are just looking for formulas, the internet is full of resources. This book presents very little interesting finding to sports fan.

Consists of 50 short sections, each giving a statistical analysis of a specific question in baseball, football, basketball and gambling thereon -- typical examples being Evaluating (baseball) fielders, Why is the NFL's overtime system fatally flawed, End-game basketball strategy, Rating sports teams. So it's useful for providing an overview of the type of questions people have studied statistically, and interesting to see the author's answers to the specific questions. But what lies between the questions and the answers strikes me as much less satisfactory. Typically the author just writes down a formula intended to predict future probabilities or ratings based on past data, explains how to do the calculations in Excel spreadsheets, and shows the results. This is fine as far as it goes, but (to me, as someone who teaches freshman statistics (FS)) it is not usefully connected to FS. Interpreting what the results of a linear regression or a test of significance actually mean, and when they are applicable, involves subtleties far beyond what any brief text explanation can provide. So a reader who doesn't already know FS will surely be unable to internalize what's going on, or to be able to start doing analyses for themselves. And a reader who has taken a good FS course such as Statistics, 4th Edition will have lots of unanswered questions about why the author does this procedure rather than that procedure and how reliable the conclusions might be.

I was a little bit disappointed in the level of mathematical sophistication in this book. Maybe I was expecting too much, but the math was really low level. I guess this is a plus for many readers, since basically anyone with a high school education and an understanding of Excel would be able to understand most of the book.

An example of the low level of math is that often times after the author introduces a mathematical model for a sport, instead of computing probabilities using a formula he resorts to Monte-Carlo simulation. I guess this makes it easier for non-mathematicians to follow, but as a probabilist it's a bit frustrating for me.

He does have some cool examples, but some of the chapters get a bit repetitive. At least half of the chapters seem to consist of
1) Gather data
2) Do a regression using Excel
This gets a bit repetitive. I think the beginning of the book is the strongest. This is the section on baseball. He does a nice job here at building up the techniques from simple analysis to more and more complicated analysis of individual players. I'm not an expert, but it seems to follow what likely was the historical development of sabermetrics.

Overall, if you like math and sports it's a fun read, just don't expect too much serious math.

This book has a lot of systems that do have value. It makes sense but you have to know a little about statistics such as taken a statistics course at a community college to understand some of the language. As a bettor's book it does provide information, but like all information IF you can use it to make your system better than the question is really up to the sports bettor himself to answer if the book is good or not. I took a few things and used them and it helped me.

I had no idea as to the quality and depth of thought that has gone into this field. Winston does an excellent job, making the material reasonably understandable, interesting, and detailed - beginning at first page. Even non-mathematical issues are addressed, such as why the regular baseball season is a relatively poor predictor of playoffs - because a team's regular season depends on the performance of five starting pitchers, vs. only 3-4 during the playoffs. Bill James' thinking on baseball is there as well, and how it has been modified to handle football and basketball. Evaluating pitchers, fielders, parks, individual players in major sports, comparing players from different eras - wow!

This is a great book for anyone looking to use mathematical techniques to analyze any sport (although, the main focus is on football, basketball, and baseball.

Should a basketball team down by two with seconds to go try to tie the game with a two-pointer or attempt to win it with a three? Are football teams too conservative when deciding whether to go for it on fourth down? How efficacious are sacrifice bunts in baseball when there is a man on first and no one out?

In "Mathletics", Wayne Winston uses mathematics to examine these and many, many more situations in baseball, football, and basketball. The author comes to many conclusions that are against current sports conventional wisdom. Some of the other more interesting questions Winston addresses are whether Joe DiMaggio's 56-game hitting streak is the all-time greatest sports record, whether college basketball games are fixed, whether the NFL or NBA has greater parity, which sports collapses are the greatest, and whether another system to determine a college football champion would be better than the current BCS.

Some of the math in the book is too advanced even for someone who took college algebra, but someone who did well in college algebra and is familiar with some concepts in statistics such as standard deviation will be able to understand much of the math the author uses to reach his conclusions. Any serious sports fan, however, would enjoy reading the book just to see the conclusions the author arrives at concerning three of America's most popular sports.

The articles here present a good survey of a number of techniques for analyzing player and team performance. I only read the baseball section, because that's the only sport I'm interested in, and I found the articles to be excellent, thought-provoking summaries of a number of key areas: the Pythagorean theorem, runs created metrics, VORP, platoons, historical analyses and comparisons, and so on.

The fielding analysis seemed to be just a bit spare, but that may be all that the state of the public art permits. I feel it may have underestimated the impact of poor fielding, for example by not accounting for plays (or double plays) not even attempted when the first-baseman is sub-par; or not accounting for the impact on a pitcher of a key late-inning error.

But there are very good references, and the book itself collects a lot of really useful and interesting data together in fun, well-written essays.

This is a great book! If you like math and sports this is the book for you.