Learn how probability, math, and statistics can be used to help baseball, football and basketball teams improve, player and lineup selection as well as in game strategy.

Loading...

來自 University of Houston System 的課程

Math behind Moneyball

43 個評分

您將在 Coursera 發現全球最優質的課程。以下是為我們為您提供的部分個性化建議

Learn how probability, math, and statistics can be used to help baseball, football and basketball teams improve, player and lineup selection as well as in game strategy.

從本節課中

Module 4

You will learn how to evaluate baseball fielding, baseball pitchers, and evaluate in game baseball decision-making. The math behind WAR (Wins above Replacement) and Park Factors will also be discussed. Modern developments such as infield shifts and pitch framing will also be discussed.

- Professor Wayne WinstonVisiting Professor

Bauer College of Business

Okay, let's talk a little bit about how catchers can frame a pitch.

You may have read something about this on Grand or

Sports Illustrated in the last year, and it's very, very interesting.

The idea is umpires are not 100% perfect in calling balls and strikes.

So, can certain catchers sort of coax or

influence an umpire into being more likely to call a ball a strike?

That can be worth a lot of runs, as we'll see, for Jonathan Lacroix of the Brewers.

He saved 85 runs estimated by the Fielding Bible over the last five years.

That's eight wins just by framing pitches,

making them look like strikes for the umpire, when they were really balls.

So, what would be the way to approach this?

I think everybody has done this a bit differently, but

I really like the approach in the field environment.

So, what you're trying to predict, and

it's a form of regression, but it's what's called logistic regression.

You're predicting if the pitch of a ball will be called a ball or strike.

Okay.

When you're trying to predict a binary, dependent variable,

like the pitch is called a ball or called a strike, you can't use radial regression.

You use something called logistic regression,

which we'll actually do an example of,

later in the course when we do sports ratings.

You may have seen on 5388.com the ELO ratings, or chess ratings, and

those are sort of based on that principle.

We'll do a little bit on that later in the course.

Okay, so

what are some independent variables that the fielding bible people found?

Well, obviously the location of the pitch.

If it's a foot outside, it's probably not going to be called a strike.

Okay.

Then, the other thing is the count which may surprise you.

On 3 and 0, pitches are more likely to be called strikes.

I guess they're trying to help out the poor pitcher.

The third variable, in addition to who's involved in the play,

is what's called command.

Now, this is not B613 from Scandal, but

command is how close the pitcher was to hitting the catcher's target.

The other variables would be, who's the pitcher, because some pitchers

might get better luck or worse luck, it's who the umpire is.

Some umpires are more likely to call strikes than others.

Then, of course, the catcher, and then they accounted for the batter,

which is really quite incredible because there are a lot of batters, but,

of course, there's a lot of pitches.

There's a couple hundred pitches in every game, and there's over a thousand games.

So, I mean, you've got at least a thousand times 200 pitchers,

you've got at least 200,000 pitchers a year.

Okay.

So, when running a model, they could see how significant a catcher was

in reducing or increasing the number of balls that would be called strikes.

So, the two best catchers at this, according to fielding bible,

or the guy generally considering the best defensive catcher in baseball,

Cardinal's star catcher Molina, and basically

they call this strike zone run save in the fielding bible.

33 over five years, but

then the big surprise here is Jonathan Lucroy of the Brewers

saved 85 runs over five years, and the period is 2010 to 2015.

Okay.

This is really a bring it model.

It took a lot of analysis, a lot of sophisticated math,

it uses something called logistic regression, which is more sophisticated,

actually, than ordinary multiple winner regression, which we've studied so far,

and we'll use again later in the class.

85 runs in 5 seasons is 8.5 wins that were

caused by John Lucroy coaxing umpires into calling

pitches that were actually balls into strikes.

The concept of framing I think comes from Kahneman,

I'm not sure I'm spelling this right.

Daniel Kahneman.

Looks like this.

One n.

It's [INAUDIBLE], I'm pretty sure.

Amos Traversky and Kahneman won the Noble Prize for economics.

Traversky did not win because,

unfortunately, he died tragically before his 60th birthday of melanoma.

Basically there's a great book, Thinking Fast and Slow,

I urge you all to read by Kahneman,

one of the best business books I think ever written, teach you how people think.

They talk about the concept of framing, in other words, how you frame

a question to people can govern the way they respond or make decisions, and

this is why it's called, I think, catchers pitch framing, because the catcher

is sort of framing the argument if the pitch is for a strike for the umpire.

In fact, I urge you to read that book because it's a great, great book.

Okay. That's it for this video, and

we'll talk about park factors next time.