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...

From the course by University of Houston System

Math behind Moneyball

41 ratings

At Coursera, you will find the best lectures in the world. Here are some of our personalized recommendations for you

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.

From the lesson

Module 1

You will learn how to predict a team’s won loss record from the number of runs, points, or goals scored by a team and its opponents. Then we will introduce you to multiple regression and show how multiple regression is used to evaluate baseball hitters. Excel data tables, VLOOKUP, MATCH, and INDEX functions will be discussed.

- Professor Wayne WinstonVisiting Professor

Bauer College of Business

Okay, in this video I will try to explain to you the baseball

maxim that 10 runs equals 1 win approximately.

Nice round number.

And, it turns out in basketball 1 point per game is close to 3.

Now the NBA and NFL, I don't know it's 30 to 40 points with the one.

And so if you can figure out how many points a player is worth you can figure

out how many wins he generates and then you can figure out what you should think.

Well last year let's say 2014, in the American League,

I think there were 677 runs scored per team and

National League, 640.

And if I average that, what do I get?

Let's say 659 runs.

And let's suppose an exponent of 2 in the Pythagorean Theorem.

Okay, so let's suppose you're an average team.

So you'd score 659 runs.

And if you started out as an average team runs against, 659.

So the question would be how many more runs do you have to score or come out

slightly different when you look at how many runs unless you have to give up.

How many more runs to score to win one more game?

Okay, so what we can do is figure out the ratio.

Which is runs for divided by runs against, which is one in this case.

And we can get a predicted winning percentage.

Now I'm going to use something called get Ctrl+1 does alignment,

Format Cells > Wrap Text.

I'd like to do that sometimes to make my headings not as wide.

I can see more columns.

All right, so predicted 1% age would be the ratio

to the 2nd, divided by 1 plus ratio.

Caret over the 6 raises things to the power.

The caret is the healthiest key on the keyboard.

Okay, now Predicted wins.

It's like total.

And again I can do that Ctrl+1 > Alignment > Wrap Text.

It would be 162 times this, which is 81.

Okay, so now, suppose that's worth 20 more runs.

Well, it would predict I'd score eight, I would win 83 games, so that's too high.

So the question is, change the yellow so the red is 82.

And this happens a lot,

not just in sports analytics or money ball math, just in business in general.

You want to change an input cell in the spreadsheet.

So a formula, it's a given value.

And this is a job for goal seek.

Okay.

So you go data what if analysis.

And you could manually change the yellow till this is 82.

And if you've ever in a presentation manually changed a number to make a cell

get a given value, you should've used Goal Seek.

It's much more impressive than just twiddling around in a cell, and

changing the number manually.

So I go Goal Seek, there are three parts.

So the Set cell is the formula, the value is 82, and the changing cell is this.

And it should be about 10 runs.

Okay so I got about 8.2 there to be honest, okay?

Now in the old days when teams scored more runs, I think the 10 runs would be needed.

But for now sort of in the post-steroid era where pitchers are dominating, and

hitters can't hit as well because they don't have their steroids.

So, basically 8.2 runs would equal one more win.

Okay, now if I would change this to a,

so I started with 659.

In the old days, teams would score 700 runs, so

let's put 710 runs when teams were really hitting, okay.

And if I would do it there again, I'd go Data > What if Analysis > Goal Seek.

I would set this cell to 82.

Sorry, I've got to change this to 710.

Now you'll see there [INAUDIBLE].

There will be a difference, Data > What If Analysis > Data Table.

I would like to, sorry, Goal Seek.

Data > What If Analysis > Goal Seek.

I want to set this cell to 82 by changing the runs scored.

And I get really nine runs.

We have average runs

= 710 9 runs.

You'd have to score more to get the ratio bumped up by the desired amount.

Okay, that gives you an idea where sort of this 10 run idea comes from,

although as the pitchers get more dominant we may have to change that 10 runs.

It could be 8 runs will be it,

and you can see intuitively if it's a lower scoring game,

you don't need as many more runs to make a difference in your winning percentage.

Because every run means more if the games are lower-scoring.

Okay, that concludes this video.

Coursera provides universal access to the world’s best education,
partnering with top universities and organizations to offer courses online.