As the graph says, probabilities of the Brewers winning X amount of games. The first 99.99% mark...

As the graph says, probabilities of the Brewers winning X amount of games. The first 99.99% mark is 52 games, then it goes to 60 and then increments by 5. This is using a binomial distribution based on my WAR projections, which give a win% of .538

Larger version here: http://img.imgcake.com/brewerprobabilities.jpg

5 months ago jhmoore Comment 40 comments 0 recs |

Story-email Email Printer Print

Comments

This is the type of stuff I love seeing.

Is there any way you could post a walkthrough on the process, so we can get some more of these over the next few weeks?

by R.J. Anderson on Jan 10, 2009 10:07 AM EST reply actions 0 recs

I like the results

by JI on Jan 10, 2009 12:30 PM EST up reply actions 0 recs

Well, once you figure out your win%

then you go into excel or open office or whatever you like and then you use the BINOMDIST function. The way I did it was I put wins in one column, and then in the other column I did *100*BINOMDIST(wins/2, 81, win%, 1)*. The program got mad at me when tried it with 162, so I just halved everything.

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by jhmoore on Jan 10, 2009 1:01 PM EST up reply actions 0 recs

stupid sbnation

anyway, the way the binomial distribution function works is like this

BINOMDIST[a, b, c, d] where a is the number of successes you want (won games), b is the number of trials (total games) , c is the probability of success (win%), and then for d you use 1 because you want to get the probability of at least the amount of games, whereas if you use 0 you’ll just get the probability of winning EXACTLY that amount of games, which is no help.

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by jhmoore on Jan 10, 2009 1:04 PM EST up reply actions 1 recs

Awesome, thanks.

I’ve yet to discover all the wonders of Excel.

by R.J. Anderson on Jan 10, 2009 1:12 PM EST up reply actions 0 recs

Thanks for the above

But how to you turn it into that sweet graph?

The Crawfishboxes
A good friend of mine used to say, "This is a very simple game. You throw the ball, you catch the ball, you hit the ball. Sometimes you win, sometimes you lose, sometimes it rains." Think about that for a while.

by DyingQuail on Jan 11, 2009 5:57 AM EST up reply actions 0 recs

even better

this is way too easy to read, could you make it smaller?

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 10, 2009 12:25 PM EST reply actions 0 recs

http://img.imgcake.com/brewerprobabilities.jpg

Scored three times and detonated an indisputable in four visits to the batting box.

by Jordan M on Jan 10, 2009 12:55 PM EST up reply actions 0 recs

heh sorry.

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 10, 2009 1:01 PM EST up reply actions 0 recs

Fantastic. Wish I'd thought of doing this.

I’ll add a chart using this method to the WAR spreadsheet next time I publish an update.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 10, 2009 6:25 PM EST reply actions 0 recs

hmmm

I wonder if that will import into google spreadsheets

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 5:52 PM EST up reply actions 0 recs

I meant a table, not a pretty graph. Something like:

Wins Prob

76 99%
81 85%
86 62%
91 34%
96 12%
101 1%

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 11, 2009 7:18 PM EST up reply actions 0 recs

finally figured the chuart out with help

from jhmore and Slyde. I should probably resign my blogging credentials…

now I just have to format it properly

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 8:21 PM EST up reply actions 0 recs

OK...

I think I got the binom funciton working… but how do I get it to graph again?

for the Royals with 79 wins, I’m using this command

=BINOMDIST

is that the right command, anyway?

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 5:57 PM EST up reply actions 0 recs

If you want to find the probability the the Royals will win x amount of games

the cell is " =100*BINOMDIST(x/2, 81, 79/162, 1) "

and you have to do it separately for each amount of games you want to look at

I use x/2 and 81 instead of x and 162 because it causes an overflow of the system when I use 162, but I don’t believe it makes a difference.

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by jhmoore on Jan 11, 2009 6:38 PM EST up reply actions 0 recs

I don't get that error, funny enough.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 11, 2009 7:19 PM EST up reply actions 0 recs

and so you just graph wins on the x and then the probabilities you get on the y

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by jhmoore on Jan 11, 2009 6:39 PM EST up reply actions 0 recs

ah... OK... I wasn't doing 100*

but I’m getting weird percentages… ahh… I see the problem, I knew what I was doing, but mixed up the trials and probability fields… yikes…

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 6:45 PM EST up reply actions 0 recs

sorry i'm such a moron

but I have from 1 to 162 on the right hand column, but the percentage just keeps goin up and up until they have a 100% chance of winning 177 games or more… I’m pretty sure that can’t be right

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 6:48 PM EST up reply actions 0 recs

make sure the only thing in the BINOMDIST function that ever changes is the x

and it’s key that there’s a 1 in the 4th spot

although you might want to check if you’re using excel because it might be different; I used openoffice for this

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by jhmoore on Jan 11, 2009 6:50 PM EST up reply actions 0 recs

excel 2003

yeah, nothing’s changing but the “x”

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 6:51 PM EST up reply actions 0 recs

ah

for “0” and “1”, Excel has “FALSE” and “TRUE”

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 6:52 PM EST up reply actions 0 recs

I think you have to use this

=100*(1-BINOMDIST(x/2, 81, 79/162, 1))

The results seem to be reversed. When you subtract it from one, the results look more realistic and the curve slope is similar.

We’’re in process of trying to a guy with a trade record of working with pitches

by Slyde on Jan 11, 2009 6:53 PM EST up reply actions 0 recs

that's possible for excel

in openoffice, the “1” makes it a sum of all the probabilities up to x/2

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by jhmoore on Jan 11, 2009 6:54 PM EST up reply actions 0 recs

the excel help on the last value

“Cumulative is a logical value that determines the form of the function. If cumulative is TRUE, then BINOMDIST returns the cumulative distribution function, which is the probability that there are at most number_s successes; if FALSE, it returns the probability mass function, which is the probability that there are number_s successes.”

I think you might be right, guys. I’m truly sorry I’m so stupid. I thikn Slyde’s 1- BINOM) might be the answer in Excel… give me a sec

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 6:56 PM EST up reply actions 0 recs

yes, I'm dumb

it should be 1-BINOM

I was just trying to go by memory because I couldn’t find my file but I found it and it’s 1-BINOM

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by jhmoore on Jan 11, 2009 6:57 PM EST up reply actions 0 recs

so the royals really are a literal lock to win at least 40?

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 6:58 PM EST up reply actions 0 recs

yeah, I'm not sure why it is

I don’t know this function well, but in excel, TRUE=1, so the results should be the same. The function appears to be a match otherwise.

We’’re in process of trying to a guy with a trade record of working with pitches

by Slyde on Jan 11, 2009 6:56 PM EST up reply actions 0 recs

but then

the first 99.99% mark is at 24 games. Does that sound right?

We’’re in process of trying to a guy with a trade record of working with pitches

by Slyde on Jan 11, 2009 6:55 PM EST up reply actions 0 recs

depends on what your win% is

maybe if it’s in the low .400s, possibly

---
Juuuust a bit outside!!
http://www.rightfieldbleachers.com

by jhmoore on Jan 11, 2009 6:57 PM EST up reply actions 0 recs

I used 79 wins

I just copied and pasted your formula.

We’’re in process of trying to a guy with a trade record of working with pitches

by Slyde on Jan 11, 2009 6:58 PM EST up reply actions 0 recs

or it's just rounding

If I drop it to 2 decimal places, it’s at 43 wins, which seems much more reasonable.

Thanks for posting this!

We’’re in process of trying to a guy with a trade record of working with pitches

by Slyde on Jan 11, 2009 6:59 PM EST up reply actions 0 recs

you the man

now I just have to remember when I was in university many moons ago and could actually generate a graph easily…

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 7:01 PM EST up reply actions 0 recs

select the data set

and hit the chart button. :)

We’’re in process of trying to a guy with a trade record of working with pitches

by Slyde on Jan 11, 2009 7:02 PM EST up reply actions 0 recs

yeah...

but for some reason I can never quickly figure out which chart type is the best

Bringing you more-or-less replacement level analysis and commentary since sometime in 2008.

by devil_fingers on Jan 11, 2009 7:03 PM EST up reply actions 0 recs

thanks to jhmoore!

I just posted a graph for the Reds at Red Reporter based on a 79 win projection for the Reds, which I still don’t quite believe to be true.

We’’re in process of trying to a guy with a trade record of working with pitches

by Slyde on Jan 11, 2009 7:36 PM EST reply actions 0 recs

Are you using Excel?

If so, how did you add the 75% and 25% red lines?

by R.J. Anderson on Jan 11, 2009 7:37 PM EST up reply actions 0 recs

yep

I added two data series, one where the 162 values were all 0.75 and one where they were all 0.25. NOTE: I did not multiply by 100, rather I converted all of the data values to percentages so they would display the same, but with ‘%’. If you are multiplying by 100, your extra series need to be whole numbers instead of decimals.

We’’re in process of trying to a guy with a trade record of working with pitches

by Slyde on Jan 11, 2009 8:09 PM EST up reply actions 0 recs