THE BOOK--Playing The Percentages In Baseball

This is also interesting when thought of in the context of Japanese baseball, where non-champion teams have to win an extra game to take the series.

In that case, if both teams are 50/50, but one team needs to win 3 games and the other needs to win 4 games, then the odds are 65.6% that the team that needs fewer wins will win the series.

If all we care about is your chance of winning all 3, can’t we just get the cumulative binomial distribution. You win as long as you win 3, 4, 5 or 6 of the first 6 games, it doesn’t matter that 6 is impossible. In excel, 1- binomdist(2,6,.67,1). The answer you get is 90.3%.

Yes, and if you change .67 to 2/3, you get 0.8999, which is what I get with my method.

But your way sure is less fun.

Excel and “less fun” = blasphemy.

This is actually just the well known negative binomial distribution:

http://en.wikipedia.org/wiki/Negative_binomial_distribution

Thank you for that… I was unaware that it had that name.

Sounds much like Adam Dunn’s 2011 season.

I while back we found that when you had a 3-0 count, a pitcher batting, and no runners in scoring position (in other words no reason not to throw a fastball right down the middle) pitchers threw a strike 71% of the time.

If pitchers throw strikes 71% of the time if they really try then you get 6% walks, 94% strikeouts.

It would be interesting to compare that outcome with some of the worst batting performances - Brandon Wood 2010, Jeff Mathis/Drew Butera 2011, bad hitting pitcher, etc.

Dean Chance was one of the worst hitting pitchers ever (.066, 2 XBH in 662 AB).  He did take a decent amount of walks for a complete non-hitter (30 in 759 PA), so that combined with his 63% strikeout rate indicates he must have thought the best he could contribute was to keep the bat on his shoulder.

Rally: I profess to not remembering that discussion at all.  If you have a handy link, can you post it?

But, yes, finding game situations where the pitcher intends to throw a ball basically as a fastball down the middle is the ideal scenario we are looking for.

http://www.insidethebook.com/ee/index.php/site/comments/how_much_control_do_pitchers_really_have/

Here’s the link to a thread from a while back.  I’m putting more text into this post so maybe the cooment system won’t think it’s spam.

Geez, a thread that I started, and was pretty specific about.  I don’t remember that at all. Maybe I post too much.

Just re-read that thread.  Great stuff in there!  If ever I were to do a “best of”, that would be in there.

Thanks for posting this. I often want to calculate this kind of thing and I’m never sure if I am doing it right. What about home/road issues? The favored team might have one expected winning percentage at home but then they have to go to the other team’s park where they will have a slightly lower chance of winning. So what happens when your p is not the same for all trials?