Thursday, January 27, 2011

Briefly Defining FIP And Presenting My Findings

I'll start off by briefly defining FIP, so you have a little bit of context behind the findings I'd like to publish here. FIP stands for Fielding Independent Pitching and in the most basic terms possible, is a replacement statistic for ERA that tries to evaluate the pitcher in a vacuum. This means when looking at this statistic we will not be taking into account the performance of the defense surrounding a pitcher, only the pitchers performance.

FIP is expressed as a number in the same way as ERA (0.00) so that we may have a good method of comparison between the two. FIP is also known as "ERA plus", and by that I mean, that's what I call it. It's a pitchers true ERA, better reflecting the run impact a pitcher had on his games that pure ERA by itself.

So, I may or may not go into more depth on FIP at a later date (I probably will) but right now I'm going to segue into my findings. Check in after the jump to see.

So, let me explain how we find FIP. The formula for finding a pitchers FIP is:

(((13*HR) + (3*(BB + HBP)) - (2*K)) / IP) + constant

In MLB this constant is generally hovering around 3.2 and is recalculated every year. The problem I ran into trying to use this statistic for the WSU baseball team is that there is no developed constant at the college level to use in this formula. Using 3.2 was giving me numbers that most likely wouldn't accurately reflect our pitchers performances. 

In MLB this constant is derived by calculating the FIP for every pitcher in the league (minus the constant component) and the ERA for each player in the league. The league average FIP is then subtracted from the league average ERA to obtain the constant to be used in the FIP formula.

Now there aren't nearly as many MLB teams as NCAA Div. 1 baseball teams, so instead of going through every pitcher for the entirety of Div. 1 I decided I would derive the FIP constant for the Pac-12. Every team except Colorado has a baseball team and each of these 11 teams has between 11-16 pitchers, so the sample size is large enough to derive a solid constant. 

To see the spreadsheet I used to derive this constant, go download it here.

Let me quickly explain this spreadsheet. Columns A-H are pretty self-explanatory, I double checked and all but two or three ERA are exactly correct. Now the first column labeled "FIP" (column I) is the FIP derived for each player without the constant. Off to the right is where the average ERA and FIP was found for the Pac-12 and then from that I found the league constant.

I used this constant to derive the FIP found in the column labeled "FIP + Constant" (column J). The numbers in this column are the true FIPs for each pitcher on a Pac-12 roster during the 2010 season.

So, any questions, comments just go ahead and post them in the comments section. Sorry, no funny in this post, just wanted to present my findings here.

1 comment:

  1. First, great job on the formula. Math sucks. Secondly, talking about sabermetrics sucks. It's complex and difficult to understand most of the time. I did it over at my blog last year, and had to do it in super basic terms so my girlfriend would understand what I was talking about. I spent a lot of time just trying to make everyone understand how useless some stats (like wins, errors, era, etc.) are. You're fighting the good fight, though. Keep it up.