**The Factor12 Rating**

The Factor12 Rating (F12) is a measurement utilizing league average performance to compare the value of all MLB pitchers for a given season.

F12 consists of the following twelve statistics, incorporating every aspect of pitching….

Innings Pitched (**IP**); Strikeouts Minus Walks (**SO-BB**); Fielding Independent Pitching (**FIP**); Earned Run Average (**ERA**); Walks plus Hits per Innings Pitched (**WHIP**); Home Runs per 9 innings (**HR/9**); Walks per 9 innings (**BB/9**); Strikeouts per 9 innings (**SO/9**); Opponents Batting Average (**OBA**); Opponents On-Base Average (**OOBA**); Opponents Slugging Average (**OSLG**); Modified Base-Out Percentage (**MBOP**) includes wild pitches and balks. For 1876-1944, Rick Swanson’s Power Pitching (**PP**) replaces OSLG due to unavailability of accurate Total Bases Allowed data.

F12 produces a rating number using the percentage difference equation for the ten ratio categories. Each pitcher is ranked according to league average performance using 2.000 as the baseline. Categories have a maximum value of 4.000 and a minimum of 0.000.

The Innings Pitched (IP) and Strikeout Minus Walks (SO-BB) categories utilize a percentage change formula, which does not contain a fixed range.

A pitcher’s F12 is the sum of the percentage difference/change value of the twelve statistical categories. Pitchers recording zero innings pitched will receive a 0.000 F12 Rating. Elite pitchers will accumulate a 30.000+ seasonal rating.

Pitchers completing less than the average yearly innings (i.e. 65.75 in 2011) will have their F12 Rating weighed by the percentage of innings completed in relation to the league average (i.e. Sergio Romo 48 IP/65.75). This adjustment enables starting pitchers and relievers to be compared.

**History of Factor12**

*(by Sven Jenkins, 4/4/2012)*

With the release of the Factor12 Rating this week, I figured I’d put together a brief history of the project. The F12 story is full of unlikely scenerios that began in Poughkeepsie, NY in 1996.

The gist of the Factor12 Rating was born in both the lounge of Dutchess Community College and the Price Chopper picket line. Ideas were brainstormed at school, and numbers were crunched at work. Originally used as a fantasy baseball scoring system, FactorTwo (as it was called then) used a variation of the percentage change formula to compare each player with the league average to determine a fantasy score.

FactorTwo led to the creation of the original Picket Line Production Fantasy Baseball League, which has since been revived thanks to Yahoo. FactorTwo included every statistic found in USAToday‘s weekly printing of MLB totals. The league used Wins, Losses, Games Finished, etc…. many stats that have since been realized nearly useless by today’s sabermetric community.

Due to the excessive man power needed to enter all the data, the PLPFBL met its maker at season’s end. After the epic collapse of the league, the FactorTwo spreadsheets sat on a 3.5 inch floppy disk, inside a plastic bin for 16 years.

Josh Robbins unearthed the disks this winter and purchased an external disk drive on eBay to reclaim the files. What he found was a curiosity at first, but after I made the comment, “I wonder where all these players ranked”, the project was reborn.

Realizing that we might have the framework for a legitimate baseball metric in our hands, two months of hard work followed that conversation. Josh and I contemplated the value of a strikeout and the philosophy of FIP (Fielding Independent Pitching). I had nightmares about division by zero and Josh memorized league innings totals back to 1950. Eventually, we discovered the *percentage difference* formula and almost became real mathematicians by the end of the process. With the Cactus League season coming to an end, the Factor12 Rating was completed just in time for Opening Day 2012.

*(Factor12 is the result of the collaborative effort of Josh Robbins and Sven Jenkins, with additional credit to each member of the 1996 Price Chopper picket line)*

Josh, thanks for posting on Replacement Level. I like F12. I think you do a great job of focusing on the metrics that matter, and I particularly enjoy seeing K-BB in there, as that might be the simplest way to gauge a pitcher’s effectiveness over a single start, a season, or a career. My one concern is that you’re combining basic elements (K/9, BB/9, etc.) with more complex metrics like FIP, which already encompass those same elements. I think F12 is going to identify the best pitchers more often than not, and that’s the goal, but I wonder if you’re going too far in trying to incorporate everything.

ERA is far from perfect, but we know what it measures: a pitcher’s ability to keep runs from scoring when his defense isn’t making obvious mistakes. WAR is complex, but we know what it measures: a player’s contributions to his team’s runs scored and allowed, as compared to a replacement level’s contributions. My goal with True Season Score is to measure a pitcher’s ability to control true outcomes, with a slight adjustment for outcomes partially within his control. Could you explain in brief what F12 tries to measure?

Keep up the good work.

The Factor12 Rating (F12) is an analytic measurement utilizing league average performance to compare the value of all MLB pitchers. Basically, we are trying to measure ‘Pitcher A’ versus the League AVG Pitcher’s actual statistics. THANKS for the nice words and feel free to ask more questions about F12.

Hey Sven- I’m sure you guys have thought about this already, but what are your thoughts about adding weights to each of the factors in Factor12? For example, if the percentage difference of a pitcher’s SO/9 contributes to the likelihood of a win more than SO-BB, you would assign the SO/9 %diff a weight of ‘3’ and the SO-BB %diff a weight of ‘2’ and adjust the ultimate formula accordingly. Do you think this is reasonable/a useful factor in increasing the predictive power of a Factor12 rating?

Garner, thanks for writing…..

This is something we didn’t really consider very long. When we came up with the 12 Factors, we decided they should all have equal value. EXCEPT, IP and SO-BB. These two use a formula that does not have a cap, which gives extra credit to starting pitchers.

Also,

the SO-BB Factor is weighed. If a pitcher’s innings total is below the number of innings needed to be a “qualifier” for ERA title, then we multiply the SO-BB F12 number by (pitcher innings / qualifying innings). What this does is keep the relievers in, what we believe to be, their proper place in the rating.Back to your suggestion, the Defense Independent statistics, like BB, SO, HRA hold more value than say, ERA, which depends on the team defense. BUT, we believe that the combination of the 12 factors levels the playing field. Every aspect of pitching is taken into account at some point in the Factor12, from Stolen Bases Allowed, balks, runs, walks, Ks, wild pitches. Everything, we think. We think this tends to even out pitchers that may be outliers in certain categories, but fail in others.

Also, in the future, I’d like to create a few different variations of F12. I’d like to run an NL only, AL only, and a park factor version.

And our biggest goal is to get this thing cranking out predictions for 2013.