Saturday, November 29, 2014

Fielding Independent Pitching (FIP) is silly.

Late to the party?  Maybe.  I was minding my own business, reading The New York Times on my tablet, when I spotted this:

How Baseball Statistics Can Help Explain the Economy
NOV. 25, 2014 by Neil Irwin  The New York Times

Peripheral statistics are the more obscure indicators that correct for these kinds of quirks and aim to give you a richer and more truthful evaluation of a pitcher, taking into account factors like ballpark dimensions. (Baseball Prospectus combines some of those measures into an index it calls Peripheral Earned Run Average, or PERA.)

I foolishly clicked the link for Baseball Prospectus and yet another link to Glossary: Baseball Prospectus Exclusive:


Fielding Independent Pitching converts a pitcher's three true outcomes into an earned run average-like number. The formula is (13*HR+3*BB-2*K)/IP, plus a constant (usually around 3.2) to put it on the same scale as earned run average.

First the small stuff:

- Replace Innings Pitched (IP) with batters retired by the pitcher, i.e., at bats (AB) minus hits.  I once again call for the same stats to be used for both batters and pitchers.  For instance, batting average, not hits per nine innings.  Earned Run Average (ERA) is obviously a stat that is specific to pitchers but much and maybe most of the other stuff correlates directly to batter stats.  Use them for pitchers, too.

- Some home runs are inside the park and could be played by fielders.  Since the number is small in recent years it probably doesn't make much difference, like including hit by pitch (HBP) along with BB as some versions of FIP do.  In 1909 Ty Cobb had the triple crown, leading in home runs with 9, all inside the park.

Measuring fielding, especially on the team level, is a good thing.  Applying it to individual pitchers in this way seems silly.  More home runs allowed seems to improve a pitcher's FIP measurement.  Good, you say.  That's the point: home runs are out of the pitcher's control.  We're measuring the pitcher's fielding support.  Say what?  How about don't feed the batter a fat one down the middle?  But that also applies to rocket shots that do not go out of the park and have about as much chance of being caught as most home runs.  Example: 200 foot line drive landing on the foul line.  It's possible that a fielder could be placed there and catch the ball.  It's also possible to place a fielder at the outfield wall and that the fielder would catch a potential home run.

More to the point are fundamental attributes of home runs:

From the original document: Radical Baseball June 9, 2006 (posted February 20, 2008)

2. The Real scandal of the last 16 years: propagation of non-uniform playing areas.

Fenway Perk satellite view, March 9, 2007
by Betp [Public domain], from Wikimedia Commons
It’s not steroids. It’s the fences. Baseball is the only American team sport in which the playing area is not uniform. Imagine a National Basketball Association (NBA) game played at Madison Square Garden. The three-point line is drawn irregularly. A player can get three points by sinking a basket from behind that line but in some places the line is 25 feet from the basket and in some places it is 15 feet away. How about a National Football League (NFL) game played on a field where the sideline is wider in some places than in others? Or the end zone is shaped oddly? Silly, right? So how come baseball gets away with it? Baseball does not merely get away with it. It’s considered cute, charming, traditional, blah, blah, blah. Here’s the real travesty: the non-uniform playing area perverts baseball’s most cherished event: the home run. It undermines the very integrity of the game that is supposedly threatened by steroid use.

Some thinking fans categorize baseball events into random and non-random. To them a home run (one hit over a barrier on a fly, not an inside the park home run) is clearly a non-random event because a fielder has no chance to catch it. A home run is a random event. Here is why. Is a 180 foot fly ball a random event? Clearly, it is random: it may be caught or it may not. But how about a fly ball hit 380 feet? The non-random advocates would be forced to ask in what direction and in what park the fly ball was hit. In other words they can only certify its randomness by waiting until it lands. The same could be done for the 180 foot fly. Like the three point shot in basketball (OK, the line is closer at the sideline to fit in bounds but that’s basketball’s problem) the only thing that should matter is how far did the fly ball go. With uniform playing areas that alone would tell us if the fly ball is a home run or not.

However, in some cases a fly ball can travel 50% further than a home run and be an out. The distances to the barriers are not just different from park to park but they are different in some parts of the outfield in the same park. A home run should reward the batter for hitting a fly ball over a barrier and for that to be fair and meaningful the barrier should be the same distance and the same height in every direction in every park. That’s pretty basic stuff. How about 375 feet to a ten foot high barrier? If you were starting baseball today and making the rules, that’s clearly how you would do it. But baseball evolved and that’s how it has always been. So? About 13 new parks have been implemented in the last 16 years (with two more coming in New York) and baseball had a rare opportunity to correct this historic inequity. Instead it allowed and even encouraged teams to replace parks that were in many cases at least symmetrical with parks that were irregular in the shape of the playing areas. Irregularities were often unavoidable in old parks because of streets and other things that required some imagination in building a park. In recent years there were no such impositions, just a warped intent to make new parks that looked old fashioned. See the Rangers park in Texas, built in an open space.

Yes, this should also apply to foul territory. Here’s something no one has considered: Fenway Park helps strike outs. Because the area in foul territory is so small it is very difficult to foul out. Also, because the fences are close in the outfield, that also helps. Let’s say Roger Clemens is going for the single game strike out record and he’s pitching in Fenway Park, a hitter’s park. Every out that is not a strike out hurts this effort. Every batted ball that results in an out also hurts. A foul pop up that drifts into the stands helps. A ball hit off the wall in left also helps.

The single season and lifetime home run records are the most important sports records in America. Yet, they are subject to the greatest randomness of any records in team sports. Forget the steroids. Fix the fences.

No comments: