Tuesday, May 29, 2007

Luck...and a little of bad luck

Somewhere, Billy Beane cringes a little each time "luck" is mentioned.

Sabermetrics has attempted for some time to quantify luck. Instead of simply saying Player A is hitting the ball well or just not seeing the ball, the numbers may show that one player is simply lucky or unlucky. It allows us a chance to prepare for a possible and pending evening out of a lucky player while staying upbeat about a struggling player.

A little sidenote, Hardball Times hasn't updated their numbers for yesterday's games so these numbers may be slightly off.

Two stats, one for pitchers and one for hitters, do that best I feel. First, we start on the mound. Fielding Independent Pitching, or FIP, tries to give an ERAish stat to a pitcher based on what he has the most control over - strikeouts, walks, homers, etc. - over hurting him because of his defense or whatever. FIP is exactly like ERA so a 3.00 FIP or lower is good and a 4.50 or higher FIP is not so good. For instance, Jorge Sosa's FIP in 2005 was 4.13, a good run and a half higher than his ERA, while his FIP last year with the Braves was around 5.83, pretty near his actual ERA of 5.46. Sosa was pitching lucky in 2005 but his numbers more accurately described his suckiness that year. For those scoring at home (congrats!), Sosa's ERA right now is 3.64. His FIP is 4.45. Will he put up lucky numbers again? Doubtful.

FIP can be used in direction relation with ERA, showing its best value. A negative number leads to one calling that pitcher "unlucky" while a positive number leads you to call that pitcher "lucky." Using a half-run either way as random ups and downs in the data, here are the lucky pitchers and unlucky pitchers.


















Name FIP ERA FIP-ERA
Chad Paronto 4.74 8.16 -3.42
Tyler Yates 2.59 4.13 -1.53

Yeah, that's about it. Two unlucky pitchers. Mark Redman's also on this list, but being how he's not a Brave anymore, I just don't care to mention the Indian more than I have to. He did just sign with the Texas Rangers proving that in the 21st century, cowboys and indians do get along.

On one side, Tyler Yates unluckiness is a bit overstated. His expected FIP is 3.49 so he's probably not mentioning up to that 2.59 FIP, but his ERA is more telling of how useless it is to look much at a reliever's ERA and draw anything from it. On the other hand, Paronto has been "unlucky" but still pretty bad.

How scary is this? Well, if those are your only two unlucky pitchers, how many lucky pitchers are we throwing out there? Well, a great bulk of players fall in the half-run higher or lower catagory, including: Kyle Davies, Oscar Villarreal, Rafael Soriano, Chuck James, Tim Hudson, and Bob Wickman. So really, what you get from those guys is pretty clear. That leaves three more players to mention. I am leaving out McBride (a fourth lucky pitcher) because of his lack of innings.
























Name FIP ERA FIP-ERA
John Smoltz 3.22 2.82 0.64
Mike Gonzalez 3.02 1.59 1.43
Peter Moylan 4.47 2.25 2.22


Smoltzie is still pretty hot so that's not much of a worry. Gonzo's ERA just needed to even out from its way too low number. On the other hand, Moylan's number is pretty telling. His number basically finds its origins in his LOB%, the same number that helped Sosa put up a sub 3.00 ERA. 85.6% of baserunners to reach against Moylan were left on base. He may be able to continue that, but it's doubtful. Eventually, runs start to score.

On the flipside and the focus of a future column...PrOPS or predicted OPS. Look for it during the week.

No comments: