We've gotten into a debate lately with MDS, of Million DollarSleeper.com, on the credits of BABIP and HR/FB with an interest into its effects on calculating FIP and xFIP. I want to take the time to go over what data is out there and where I think this stands. I will restate that I believe pitchers have little to no control over these two stats, and this possible control can be ignored once you consider year-to-year variation in a fantasy setting.
Controlling BABIP
The idea of BABIP largely began from Voros McCracken, who proposed the idea that pitchers had no control over the result of a batted ball in play. He has backed off slightly in recent years on this statement, but the leading ERA calculations, FIP and xFIP, take this largely into account.
One of the first to question this with actual research was someone MDS mentioned in his article: Tom Tippett. I like Tom, and some of his stuff is great, but I have some issues with this research. The biggest problem is he used teammates to compare pitchers to account for defense and park factors.
That leads to quite a few possible problems. That assumes that pitchers give up the same type of hits, but we all know that a groundball pitcher will have to count on different parts of the defense than a flyball pitcher. The same could be said for control pitcher vs hard thrower or lefty vs righty.
Not only do different pitchers have to deal with different parts of the defense, but they have to deal with different parts of the ballpark. I understand what Tom was trying to do, but this still didn't prove that pitchers had a large effect on BABIP.
There is one thing that has changed since the article. Knuckleballers have been shown to beat the BABIP as well as HR/FB. This is fairly evident when you look at Tom's best BABIP example, Charlie Hough, and then close behind was Phil Niekro.
Outside of the knuckleballers, there is little evidence that pitchers can control BABIP. The pitchers with the best BABIP over the past three years all have pitched for the best defenses. Matt Cain not only has a great HR/FB, which we will discuss later, but also a great BABIP. It just happens that the Giants have had one of the best team UZR/150 in all those years.
The team defense is number one on BABIP, and by the time you have enough data to show a pitcher has real BABIP skill, he's past his prime. Trying to pick pitchers based on BABIP skill a) does not have a great return if you're correct and b) is like playing the lottery if it's not backed by team defense.
Understanding DIPS - Here is a solid discussion of the limited influence of pitchers on BABIP.
Controlling HR/FB%
Moving on to HR/FB, there has been a lot of content on this recently. Over at FanGraphs there has been the return of the Matt Cain discussion as well as a Dave Righetti discussion. These studies always try to find how Matt can control this, as well as the whole Giants staff, but I think they always fail to account for one thing or another.
In this case park factors were taken for home field, but we should remember these pitchers get an extra few games in Dodger Stadium and PETCO park each year to help. That's not the big one, though. The big issue is the power of the hitters in the NL West as well.
The NL West had only 4 players with HR/FB over 15% and who did not play in Colorado. The Padres and Dodgers both rank in the bottom of the league in HR/FB% and also have most of their power in right handed bats. This would lead me to believe if Dave Cameron had taken into account the NL West hitters, he might have found more reason for right handers in San Francisco to beat the HR/FB% average.
So the debate continued yesterday with Jesse Wolfersberger looking at factors that affect HR/FB. He certainly showed some interesting data, but how much these numbers change the overall HR/FB wasn't explained. It's been looked at in this way before, and as shown by his final r-squared of 10 percent, it shows the effect is small, as we already knew.
Conclusion
Reminding ourselves that pitchers can control this stuff is always good, but also remember that this effect is so small it's not something that's usable for a single season of data. The variability of BABIP, HR/FB and ERA is large enough to make this control irrelevant to a fantasy owner.
In the long run, knowing who could be a pitcher capable of controlling these numbers would be good, but they can only keep this skill so long. The big swings like from Matt Cain's and Clayton Kershaw's HR/FB are mostly park factors. If you understand this, you can take advantage of them, but also be cautious of pitchers being traded.
If Cain was sent to another team like the Phillies or Blue Jays, his HR/FB would reach at least the average HR/FB%, if not exceed it in those home parks. This would negate any value he had and make him the 4.00 ERA pitcher his xFIP suggests.
Follow the new @RotoSavants tweet account.
