Monday Morning Ten Pac... (05/13)
Sporer Report: Has Coo... (05/21)
the archives are now free.
4 comments have been left for this article. BP Comment Quick Links
| |
|
|
May 14, 2013 Sporer ReportEvaluating Early-Season Outliers
We are about to hit a big date in the fantasy baseball world: May 15. While it doesn’t carry any inherent significance on the calendar (unless you’re John Smoltz, George Brett, Josh Beckett, Justin Morneau, Michael Brantley, Brian Dozier, or Brandon Barnes; happy birthday, gents!), it is another mile marker on the fantasy landscape, as many see it as a point beyond which they should begin addressing needs and making serious moves with their ballclubs. With the general chorus of advice being that you should be patient with the players you drafted, there are a handful of these early mile markers at which fantasy managers feel they have been patient enough based on their perception of what it means. The first is usually Tax Day, April 15. Yes, two weeks is long enough for some, and then it’s time to get in there and start rearranging the puzzle pieces. For others, May 1 is a big day. Essentially a sixth of the season has passed, we’ve flipped the calendar, and now trading large chunks of your original squad doesn’t seem like such a panic move. The other two biggies are May 15 and Memorial Day (or June 1, but those are close enough that we can lump ‘em together). Not only are these markers used for deciding whether or not you should make that big eight-player deal with Fred from accounting, but the fantasy community also uses these dates to start judging the legitimacy of performances. Whether they are going by gut and just saying that mid-May is an adequate sample for them to pass judgment, or if they are more scientific about it and adhere to Russell Carleton’s excellent work on sample sizes, even the most patient fantasy managers are ready to start believing in hot starts or freaking out about cold ones. With that, I’d like to dive into the performances of two surprise standout pitchers who have burst onto the fantasy landscape so far in 2013 and see whether or not we have something worth believing in or merely a flash in the pan. Travis Wood, CHC (2.03 ERA, 0.90 WHIP in 53 1/3 innings) After another brilliant start on Monday night, Wood now has a quality start in all eight of his tries having allowed more than two runs just once (in 7 2/3 against the Padres). His component numbers haven’t been terribly special with a 17.7 percent strikeout rate and 8.4 percent walk rate, which yield just a 2.1 K:BB. However, Wood does have an 83 percent left-on-base rate and a .186 BABIP so far, both of which exceed league-average rates by leaps and bounds. Now, one of my least favorite things on the fantasy baseball landscape of analysis is when someone simply looks at these two rates, sees that they are much better than or much worse than league average and instantly suggests that the pitcher has been lucky or unlucky. It’s as annoying as it is lazy, and it doesn’t do anyone a single bit of good. Yes, Wood is living outside of the norms, but he also has components within his profile that give an idea as to why he has been able to eclipse the average by such a wide margin. With regard to his BABIP, you should first look at the pitcher’s batted-ball profile before assigning gobs of luck or handfuls of misfortune. In the case of Wood, his .244 BABIP from 2012 was a bit depressed by the fact that he allowed a ton of home runs, which don’t count against the metric. In 2013, his homers are way down, his line-drive rate is down by seven percent, and his 45 percent fly-ball rate includes a hefty 15 percent infield-fly-ball rate. Line drives are mostly likely to turn into hits, whereas infield fliers are the least likely batted balls to become a hits. As for the left-on-base rate, it is a lot easier to sustain a reliever-esque 83 percent mark when you are putting next to no one on base in the first place. The LOB rate is more a byproduct of the tiny BABIP and resulting 5.2 H/9 rate than anything else. I don’t look at Wood’s success to date and think he has simply been a luckbox. He is pitching incredibly well right now, and he’s earned the golden numbers he has through eight starts. Sustaining this level of success, however, will be a different matter altogether. He’s done a great job inducing poor contact, but he allows far too much of it to sustain this level of excellence. First off, no pitcher has a qualifying ERA season (162+ IP) with a hit rate better than 6.24 H/9 since 2010 (Justin Verlander, 2011), and then only 12 pitchers have been at seven or better in that same time period. Of those 12, only Jeremy Hellickson had a strikeout rate under 21 percent. He pulled it off with a 15 percent mark while the majority of the group sat between 23 and 27 percent. Remember that when we are projecting, we are relying on probabilities (what is most likely to happen), and so while there is a non-zero chance that Wood stays excellent for a transcendent, record-breaking season, the probability that he regresses—given his high amounts of contact allowed and lack of overwhelming pure stuff—is quite a bit higher. The lesson: Don’t look at a guy’s ERA and WHIP, compare it against expectations/what you know of the pitcher, and automatically discount the success. Same goes for BABIP, LOB rate, and HR/FB being well above or well below league norms; that doesn’t automatically add up to severe luck one way or the other. Now, let’s look at another guy with some incredible early-season success whose profile tells a different story about it. Jeff Locke, PIT (2.95 ERA, 1.21 WHIP in 39 2/3 innings) Locke, like Wood, has outlier BABIP (.214) and LOB (81.5) rates, but a deeper look outlines clear trouble ahead. His batted-ball profile is a mess, starting with a 24 percent line-drive and… well, that’s it, but it’s so horrible that it’s enough to terrify me. He has a .567 BABIP on his line drives compared to a league average of .694, saving him four hits. In fact, all three of his batted-ball type (line, ground, fly) BABIPs are well below league average, saving him a total of 11 hits to date. Adding those hits to his ledger would send him from a 6.6 H/9 up to a 9.1 H/9. The reason I’m particularly skeptical of his BABIP is the exorbitant line-drive rate. There is just no way he can continue to let the league barrel him up like that and avoid trouble. It gets worse. If that weren’t bad enough, his 13.3 strikeout is not only terrible, but it’s far too close to his also-awful 11.5 percent walk rate, yielding a 1.2 K:BB. That is a recipe for disaster. Whereas there is a scenario Wood has a huge season and surprises us all, there is no way that Locke can keep this profile and stay anywhere near his sub-3.00 ERA. Recap On the surface, these are two similar guys: a couple of nondescript lefties who have seemed to absorb David Price’s ERA ability. But, a deeper look shows two vastly different roads to the early success with one having a far better chance at sustaining at least a modicum of it. Whether you’re looking at the success of unexpected arms like Wood and Locke or the shortcomings of established studs like Price and Matt Cain, you have to dig into the underlying numbers to get a better idea of what is going on with them. There are no one-size-fits-all rules of thumb where you can look at one metric, compare it against an average, and understand what is happening with the pitcher. The data is out there to help you become knowledgeable. Hopefully, this gives you an idea of how to best use it when analyzing early outliers.
Paul Sporer is an author of Baseball Prospectus. Follow @sporer
|
Paul,
Interesting article. I have one question, however---can a pitcher control the LD% against or is that just a luck factor? If it is just a luck factor, then I don't know if your analysis is valid---i.e. if Locke's high LD% is just as likely to fall as it is to continue then I wouldn't worry as much about his luck in avoid hits on the line drives so far. For example, Felix Hernandez's LD% has flucuated between 16.1 and 22.5 over his career without any apparent rhyme or reason. The pitcher with the lowest LD% last year was Trevor Cahill, and he was pretty mediocre. I remember reading several years back that pitchers did not control this (the writer's evidence was that Lincecum (in a Cy Young year) has a worse LD% than Sidney Ponson (in a normal Sidney Ponson year), and his theory was that it was a luck factor that tended to balance out (as opposed to say, hitter's LD%). Do you have any more scientific knowledge one way or the other as to whether pitcher LD% is a repeatable skill?
I have a hard time believing that guys squaring a pitcher up at a 25% clip is simply bad luck, especially when they don't have particularly impressive stuff like Locke. I feel like we've gotten away from the "pitchers have no control over their batted ball profile" notion in recent years as we've seen guys continually exert force on it whether via heavy GB rates or heavy FB rates (specifically heavy IFFB rates).
Definitely something to think about in-depth.
I think the evidence on GB% and FB% is pretty solid; guys consistently have a bigger percentage of one or the other. I wonder if LD% looks more random (it does from the stats) because there is more of a subjective element to what is a line drive and what isn't. That could add to randomization---some guys might have a higher LD% if their official scorer terms higher things line drives instead of fly balls when they are not particularly likely to be hits.
It is a question to explore. I note that the evidence goes both ways. Someone like King Felix is pretty random and he does well regardless, but then again, some names on the leaderboard for LD% like Joe Blanton obviously make one think it is a skill.