|Analyzing next-level pitching stats|
|Theory and Strategy - Platinum|
|Written by Todd Zola|
|Tuesday, 19 June 2012 11:58|
Is this Brandon Beachy kid for real? What's wrong with Max Scherzer? Was Jeff Samardzija's recent shelling the beginning of the end for him? Is it time to finally give up on Phil Hughes?
The fantasy community is getting much better at discerning how much of a pitcher's result is due to skill (or lack thereof) and how much is simple happenstance. In that respect, Voros McCracken's DIPS (defense-independent pitching statistics) Theory is truly epic. The fact that a pitcher has, at best, limited control over the fate of a batted ball has forever altered the way we look at the game of baseball and has implications that transcend deep into the fantasy realm. If you are not incorporating at least some of the advanced means of analyzing pitching in your team management, you are at a competitive disadvantage.
Today, we will put pitching "under the microscope" and review the first level of this analysis. We will then put the theory into action by looking at 10 starting pitchers in terms of what they have done and what they are likely to do going forward.
Skill versus luck
It used to be that a pitcher's ERA was the chief indicator of his talent. We now know to separate the skills from luck and focus on the skills as the primary means of projecting future performance. If a hurler has been lucky or unlucky, we anticipate regression, which means that going forward, the pitcher should realize neutral luck. What follows is a brief review of the components that influence what the pitcher can and cannot control.
Strikeouts and walks
These two elements will be discussed as a pair, as they are the entities for which the pitcher exerts the most control. Below is a three-year review of strikeout and walk rates broken out by starter and reliever, with a look at where we stand so far in 2012:
Other than to corroborate that pitching is indeed on the upswing (or hitting is declining), the primary purpose of this chart is to provide a number we can use to compare to the league-average. Perhaps the initial mainstream application of advanced pitching metrics in fantasy analysis was the LIMA (Low Investment Mound Aces) plan, which was the brainchild of fantasy baseball pioneer Ron Shandler. While this is an oversimplification, a filter of 6.0 K/9 was set as the minimum acceptable level for a viable "LIMA" pitcher. Even if we don't directly follow the LIMA Plan, 6.0 has become our internal litmus test. Back in 2002, the league-average K/9 for starting pitching was 6.2, almost a full punchout lower than today. It's time we adjusted our eyeball filter to 7.0 to match the league-average. For what it's worth, the BB/9 for starters 10 years ago was 3.2, basically the same as now.
Although the metrics K/9 and BB/9 are conventionally used, K% and BB% might in fact be better for precision. K% and BB% is strikeout or walk per plate appearance. As will be discussed shortly, sometimes a pitcher enjoys a better-than-average rate of retiring hitters on batted balls in play. Not only does the pitcher need fewer outs to get out of the inning, he also is facing fewer batters. Therefore, if a pitcher has been fortunate with respect to hits on balls in play, his K/9 may be lower, even though his K% is the same. The reverse is true if a greater-than-normal number of batted balls become hits. A pitcher needs that extra out to finish the inning, which could lead to a higher K/9 but a similar K%. The same holds true for walks.
Years ago, a pitcher's ability to limit homers was thought to be primarily a skill. More recently, it has been suggested that there is also a great deal of fate involved. Here is a look at home run rates from the previous three seasons compared to thus far in 2012:
In a nutshell, the number of homers a pitcher surrenders is a function of the number of fly balls given up and the percentage of fly balls that clear the fence (HR/FB). The number of fly balls is certainly within the pitcher's control. The best means to limit fly balls is minimizing contact, while inducing as many grounders as possible.
HR/FB has been shown to be out of the pitcher's control. The past few seasons, the league-average has been about 10 percent. So far in the present campaign, it's 11 percent, but since this number is influenced by weather and venue, we need to complete the season before drawing any conclusions. If a pitcher is above this mark, they have been unfortunate with respect to homers allowed, and vice versa.
Perhaps the best way to explain HR/9 is if every pitcher worked in a neutral park, unaffected by weather, their seasonal HR/FB should be 10-11 percent. Their actual HR/FB is influenced by their home park, and the ensuing HR/9 impacted by their ground ball rate.
Batting average on balls in play (BABIP)
BABIP emanates directly from McCracken's DIPS Theory cited above. Pitching analysis forever changed once it was demonstrated that whether a fairly batted ball went for a hit or an out was largely happenstance. Here is the league-average BABIP from 2009-2011, along with the present mark this year:
Keeping in mind that DIPS theory suggests the fate of a batted ball in play is largely luck, the fact that these numbers display the same pattern as K/9 and BB/9 is at minimum curious, with the possibility there is more skill involved than is presently believed. That said, improved defense and positioning is also a possibility.
Since fly balls that stay in the yard are caught at a higher frequency than ground balls are converted into outs, a fly ball pitcher should sport a WHIP lower than league-average. However, the recent trend is more ground balls are being induced, which further suggests the BABIP trend may be a result of overall increased pitching skills.
BABIP is worthy of its own column -- those details will be shared another time -- but here's a brief argument in favor of BABIP being more of a skill than is presently thought. If BABIP is indeed a skill, it should display the same patterns as K/9 and BB/9. If you examine handedness of matchup and home versus away, BABIP indeed has identical traits as K/9 and BB/9 rates. To wit, a southpaw facing a lefty hitter sports the highest K/9 and the lowest BABIP. A righty facing a lefty batter has the worst K/9 and highest BABIP. The home K/9 exceeds that on the road, while the home BABIP is a few points lower than on the road. There is growing evidence that a pitcher, in fact, exerts a measure of control over BABIP, beyond that influenced by fly ball-versus-grounder, and it is something I am comfortable using in analysis.
Ground ball rate
As already discussed, ground ball pitchers tend to allow more hits, but serve up fewer homers. The result is an extreme ground ball pitcher will sport an ERA lower than the league-average (good for the real game and fantasy) but a higher WHIP (not so good for fantasy). An extreme fly ball pitcher will give up fewer hits but more homers, resulting in a higher ERA (bad for real baseball and fantasy) but a lower WHIP (good for fantasy). A fly ball pitcher carrying a K/9 above and a BB/9 lower than the norm, while pitching half his games in a big park with strong outfield defense, can be just as effective as a ground ball specialist.
For comparison purposes, here are the league-average ground ball rates since 2006:
The next level
This also is a topic deserving its own treatment, but here is a cursory look at where pitching analysis is headed. The primary reasons this type of analysis is not yet in the mainstream is due to a lack of readily available data, along with ample years of the necessary data, to statistically draw significant conclusions.
1. Batted-ball data is being further classified by how well it was struck: hard, medium and weak contact. As expected, BABIPs differ according to how well the ball is hit, with hard-hit balls obviously carrying the highest number. If it can be proven a pitcher can control the type of contact, a lower BABIP can indeed be deemed a skill.
2. Pitches are now being tracked as balls, called strikes and swinging strikes, with the pitch being in and out of the strike zone. Contact is also tracked on pitches in and out of the zone. One application of this could be to help determine if a pitcher's strikeout rate that's different from his career norm is a new skill level, good or bad. There are also implications regarding quality of contact.
3. Electronic collection and computer analysis has refined the archiving of the speed, break and classification of a pitched ball. This will add to the arsenal for evaluation, delineating a new skill level from randomness.
Putting theory into action
What follows is an evaluation of the performance of selected starting pitchers, chosen to best illustrate the tools described herein. Specimens are restricted to starters since the small number of innings registered by relievers tends to exaggerate the luck element. Stats are through games as of Sunday, June 10.
James McDonald, Pittsburgh Pirates(2.72 ERA, 1.00 WHIP, 8.8 K/9, 2.6 BB/9, .48 HR/9, 39.7 GB%, 5.2% HR/FB, .257 BABIP): McDonald has been both good and lucky. His BB/9 has been especially stellar, as it's well below his career mark of 3.8. The K/9 is less surprising, although it's still a career high. While McDonald's BABIP is well below the league norm, meaning he has been a recipient of some good fortune, his extremely low HR/FB is the real charm. Since the Pirates' new ace is a fly ball pitcher, this luck has an extra impact, resulting in a sub-3.00 ERA.
There's no doubt McDonald is a different pitcher than previous seasons, but it remains to be seen if he can sustain his newfound control. In addition, while it has been suggested that not all of the reason for the low BABIP and HR/FB is luck, some regression must be anticipated. While McDonald is likely going to be the poster child for selling high, my inclination is to buy high, feeling that while he will see his ERA climb, there is something positive going on here, and the correction will not be as steep as some may assume. At least some portion of McDonald's low BABIP and HR/FB is the result of good pitching, which should remain even after his hit and home run rates regress.
Jeff Samardzija, Chicago Cubs (3.96 ERA, 1.35 WHIP, 8.9K/9, 3.1 BB/9, .62 HR/9, 47.5 GB%, 8.3% HR/FB, .332 BABIP): After the recent pasting Samardzija took at the hands of the crosstown rival Chicago White Sox, many are wondering if his 15 minutes of fame are over and the wild Samardzija resurfaces. In contrast to McDonald, Samardzija has been victimized by an unlucky hit rate, although his home run rate teeters on the lucky side. He has also reduced his free passes, while continuing to rack up the whiffs.
There's just something a bit more tenuous about what Samardzija is doing as compared to McDonald. There is a very real chance Samardzija is an improved hurler; it just feels like he is less likely to sustain his newfound control than McDonald. That said, his hit rate is high and should regress, helping to compensate for the chance of increased free passes. Samardzija is not a guy I would target unless he came at a discount and you need to take a risk, as the reward could be substantial.
Gio Gonzalez, Washington Nationals (2.35 ERA, 1.00 WHIP, 11.0 K/9, 3.7 BB/9, .12 HR/9, 49.1 GB%, 1.9% HR/FB, .244 BABIP): An improvement in peripherals was expected for Gonzalez as he moved to the Senior Circuit. That said, his strikeout rate is even higher than expected. Some of that is because he's facing hitters who aren't yet familiar with his stuff and delivery; even so, Gonzalez owners are rightfully thrilled with his performance so far.
The primary incongruity is his HR/FB rate. As in, there's no way that can remain so miniscule. Expect a few more big flies as the season wears on, with a corresponding ERA correction. And as suggested, Gonzalez is likely to see a dip in K/9 rate as the season progresses.
While I would not be eager to sell high on Gonzalez because he's going to continue to be a top starter even when his K/9 and HR/9 rates adjust, if you can fetch a top bat, now might be the time to do so. I'm talking about obtaining a hitter of the ilk that a Cole Hamels or C.J. Wilson would bring you, as I expect Hamels and Wilson to be better than Gonzalez going forward.
Brandon Beachy, Atlanta Braves (1.98 ERA, 1.00 WHIP, 7.3 K/9, 3.3 BB/9, .70 HR/9, 41.5 GB%, 7.0% HR/FB, .206 BABIP): Sell, Mortimer, sell! In the spring, Beachy was on a ton of sleeper lists for his exceptional strikeout potential. Despite a precipitous decline in strikeouts this season, Beachy is sporting one of the best ERAs in the league. The thing is, it's completely a result of an obscenely low BABIP that is going to correct. This is just old-fashioned good, make that great, luck.
Unless Beachy is able to regain his strikeout prowess (his K percentage has dropped the same amount as his K/9 rate, so the decline is real), his ERA is going to double. Did I mention you should sell?
Daniel Hudson, Arizona Diamondbacks (5.79 ERA, 1.53 WHIP, 5.8 K/9, 2.8 BB/9, 1.4 HR/9, 40.7 GB%, 14.3% HR/FB, .333 BABIP): If good pitching can positively influence BABIP and HR/FB, then bad pitching can exert a negative effect, which is likely happening here. That is, not all of Hudson's elevated BABIP and HR/FB is just bad luck. The wild card is how much of this poor performance is injury-related, although the instances where health is not a concern are few and far between.
Historically, Hudson is a guy who whiffs more and walks fewer than he has this season, so there's a strong chance that, as he continues to recover from early shoulder woes, he will return to form, a mid-3s ERA guy. But the problem is, even when healthy, Hudson is not especially dominant, as his career K/9 is just 7.0, which we now know is only league-average. His margin of error is thus reduced.
If I owned Hudson and was able to stash him on reserve until his K/9 ramped back up, I am reasonably confident he will have a much better second half. However, he is not the type to target, even if you are in dire need of a shot in the arm and buying low, as his upside is limited. Samardzija would be a better option.
Phil Hughes, New York Yankees (4.76 ERA, 1.34 WHIP, 8.3 K/9, 2.4 BB/9, 1.99 HR/9, 31.9 GB%, 14.3% HR/FB, .293 BABIP): My first reaction to Hughes' line is if you put this guy in a big park, you'd have a Cy Young Award candidate. His strikeout and walk rates are both better than league-average. His BABIP is right where it's supposed to be. But wow, those home runs are unsightly.
Hughes has everything going against him with respect to HR/9. He's a fly ball pitcher with a launching pad for a home park. His high HR/FB rate is doing him no favors. That mark should be above league-average due to Yankee Stadium, but not that much.
The park limits Hughes' fantasy value, but if situated properly on a fantasy staff, he can be quite helpful. The key is planning your staff such that Hughes' ERA and WHIP are akin to a third, fourth or even fifth fantasy starter, but he is more valuable than your typical mid-staff guy due to his greater wins and strikeout potential. So if you can deal for Hughes at the cost of a back-end starter and don't expect too much in terms of ratios, you can come out ahead. Not all your pitchers are going to have a great ratio; some are going to be mediocre. At least with Hughes, you get the greater potential of help in other areas.
Max Scherzer, Detroit Tigers (5.88 ERA, 1.59 WHIP, 11.2 K/9, 3.4 BB/9, 1.8 HR/9, 37.6 GB%, 18.1% HR/FB, .378 BABIP): There might not be a more maddening pitcher on the planet to dissect than Scherzer. While his walk rate is a touch above league-average when compared to his strikeout rate, it is fine, suggesting strong skills. But his BABIP and HR/FB suggest there has to be some bad pitching happening as well.
Scherzer is likely a prime candidate for the next level of analysis to try to figure out what's really occurring. Something that sticks out is, while Scherzer's fastball velocity is normal, he is throwing his slider and changeup three mph faster than usual. In this instance, faster is not better, especially regarding the changeup. The increased velocity on his offerings could be inhibiting movement, and it's surely not disrupting timing as much as normal.
I would not suggest acquiring Scherzer unless you can stash him away. If you are able to stash him, the indicator I would track is how fast he is throwing his off-speed stuff. Those strikeouts may look good, but right now they come with too much collateral damage.
Felix Doubront, Boston Red Sox (4.34 ERA, 1.42 WHIP, 9.8 K/9, 3.7 BB/9, 1.2 HR/9, 43.1 GB%, 14.8% HR/FB, .317 BABIP): Doubront is different than the other pitchers being examined in that not much was expected from him, so you have a bit of wiggle room with respect to what you actually get if you invested early. The more pertinent question: Is he worthy of a waiver pickup if you are in need of strikeouts and wins?
Doubront's K/9 is obviously excellent. The BB/9 is high, but within reason, considering the stellar K rate. With his limited track record, it's necessary to question whether Doubront will be able to sustain such a lofty K rate as he makes his way through the league a few more times.
That brings us to his BABIP and HR/FB and the conundrum of how much is bad luck and how much is bad pitching? Though Doubront's BABIP is high, it is not exceedingly so, especially since the Green Monster in Fenway Park turns outs into hits, elevating BABIP. The HR/FB, however, is disturbing and probably a result of some fat pitches coming from an inexperienced left arm.
If you took a flier on Doubront in the spring and can spot him in favorable matchups, he can be useful to a degree, since the expectation is low. But even if you need strikeouts, he's too risky too trust until he shows he can avoid the home run ball.
Ricky Romero, Toronto Blue Jays (4.15 ERA, 1.32 WHIP, 6.7 K/9, 4.5 BB/9, 1.1 HR/9, 54.7 GB%, 15.6% HR/FB, .248 BABIP): There are some that will contend that all that's happening to Romero is payback for the good fortune he enjoyed last season. Unfortunately, there is more to it than that, and it's not particularly encouraging.
Romero's biggest issue is giving back all the gains he made with respect to control. His walk rate has ballooned to a career high, while he has fanned fewer. So even if the fantasy gods are getting even with him, he is not helping matters with his free passes. Even when he's on, Romero is a below-average strikeout pitcher. If he is contributing below-average ratios, Romero is a fantasy detriment. Many drafted the Blue Jays southpaw as their third or fourth starter and have been sorely disappointed with little hope for a turnaround due to his lack of dominance. If possible, Romero should be benched until he demonstrates the ability to throw more strikes.
Thus concludes today's examination. If there is a pitcher you would like analyzed, please post the name in Conversation (below), and I will do my best to drop by the lab and fire up my
|Last Updated on Tuesday, 19 June 2012 12:18|