Putting Derek Lowe’s 2009 in Context
October 26, 2009 at 8:00 am by Capitol Avenue Club under Atlanta Braves, Pitching, Player Analysis, Statistical Analysis
Everyone knows (I hope) that pitchers’ ERA’s fluctuate a great deal more than their fundamental skills. Due to the pseudo-randomness of hitting singles, pitchers more or less have no control over whether or not a batted ball in play results in an out or a single. Doubles and triples are clearly different, but singles are–whether you want to believe it or not–mostly just luck. Whether a ground ball finds the hole between the 1B and the 2B or rolls right to the 2B, whether a shallow fly ball drops in front of the center fielder or falls into his glove, whether a broken-bat blooper gets over the SS or he’s able to make a jumping catch, these are all luck-based events. They’re lucky for a few reasons 1) the hitter wasn’t trying to do that. There wasn’t any skill, other than putting the ball in play, involved from a hitter’s standpoint. No hitter goes to the plate with the thought “I’m going to hit a weak fly ball and make it drop in front of the center fielder”. 2) It’s entirely up to the defense. The hitter and pitcher are out of the equation once the ball is put in play.
Therefore, the amount and rate of singles a pitcher gives up will fluctuate from year to year. Not because the pitcher has fundamentally gotten better or worse, but because of fluctuations of chance. Therefore, a pitcher’s ERA fluctuates similarly and parallel to the fluctuations in singles.
A metric called BABIP was created to model this phenomenon. Studies show that pitchers can’t control their BABIP, it’s a function of nothing they’re able to influence. It is calculated (H-HR)/(Balls put in play).
Which brings us to Derek Lowe and the hit-unlucky numbers he posted in 2009. They aren’t good, 4.67 ERA, 1.515 WHIP, 1.76 K/BB, etc. But Derek Lowe posted the 2nd highest BABIP in the NL in 2009–.333–leading to a league-high 232 hits allowed. While this isn’t a good thing by any means, it tells us something. If he can do everything he did and get a little luckier with the balls he allows in play next year, his numbers are likely to improve.
Bill James invented a metric called Component ERA, which attempts to determine what a pitcher’s ERA would be just from hits, walks, homers, etc. It’s abbreviated ERC. Derek Lowe’s ERC was 4.80 last season. However, if you adjust his BABIP from .333 to his career average (.295), he only allows 208 hits instead of 232. 24 hits doesn’t seem like a whole lot, but his ERC drops to 3.98 simply by removing those 24 hits. I’ve dubbed this aERC (adjusted ERC).
I think that 3.98 figure is probably what we should expect from Derek Lowe in 2010, though not just aERC, ERA as well. His FIP was 4.00 this season and if his fundamental skills hold steady and he has any luck at all, it’s reasonable to expect an ERA around 4.00.
2009′s numbers, on the surface, aren’t a true representation of Lowe’s fundamental skills. The smart money is on him bouncing back. Entropy exists in the universe, things have a natural tendency to regress towards the mean. The mean for Lowe isn’t 4.67, and we should see regression towards the mean–around 4.00–in 2010.
Loading ...
Just wanted to jump in off topic. Was just talking to a friend, who’s a friend of Tim Hudson’s wife’s, and according to her, he’s definitely staying with the Braves.
I would be extremely surprised if he goes anywhere else. Of course, that’s pretty much always been my stance.
Lowe wasn’t at bad as his ERA would suggest, but I fear you may be glossing over the fact that his peripherals are all headed in the wrong direction:
His K/9, after being above 6 the last 2 years, fell to 5.13.
His BB/9 in 2009 was his highest since 2004.
His GB% is dropping: 67.0, 65.0, 60.3, to 56.3 last year.
And the metric that combines these three: 2005-2008 xFIP in LA: 3.53, 3.80, 3.50, 3.43. 2009 in ATL: 4.18.
and yes, he had a high BABIP, but his HR/FB was also a little low, which depresses his FIP.
The other thing to consider for a pitcher like Lowe is the quality of his IF defense. Yunel is above average; the rest aren’t helping much.
I thought the Lowe signing was a good idea at the time (definitely preferred him to Burnett), and I’m still hopeful, but his rate stats are not encouraging.
So? The peripherals may be headed in the wrong direction. But there’s no evidence there’s a trend here and not just an anomaly. Plus, FIP and ERC both account for his 2009 peripherals.
He was fundamentally worse in 2009 than 2008, but it was more like half a run worse, not 1 and a half runs worse.
Hizouse,
I think your points reinforce that we overpayed for Lowe, and that he’s on a natural age-related decline. I think Peter’s point is that the decline is not one so precipitous as to be useless, that he’s even more useful than he appeared to be in 2009, that it would be more reasonable that 2010 look more like ’07 – ’08 (or really, maybe 2006). I think his point is that the data seems to show Lowe was exceptionally unlucky in ’09 on events outside of his control, and that if he reverts to something nearer his career norms, that we can expect a better Lowe, that we’re still over-paying some for.
Its a little troublesome that the K rate is dropping like that, but if the GB rate, even at 56.3, was the second best among starters in MLB this year. Joel Piniero paced everyone at 60.5, then Lowe, then Jason Marquis and Carpenter in the 55% range. So long as his GB rate is so high, Lowe is a rare sort of pitcher who can still produce at a high level with a middling K rate.
While we’re at it: Lowe’s 56.3 GB% in ’09 would have been good also for 2nd in the majors in ’08 behind Brandon Webb’s. Aaron Cook and Ubaldo Jiminez were 3 and 4 in ’08, Marquis #3 in ’09. Lotta groundballs going on in Colorado. Makes one understand why the A’s have been stocking up on young groundball-throwing SPs. They’re nearly as good as and cheaper than strikeouts.
Also while we’re here: Jurrjens had the 10th best GB% in the majors in ’08, 51.5%. Dropped a good bit this year, but if that pops back up, then with he, Lowe and Hudson we could have 3 of the top 10 – 15 groundball pitchers in the majors, along with two of the top K rate starters in Vazquez and Hanson. That’s a tremendous, tremendous depth of out-generation. Please, Frank, find some offense. Even if it does mean Dan Stinkin’ Uggla.
FIP doesn’t account for his shrinking GB rate, but it’s not a big deal here.
I guess my question is, can we trust in “his fundamental skills hold steady”? Dude is 36 (but a workout freak), and his peripherals have been headed in the wrong direction. I don’t know whether his true talent is better than the recent peripherals or if they will continue to decline with age; I’m looking forward to seeing some projections.
Peter, you mentioned BJ’s KJ projection the other day; does he have his pitching projections online?
At any rate, I pretty much agree with your and Nevin’s conclusions; 4 seems like a pretty good under/under for his ERA next year.
Nevin: good info re GB and K rates. Thanks.
Yeah, I posted the BJOL pitcher projections for the Braves on Friday in the Friday Links segment. There’s also a link to JC’s site where you can download the entire handbook. If you ask me, it’s not worth buying, but definitely worth downloading.
With regards to Derek Lowe’s GB rates, let’s take a look at his adjusted tRA. tRA is the only defense-independent pitching metric that treats each different type of batted ball differently. I’m a huge fan of the theory and anxious to see what they do when we have more hit F/X data.
Year – tRA+
2009 – 104
2008 – 132
2007 – 118
2006 – 121
2005 – 113
2004 – 109
2003 – 108
There’s not a clear pattern here. If we can conclude anything regarding trends, 2009 is an anomaly, there’s not a lot of evidence that his skills are in decline. The sharp decline in 2009 can be explained by his GB% which can be explained by the ineffectiveness of his sinker at times. While that’s not good, I don’t think it’s an uncorrectable problem. He experimented with a few different sinker grips this season because the ball was moving more laterally than vertically. I don’t think it’ll be a lingering problem in 2010 if it’s as simple as shifting the grip a bit. Plus he’s got one of the better sinker ballers in the history of the game as his pitching coach.
Even still, his tRA was above-average in 2009. A 104 ERA+ is perfectly respectable. A few NL players that posted ERA+ around 104 in 2009:
Ross Ohlendorf – 3.92 ERA, 103 ERA+
Chad Billingsley – 4.03 ERA, 103 ERA+
Jason Hammel – 4.33 ERA, 104 ERA+ (Coors Field)
Joe Blanton – 4.05 ERA, 106 ERA+
Barry Zito – 4.03 ERA, 106 ERA+
I’m trying to decide whether and if so how GB rate is or isn’t accounted for, and looking now at the formula for FIP (and enjoying myself):
(HR*13+(BB+HBP-IBB)*3-K*2)/IP, plus a league-specific factor (usually around 3.2) to round out the number to an equivalent ERA number.
[tip of the hat: Hardball Times / TangoTiger]
(or “H/T: THT / TT)
So, it explicitly omits all batted ball types, since balls in play, other than HRs, have results outside of the pitcher’s control. So you’re correct, FIP/xFIP itself does not account for a changing GB rate, and so we might question its usefulness as predictive datapoint here.
So… its usefulness is in helping me understand what’s going on– or what has gone on– with Lowe.
His ERA (4.67); we can see an ERA-like number that more directly drills in on what he was responsible for (xFIP: 4.18); we can see the ERA-like numbers that he’s been responsible for career-wise and recent-year wise (as you provide, since 05: 3.53, 3.80, 3.50, 3.43, 4.18);
and we can see his BABIP for recent years.
So, the high BABIP doesn’t impact xFIP, but does serve to give reason to believe that his real ERA later will be closer to his current FIP or xFIP. Which, I think we all three agree, is in the 4-ish range. And his xFIP would seem to be higher than in recent years due to the higher walk rate.
Since moving to a full-time SP, he’s had K/9 rates as low as 5.2 (’02, when he was 3rd in Cy Young balloting), 4.9 (’03), 5.2 (’04), 5.1 (’06), 5.1 (’09). So while his K rates have fallen from ’07 to ’08 to ’09, those first tow years represented career highs, but some margin, since he became a starter, and what we’re seeing is not so removed from his norm (as a starter. Career number, 5.8, is skewed from his time as a reliever. But if you use the Rule of 17 to normalise his RP numbers, you come right on at a career K/9 of 5.1, right were we saw him).
So… we’re seeing the real Derek Lowe, albeit a slightly unlucky one. We should expect to see his ERA improve to closer to his FIP/xFIP, and if he improves his walk rate, which was poor by his standards, he could be even better, sub-4. And since I believe there’s a trend for groundballers to regularly outperform their FIP due to the extreme low run value of groundballs, something in the 3.8 realm wouldn’t be out of line.
Also, I’m paid too well to be slagging off this much. If I could put all this into my real job, I could probably afford to better educate my children some day.
Sorry, kids. Daddy’s AD…Hey look! A Ball!
Well, FIP does account for his BABIP in a round-about way. It completely disregards BABIP. If two players posts the same K, BB, IBB, and HR rates but one posts a .333 BABIP and one posts a .265 BABIP, their FIP is the exact same. The one with the .333 FIP will most likely have a much higher ERA, though. The BABIP is expressed in his ERA, but not the FIP, so large differences indicate a player will regress in either direction, away from his ERA and towards his FIP.
Just a little clarification: the only factors that go into FIP are K, BB, and HR. So it does not count balls in play at all.
xFIP is slightly different; it replaces the HR factor with a factor based on fly ball % (it assumes a league average HR/FB rate). so xFIP would reflect a lower GB%, to the extent the FB% is increased (and not LD%, I think). And since there’s generally a lower average on fly balls, it would indirectly (and weakly) bear some correlation to BABIP.
BP has a similar metric, I forget what it’s called, that uses GB% as the third factor instead of FB%.
tRA is the more complex stat, as Peter describes.
is increased BP has a similar metric.
Here is a helpful article discussing how well each predicts ERA: http://www.hardballtimes.com/main/article/how-well-can-we-predict-era/