Regression and Maintenance: The story of Javy and Andy

This morning Joe took us through the tale of Frankie Cervelli and Brett Gardner, in which two young Yankee players came out of the gate hot but have since gone in different directions. Cervelli’s not the only player on the team to experience some sort of regression and Gardner’s not the only one to have maintained an unexpected level of performance. Two veteran members of the starting rotation also fit the bill.

The word “regression” has a negative connotation to it, but remember that in the world of statistics it basically means reverting to the mean. That can be a positive thing, such as a player who performed poorly early improving later on. To steal Joe’s example, think Mark Teixeira last year.

At the beginning of the season, Javy Vazquez was giving the Yankees exactly the opposite of what they had expected. He was getting hit around and doing everything but soaking up innings, and it was quite ugly at times. Andy Pettitte on the other hand, managed to exceed all expectations and emerge as the team’s best starter. Cervelli and Gardner started at basically the same place in terms of performance before hitting the fork in the road, but Vazquez and Pettitte did the opposite. They started at different ends of the spectrum and met up later on.

Photo Credit: Paul J. Bereswill, AP

When Javy struggled early this season, the all too simple “he can’t handle New York” narrative was everywhere we looked. He was booed unmercifully at home, not only because he was pitching poorly in 2010, but also because of the perception that he cost the Yankees the 2004 pennant. Vazquez has since done more than just right the ship, he’s been the team’s best pitcher for the last month, compiling a 2.93 ERA with a .596 OPS against. Prior to his start on Saturday, the Yankees had scored just 11 runs in Vazquez’s last five starts, but he still managed to win three of those games.

During his first five starts, Javy put a total of 39 men on base in 23 IP, which is certainly a ton. His batting average on balls in play stood at .358, but even more damning was a ~55% strand rate, an unfathomably low number that certainly contributed to his 9.78 ERA. His velocity had dipped from his usual low-90’s into the upper-80’s and he appeared to nibble in an attempt to compensate, walking 5.87 men per nine innings in his first five starts (more than double his 2.4 career mark). Twenty-three innings is about 11% of a typical season for Vazquez, a bona fide workhorse with a track record of being no worse than a league average strikeout machine.

Clearly Vazquez wasn’t pitching to his capabilities, but he wasn’t getting any help either. His BABIP and LOB% luck were simply terrible, unsustainably bad. Sure enough, what followed was a combination of Javy appearing to regain confidence and a statistical correction. In his six starts since (we’re throwing away that one batter relief appearance), he’s posted a .215 BABIP and a ~80% LOB%, bringing his season totals to a much more respectable .269 and 70.8%, respectively. The luck literally went from one extreme to the other.

Vazquez’s track record is long enough that we pretty much know what to expect. He’s been a dead average (literally, 100 ERA+) pitcher in the American League, and we’ll gladly take that going forward. Javy isn’t as bad as he was in his first five starts and he isn’t as good as he has been in his last six — his true talent level likely lies somewhere in between. And there ain’t a damn thing wrong with that.

Photo Credit: Kathy Willens, AP

A soon-to-be 38-year-old starter in the AL East isn’t supposed to be a staff ace. He’s supposed to be a veteran leader that gives innings and keeps the team in the game. Expect anything more and you’re likely to be disappointed. Well, unless that pitcher is the 2010 version of Andy Pettitte.

After three ever so slightly better than league average campaigns (106 ERA+) since coming back to the Yanks, Pettitte ripped a page out of the 1997 playbook this year and has been the Yanks’ best and most consistent starter. In his first six starts (prior to his little bout with elbow inflammation), he held opponents to a .619 OPS with a 2.08 ERA. In his six starts since, those numbers are .660 and 2.83. Slightly worse, sure, but still stellar.

Looking at some not necessarily more advanced stats, but ones that better represent underlying performance, helps back up Pettitte’s consistency. He got batters to swing and miss just 5% of the time in his first six starts, but that number has jumped up to 9% since. He stranded about 82% of baserunners in the first six, and about 80% in the second six. His GB/FB rate went from 1.02 to 1.47. Strikeouts? 6.23 K/9 before, 6.10 after. Walks? 3.00 B/9 before, 1.96 after. I could go on and on.

I’m comfortable saying that Pettitte’s best years are behind him, but his 2010 season has a little “last hurrah” to it. I find it to be very 2008 Mike Mussina-esque. You don’t expect him to keep performing this well, but the season is close to 40% complete, and he just keeps doing it. Season totals of a .256 BABIP and 82.1% LOB% tell us to expect a regression, but I get the sense that we might be waiting a while. Sometimes unexplainable things happen to extraordinary players, which Pettitte certainly is.

Regression and Maintenance: The story of Gardner and Cervelli

Regression to the mean. It’s a buzz term in the world of baseball statistical analysis. The premise is that baseball is such a complex game that we’re bound to see outlying performances all the time. Eventually, after a good number of plate appearances or innings, a player’s numbers will start to resemble his true talent level. Sometimes that will involve a complete correction, in which a player performs worse than his true talent and the numbers even out at the end of the year. Other times the player will just revert to his true talent and his season will look a degree better because of the hot streak. Either way, the player eventually regresses.

(And, just to eliminate any confusion, regression don’t necessarily mean downward. A player who starts cold and brings his numbers up also regresses. In 2009 Mark Teixeira regressed after his April performance.)

At the beginning of the season the Yankees were receiving production from two unexpected players: Francisco Cervelli and Brett Gardner. While some were greatly optimistic about Gardner heading into the season, I don’t think anyone could have predicted his insanely hot start. Absolutely no one predicted Cervelli’s torrential production early on. These two Yankees make for an interesting storyline, especially considering what they’ve done lately.

Cervelli: The story of regression

Photo credit: Rob Carr/AP

Early in the season opposing pitchers just couldn’t figure out how to make Francisco Cervelli hit his dinks and dunks to fielders. Everything seemed to find a patch of grass, or the dirt between fielders. On May 14, after a 2 for 4 day against Minnesota, Cervelli was hitting .415/.583/.528. No one in their right mind predicted he’d maintain that level of production. Yet many bought into the Cervelli hype. He did, after all, hit .298 last season. Plus, only the strongest of heart and mind can avoid being seduced by his wine bar eyes. But because he had only 101 plate appearances last year, we never got a sense of his true talent level.

Since that day in May — after which, coincidentally, he took over catching duties full-time while Jorge Posada hit the DL — Cervelli has fallen hard. In 93 PA, which covers more PA than his early season hot streak, he is hitting .200/.315/.227. That, too, is not necessarily indicative of his true talent level, but it certainly is a correction of sorts. It has brought his season numbers down to a much more believable .282/.374/.344. If he hit that in 350 PA as the sometimes-starter, I think everyone would be satisfied.

The good news: he actually increased his walk rate during this span. It’s why his .200 BA hasn’t been absolutely awful during this span. Instead it’s been tolerably awful.

Gardner: Maintenance, my man

Photo credit: Frank Franklin II/AP

At the start of the season, if you’ll remember back that far, Joe Girardi actually platooned Brett Gardner with Marcus Thames in left field. That didn’t last too long, not only because of Thames’s atrocious defense, but also because of Gardner’s stellar bat. In his first 108 PA, ending with the Boston series in early May, Garnder was hitting .344/.425/.419. Again, this isn’t something anyone could have expected him to maintain, but even if he settled into his true talent he would have quite an impressive season.

What we learned since, I think, is that no one has a good idea of Gardner’s true talent level. In his next 95 PA he dipped a bit, hitting .250/.326/.333, but then he got hot again. A thumb injury has taken away some playing time, but it hasn’t affected his performance. In June he is 10 for 22 with five walks, two doubles, a triple, and a homer. That has brought his season line up to .317/.400/.422, so while that’s a bit lower than his peak, it hardly represents a severe regression.

Where will Gardner end? No one, absolutely no one, can tell you with any degree of certainty. Like most players he has hot and cold spells, so it’s tough to tell what’s real and what’s a hot streak. It does appear, however, that Gardner can be more than the fourth outfielder which many people pegged him as. If that’s the case, he might have saved the Yanks the millions they would have otherwise spent on Carl Crawford.

A tale of two Phils

In the span of two batters on Sunday afternoon, Phil Hughes’ start against the punchless Astros went from great to mediocre. Tommy Manzella singled just past the reach of Derek Jeter to drive in two runs; former Yankee Kevin Cash homered for just the second time this year to plate another pair; and instead of a six-inning, one-run performance, Hughes found himself tagged for five earned runs in 5.2 innings against the NL’s worst offense.

For Hughes, the last two batters served as something of a microcosm for his recent performances. He had Manzella at 2-2 but then had to throw five more pitches — four fouls and a called ball — before the weak-hitting Astros’ short stop came through. During the AB, the pitches weren’t bad, but Hughes couldn’t locate his out-pitch. Manzella battled against the fastball and fouled off a curve. The home run by Cash, on the other hand, was off an 88-mph cutter that had too much of the plate.

With the book closed on yesterday’s outing, Hughes stands at an impressive 9-1 with a 3.11 ERA, a 3.69 xFIP, a K/9 of 8.84 and a BB/9 of 2.63. Every single one of those numbers is better than Hughes’ career averages, and as a 23-year-old, he’s showing us why he’s constantly been regarded as one of the game’s top young pitchers. But complacency can come at a price, and Hughes needs to be a few adjustments. Let’s break it down.

Over his first six starts of the season, Hughes went 5-0 with a 1.38 ERA. With a one-hit, 7.1-inning performance weighing heavily on the numbers, opponents hit .165/.243/.203 vs. Hughes as the righty struck out 39 hitters in 39 innings and walked 14. Tellingly, he gave up just one long ball. Over his next six starts, Hughes went 4-1 but with a 4.95 ERA. His strike out and walk rates — 35 K and 8 BB in 36.1 innings — are where they should be, but he has allowed four home runs over this span. Opponents have hit .277/.312/.419 against Hughes.

So what’s going on here? First, we have to address the variances in Hughes’ BABIP. As the second chart on this page shows, Hughes has seen his individual game BABIP trend constantly upward. At the same time, opponents have been hitting more fly balls and fewer ground balls off Hughes than they were at the start of the season. As such, a few more of those balls have left the park, and the ERA has ticked upward.

But BABIP and its cousin, while convenient scapegoats, aren’t the only culprits. Hughes has been getting hit because teams have been picking up his pitch selections. He’s had trouble throwing the curveball for strikes — only seven of the 18 he threw Sunday were strikes and all were called. He threw 22 cutters and no change-ups, by now a routine performance for Phil. By the time Cash teed off on a cutter, the Astros had gotten a good, long look at Hughes’ weaker pitch.

As the season has gone, Hughes has been a revelation. His peripherals are fantastic; his fastball electric; his curveball devastating. He’s pitching himself toward a spot on the All Star Game and justifying the Yanks’ faith in him over the last few years. But he’s still a work in progress. He still needs a change-up he can use as an out-pitch when hitters aren’t missing the fastball. He still needs to improve his pitch selection.

If the Yanks keep the rotation as is, he’ll face the Mets — a team against which he struggled in May — this weekend, and that could be a test for Hughes, making his last starts before he turns 24. But at that age and under the Bronx glare, every start is a test.

Jesus! Brackman & Heathcott!

The Yanks have signed St. Louis University righty Bryant Cotton as an undrafted free agent. It’s just a lower level depth move, guy to eat up innings. Meanwhile, Robert Pimpsner posted an updated projection of Short Season Staten Island’s roster, with the official one set to be released Friday.

Triple-A Scranton (8-6 win over Indianapolis)
Reid Gorecki, RF & Juan Miranda, 1B: both 1 for 5, 1 R – Gorecki tripled & K’ed … Miranda doubled & K’ed thrice
Colin Curtis, LF: 4 for 5, 2 R, 1 2B, 2 RBI, 1 K – had four hits in his previous 31 at-bats (.129)
Eduardo Nunez, SS: 3 for 5, 1 R, 2 2B, 1 3B, 2 RBI, 1 SB – that’s a helluva game right there … he got taken out hard on a clean slide at second while turning a double play, but he walked it off
Jorge Vazquez, 3B: 2 for 4, 2 R, 1 HR, 3 RBI, 1 K – not surprised he went deep, he hit some bombs in batting practice
Jesus Montero, C: 3 for 4, 1 2B, 1 RBI, 1 K – eight for his last 15 (.533) with four doubles … appeared to hurt his knee blocking a ball in the dirt, but he stayed in the game & even made a nice defensive play
David Winfree, DH: 0 for 4, 3 K – couldn’t join the party
Reegie Corona, 2BL 2 for 4, 2B – also made some kind of great defensive play, which has become his niche
Greg Golson, CF: 1 for 4, 1 R, 1 2B, 2 K
Romulo Sanchez: 5.2 IP, 4 H, 3 R, 3 ER, 1 BB, 7 K, 4-6 GB/FB – 58 of 91 pitches were strikes (63.7%)
Royce Ring: 0.1 IP, 1 H, 1 R, 1 ER, 0 BB, 0 K, 0-1 GB/FB – six of his ten pitches were strikes … gave up a homer to a lefty batter
Zack Segovia: 1.1 IP, 2 H, 2 R, 2 ER, 0 BB, 0 K, 1-3 GB/FB – 17 of his 24 pitches were strikes (70.8%)
Mark Melancon: 1.2 IP, 1 H, 0 R, 0 ER, 0 BB, 1 K, 1 WP, 3-2 GB/FB – 14 of 24 pitches were strikes (58.3%) … getting better, but still not back to his usual self

[Read more…]

Open Thread: Ready for a rematch

Sad Phillies are sad. (Photo Credit: Eric Gay, AP)

The Phillies return to the Bronx tomorrow for the first time since getting beat in the 2009 World Series by our beloved Bombers, but this current version of the Phightin’s is in quite the rut. They’re just 6-14 in their last 20 games, and have been outscored 98-48. As a team, they’re hitting just .221-.291-.309 during that stretch, and they were shut out six times. Suddenly all that talk about the Phils’ having an American League offense seems rather comical.

Anyway, here’s tonight’s open thread. The Mariners and Cardinals (Luke French vs. Adam Wainwright) are on ESPN, and I’m sure there’s some World Cup stuff going on somewhere. Whatever you choose, talk about it here.

Pinstripe Alley interviews Austin Romine

Travis G. at Pinstripe Alley recently sat down with Yankees’ prospect Austin Romine (part one, part two) for a quick and informative interview. They discussed a whole bunch of stuff, including the generic questions about what a typical day is like, what he does in the offseason, stuff like that, but I most enjoyed the questions about how he’s changed his approached his plate this year, and what he has to work on defensively. It’s a great a relatively quick lead despite being a two-parter, and I highly recommend it.