May 26, 2006
Cubs, Devil Rays
Coming into the season, the game plan for the Cubs was simple. On the mound, let a few youngsters and innings-eaters hold down the back of the rotation until the injured aces return with a flourish. Rely on a solidified bullpen to hold on to a few more leads. At the plate, expect more run production out of a lineup with speedster Juan Pierre at the top and enjoy the results of Derrek Lee and Aramis Ramirez getting more at-bats with ducks on the pond.
With the northsiders at 18-28, it's no surprise that the plan has gone awry--only the bullpen has lived up to expectations thus far. Bob Howry and Scott Eyre, Jim Hendry's pricey pair of pickups, are each among the NL's top 30 relievers by WXRL. Ryan Dempster and Scott Williamson have been reliable as well, holding opponents to batting averages under .200. The offense, of course, has been gutted by Lee's absence and hamstrung by Pierre's .269 OBP.
The first five innings (or fewer) have been baffling, though not always in a negative way. Going 5-0 in April, Greg Maddux not only stood out in contrast to the rest of the Cubs rotation, but also against his own track record:
Year April ERA Season ERA 2002 4.37 2.62 2003 5.13 3.96 2004 5.65 4.02 2005 4.20 4.24 2006 1.35 3.89 (weighted mean PECOTA projection)His April performance this year wasn't supported by his peripherals, and the law of averages has attacked with a vengeance: Maddux has given up five or more runs in four of his five May starts thus far.
The gaggle of arms behind Maddux and Zambrano has been equally unpredictable. Glendon Rusch hadn't been truly dreadful for a couple of years, but he pitched his way out of the rotation by May 1. Jerome Williams, one of Hendry's more intelligent '05 acquisitions earned himself a plane ticket back to Des Moines. Rich Hill lasted four starts in the bigs, losing each one and tarnishing his prospect status a little more with each outing. In fact, the Cubs have managed only 18 quality starts in 47 games. (By contrast, the Cardinals have 27 and the Reds 24.) Twelve of those are equally split between Maddux and Carlos Zambrano. The only other Chicago hurler with more than one is 23-year-old Sean Marshall, who has four.
Marshall made the starting five in Spring Training despite never pitching an inning above Double-A, and he'll probably join the Iowa rotation if ever Wood, Mark Prior, and Wade Miller are all healthy at the same time. In the meantime, the Cubs have won six of his nine starts. His performance thus far has been enigmatic: from the 47 2/3 innings we have to go on, the numbers could make a case for more Major League success, or for a rapid demotion.
Marshall misses a lot of bats, but not always in the way he'd prefer. He's among the top third of qualified NL pitchers in strikeouts per nine innings, but well into the bottom quartile in K/BB. His walk, strikeout, and hit rates closely resemble those of another promising NL starter, Jeff Francis, but while Francis has given up an equal balance of grounders and fly balls, Marshall's G/F ratio is 1.55, again in the top third of qualified starters.
Adding to the enigma, most of Marshall's starts have come in pitcher-friendly parks or against weak offensive teams. A 5.29 ERA put together in favorable circumstances isn't the sort of thing Cub fans would like from their #3 starter, but there's room for hope. Best of all, he's healthy, and as long as he continues averaging 85 pitches per start, he just might--knock on Dusty's toothpick--stay that way.
If you've ever asked a highly refined, somewhat snobbish but respected friend to rank his/her favorite books, meals, or TV shows, you've no doubt had the following experience. One of the list items was probably something of a guilty pleasure, totally unexpected from such an esteemed source. He likes Six Feet Under, Arrested Development, Twin Peaks...and World's Scariest Police Chases. She likes anything by Camus, Dostoevsky, Nabokov...and Mary Higgins Clark. Yes, prosciutto pizza with arugula, shaved asiago and a drizzle of olive oil is divine...but there's nothing quite like stretching out on the couch with a bowl of Product 19 with your bare feet on the coffee table. To watch Cops.
This is somewhat equivalent to Bronson Arroyo's place on most pitching leaderboards. The ranking is totally unexpected, it confuses you, and there's a part of you that thinks it's a phase to be outgrown (Cops just can't be a favorite show, can it?). But also somewhere on the list is the obscure William Gaddis novel, the up-and-comer that no one else really knows about, except you felt that it belonged on any list of great books ever since you finished it, and its presence on your friend's list is sort of validating.
This is Tampa Bay's Scott Kazmir, who has been one of the best starters in the majors this year. Ordered by expected wins, he's been second-best, behind Arroyo. Kazmir's youth, rawness, and mechanics all conspired to have the Mets send him to the Devil Rays in 2004 (for two players currently on their 60-day DL--salt, meet wound).
Last year he walked 4.8 men per nine innings, which is, to put it mildly, worrisome (only Kip Wells ranked higher among full-time starters). His walk trouble has been so pronounced in his young career that the most optimistic projection for 2006--his 90th percentile PECOTA projection--still featured a well-above average walk rate (3.8/9 IP). So that Kazmir's current walk rate sits at just 2.95/9 IP is both inspiring and suspicious.
How suspicious? To start, let's turn to his PECOTA card. One of the many cool features of each PECOTA card is the chart at the top detailing that player's profile. The profile is based on actual past performance, not projection. In this way, it's descriptive, not necessarily predictive. We'll reproduce Kazmir's chart here to cut down on excessive back-and-forth clicking:
Since everything's expressed as a percentile, you can read the chart as follows: Kazmir has been in the top 90% in terms of strikeouts per batter faced, almost the top 80% in ISO against, etc. But it's the walk column that looks obviously out of place next to the rest. Kazmir has been serviceable-to-dominant in every category here save control, where he's been downright abysmal, as approximately 95% of all pitchers have exhibited better control per batter faced than Kazmir has.
Here's where we turn to history. The criteria we're using are pretty restrictive, and it's almost-but-not-quite-like asking "how many players in history have duplicated the precise circumstances of Scott Kazmir's career?" Essentially, we're looking for pitchers who had a walk problem in their first full season (defined as "more than 4.5 per nine IP"), were better than average in their second (defined as "fewer than 3 BB/9"), and who were starters both years (defined as pitching 175+ innings AND making 30 starts--and you were warned about the restrictive criteria).
As it turns out, this is a pretty easy list to talk about. Since 1990:
Nobody. It just hasn't been done.
How about since 1980:
Going back further than this introduces all sorts of additional variables, like different offensive environments, different pitcher workloads, etc. So while we can't say "it's never been done," we can say that what Kazmir's currently doing, should it continue, is unheard of at least in the last 26 years.
There is also something of a selection bias in place here that we should identify, as a player with a walk rate approaching five in his first full season either:
But one name from the "almost satisfied the criteria" list jumps right out at you for some obvious reasons:
<--------> <--------> Name Year1 BB/9 Year2 BB/9 --------------------------------------- C.C. Sabathia 2001 4.74 2002 3.77It's not quite the same control improvement, but C.C. Sabathia's rookie year was similar to Kazmir's in many respects. Both were among the youngest players in the league, both had some observers cringing at their mechanics, and both walked in the neighborhood of five batters per nine innings. For Sabathia, he didn't become an overnight control freak, but he did cut his walks enough so that his career walk rate looks pretty average, even when you include his rookie year.
So pitchers with significantly above-average walk rates in their rookie year haven't improved to an average-to-below-average rate in just one year; it either takes more than one season to work through the command issues (think Randy Johnson) or else they give back whatever gains they eventually make (think Matt Young), which is by far the more common option of the two.
One thing we may be seeing here is a Big Leap Forward that only looks that way because of an aberrational rookie season. In Kazmir's admittedly brief minor-league career, his walk rate was a far more reasonable 3.6 BB/9. His 2005 walk rate probably clouded perception of him since it was really the only perception we had. It was also Kazmir's only sustained appearance at any level. Previously, his high in innings pitched at a single level was just 76, which came in A ball in 2003.
If Kazmir were to continue his ~3 BB/9 rate, it would be technically without precedent in a quarter-century, but the nuances of his specific case would make it less groundbreaking. Regardless, the improvement has propelled him into the top tier of AL starting pitchers, giving the Rays' the first legitimate young ace they've ever had.