April 3, 2009
Wait 'Til Next Year
[This research article appears in lieu of BP's regular college coverage. For an updated NCAA Top 25, please see yesterday's post to Unfiltered.]
In my first article as a regular at BP almost two years ago, I wrote about the discrepancy that the best pitchers in college baseball had between their numbers against the first five hitters in a batting order and the hitters in the sixth through ninth spots. My hypothesis was that since most college programs would never send any players from the bottom of their order to professional baseball, numbers against the bottom half of the order were essentially meaningless. In breaking down the eight best draft-eligible pitchers that season, the separation between how well pitchers dealt with the two halves of the lineup essentially mirrored that of major league pitchers, though with two exceptions: Andrew Brackman, who had been beaten down mercilessly by the middle of ACC lineups and was making up his (mediocre) numbers against the bottom half of batting orders, and Jake Arrieta, who oddly was better against the first five hitters in a lineup than the bottom four. I didn't know what to make of this then, and I don't know what to make of it now.
I think there is something to that type of analysis, however, and I think we can improve upon it. Last year's draft crop had its fair share of seniors and redshirt juniors that started in college baseball in 2005-in fact, with a few exceptions (fifth-year seniors), everyone playing college baseball in 2005 is no longer doing so. This is significant, because in my searches through school archives, 2005 is the earliest season that schools began to publish play-by-play logs (amazingly, some still don't). When I noticed this, I realized that the combination of '05 PBP logs and the '08 draft allowed for something rather significant: we can see how 2005 draftees did against players that went on to be drafted versus players that ended their careers upon graduation. Still, the scarcity of schools that have maintained the requisite information for those four years negated any chance of finding out if there was something we might not have realized about Matt Garza, or Ricky Romero, or Cesar Carillo.
Until we have the data necessary for every first-rounder (and every 50th-rounder, for that matter) to do the analysis, it's impossible to draw any conclusions about the importance of facing future draftees. As an introductory exercise, I want to analyze the case of Lance Broadway, the Chicago White Sox' 2005 first-round pick. Broadway, if you recall, used a late-season surge to propel himself into the top 30 picks. In all, Broadway earned Player of the Year consideration for his marvelous season: 15-1, 1.62 ERA, and 151 Ks and 35 walks in 117 IP. Momentum was Broadway's best friend in 2005, as he went from second- or third-rounder in February to the 17th overall pick in June thanks to his last six starts: 49 IP, 29 H, 61 Ks, 11 walks, and, oh yeah... just three earned runs (or a 0.55 ERA!).
We already have some sabermetric tools at our disposal, thanks to the invaluable Boyd Nation. Nation's analysis told us that Texas Christian's average park factor (of all games in 2005) was 95, so Broadway was pitching in a slightly favorable environment. We know he had four starts with more than 120 pitches thrown. We know that his defense-independent ERA was 2.14. We know, according to Boyd Nation's ISR, that Broadway's average start was against the 93.7th-best team in the nation, and he made only five starts against top 50 teams. However, none of these numbers really do much to solve the biggest issue regarding college statistics-talent disparity. While dozens of major league players have a history of dominant college numbers, so do an equal amount of players that never rose far above A-ball.
The separation between players with pro potential and players who max out in Division I creates a disparity that numbers can't catch. For instance, there were probably area scouts that filed a report on April 15, when Broadway pitched seven solid innings against the University of Memphis. The problem is, the performance is suspect; not a single one of the 31 batters that stepped to the plate that day was ever drafted. By contrast, in his last start of the season-during which the White Sox and others probably had scouting directors and cross-checkers present-Broadway was brilliant against Stanford on the road in the Super Regional, allowing one earned run while striking out seven in nine frames. Add to that line the fact that 29 of the 32 Cardinal batters that Broadway faced that day went on to be drafted, and the feat looks much more impressive.
Overall, if I tell you that 37.5 percent of the 480 batters that Broadway faced during his junior season were eventually drafted, we can create a split (using a combination of play-by-play logs and BaseballReference.com's draft database) that should identify this disparity. Please note, this analysis will fail to identify three rare instances of drafted players:
The last group excludes Joe Savery, Brad Lincoln, and Micah Owings, who I identified in my analysis. For the record, this athletic crop went a cumulative 2-for-11 against Broadway in 2005. With these caveats out of the way, this is Broadway's split against the drafted and undrafted college players that he faced in his final college season:
I didn't expect this at all. In the sample-sized case of Broadway, the first-rounder was actually better against his toughest set of opponents. Players whose baseball careers ended after college hit for a better average, with slightly more power and fewer strikeouts against Broadway. There were occasional exceptions to this trend, like Broadway's April 29 start against the University of Houston. In this complete-game, 12-strikeout victory, Broadway faced 22 batters that would never play minor league baseball and retired 20, with the lone survivors drawing a walk and getting hit by a pitch. In the 17 plate appearances against drafted players, Broadway gave up seven hits and two walks. However, there are plenty of anecdotal examples that correlate with the findings above: Broadway geared up when it counted.
How about the very best? Here's a look at the players that Broadway faced during his junior season that went on to become draft picks taken in the first 10 rounds:
February 18 vs. Florida Atlantic: Michael McBryde (fifth round, 2006)
Say what you will about the minor league performances of the players on this list-Owings and Lowrie notwithstanding-these were the best talents that Broadway pitched against in front of scouts. In all, this represented 49 plate appearances, or roughly 10 percent of his season. This best crop hit .159/.245/.205 against Broadway, striking out in nearly a third (16) of their plate appearances. It's a striking suggestion that Broadway knew when to shine, effectively turning on the gas when scouts were paying the most attention.
I close today without any grand conclusions, or even any blind guess at a hypothesis, but it's a practice that I think serves a purpose, and yet another statistic that teams should be thinking about as they try to weed out the busts from the steals. Knowing all of these numbers, I have no doubt that the White Sox still would have drafted Lance Broadway almost four years ago. While the numbers never lie, there are times that they don't make sense.