July 24, 2007
Wait 'Til Next Year
It's Raining (Leadoff) Men
Once a player turns in a 6.6 60 time, or goes from home-to-first in less than four seconds, he gets branded a future leadoff hitter. Speed is the prerequisite, and--sabermetrics be damned--it is the first trait scouting directors and managers look for in a leadoff hitter. If this can't change, surely we can pair speed with other traits to improve the percentage of successful leadoff draftees.
Last week, I began this study by grouping 33 players drafted with leadoff intentions into three categories: the good (think Jose Reyes), the bad (think Doug Glanville) and the ugly (think Drew Meyer). Sticking with the players' biographical info, I noticed that the ability for a teenager to add weight is a worthwhile predictor of future success. I also found that numbers from good college programs can cloud perspective on a player's talents. I did continued research in this and searched the first five rounds of the 1996-2003 drafts, where 43 players were taken from big college programs. Of that group, only two are good starters (Brian Roberts and David DeJesus), three more have 1,000 career at-bats (Jason Tyner, Alex Cora, Willie Bloomquist), and 20 have not reached the Major Leagues. Certainly, an emphasis on tools over college numbers seems important for future leadoff hitters.
My next step is to see what happened between a player's being drafted and his earning his distinction. What can we look at to potentially warn us of a bust, or signal a sleeper? Every player studied spent time in the minor leagues, however, the hardest thing to do would be to compare the players' performances across leagues; we're talking about 33 different players playing in about 15 leagues over quite a few seasons. That comparison would be difficult if not for Clay Davenport, who sent me his basic translated numbers of each player in the study, "as if you were bringing them up to the majors that year."
I decided to not use short-season numbers in this study because I don't find the numbers particularly useful. For the most part, college players dominate, high school players struggle, and you won't find a good correlation between short- and full-season success. Given the difficulties translating these numbers, it's best to skip to a player's first taste of full-season ball. What's interesting is that for the bad group--leadoff hitters who have been relatively worthless in the majors--only five of 11 players started full-season ball in Low-A. Another five players skipped the level and moved to High-A, and Roger Cedeno went right to the Texas League. This is mildly noteworthy, because only three players from the good group and one player (Steve Stanley) from the ugly group skipped Low-A.
For the levels, I attempted to only use a player's first extended exposure to the league. If you allow the "ugly" group to have every season they played in Low-A, they have nearly twice the amount of at-bats as the "good" group. However, I thought it would be most prudent to only look at the first time Josh Womack played in the Midwest League, and conversely, to eliminate Derek Jeter's 39 at-bat Low-A trial in 1992. With that, here are the group's translated numbers:
Group AB AVG OBP SLG K% Good 3099 .246 .305 .361 21.6 Bad 2406 .227 .271 .285 17.5 Ugly 3987 .203 .256 .290 26.0
What jumped out first was the batting average column, which really speaks to some of the overall difference in OPS. Upon further inspection, I saw the translated BABIP for each group told part of the story--these were .300, .272, and .265, respectively. This likely means the good group had a higher line-drive percentage in the minor leagues; certainly, their dominance in Isolated Slugging would agree they make better contact. Unfortunately, Minor League Splits hasn't been around for 20 years to prove it, so we're going to have to stick with that informed guess. However, the site's existence should allow us to double-check this idea very soon.
Given how much BABIP and slugging separate the good group from the rest, I think it's informative to further pursue the power department. Here's a look at what percentage of total at-bats each group delivers an extra-base hit in:
Group 2B 3B HR XBH Good .047 .012 .015 .074 Bad .035 .008 .002 .045 Ugly .041 .009 .009 .059
By far, the most shocking number in that table is that the bad group only hit home runs in 0.2% of their at-bats. Upon further inspection, I found of the 23 player seasons in Low-A, a player had a rate of HR/AB at less than .005 nine times. Only one of these players (Grady Sizemore) is in the good category, four are in the bad category, and four in the ugly category. This leads to our next rule:
The Tyrell Godwin Rule to Identifying Leadoff Busts: The early onset of home run power is important; a player hitting one home run per 200 at-bats (or less) in Low-A is a likely bust.
To take a quick trip through the other numbers, I know some will be curious about stolen bases at the level. Overall, stolen base success rates hovered around 65 percent for all three groups, with just one player as low as 50 percent and one as high as 75 percent. Overall, like height last week, stolen base success rates were not predictive at all.
The strikeout rates might surprise some people; I'm not shocked. The good group is composed of guys with a little power that learned the leadoff role, the bad group is made of athletes that were always taught to put the ball in play and let their legs do the work. The ugly group has a lot of strikeout rates that intertwine with the good group; however, five of the ten players in the group with Low-A experience struck out in 25 percent of their at-bats. Swinging and missing that often is pretty damning for a leadoff hitter, but the real concern shouldn't come until they progress up the minor league ladder.
Group AB AVG OBP SLG K% Good 2604 .257 .320 .378 19.7 Bad 3628 .239 .286 .318 20.1 Ugly 3972 .229 .292 .336 23.6
If strikeouts aren't completely worrisome in Low-A, the next level should be the time when triple-digit strikeout numbers become a concern. The good group drops 11.6 percent and strikes out every five at-bats, with stars in Derek Jeter, Jose Reyes, and Grady Sizemore all improving their strikeout rate between Low- and High-A. The ugly group drops as well, but not substantially, as the average is brought down by three players posting a rate under 15 percent. While Low-A didn't have many players from the Bad group, nine players from the group played in High-A, so players like Chris Duffy help raise the strikeout average by fifteen percent. This can lead us to our next rule:
The Josh Burrus Rule to Identifying Leadoff Busts: Strikeout rates should be declining, not staying stagnant (or increasing) as a player goes up the ladder. Striking out in 25 percent of at-bats at any level past Low-A is a problem.
Finally, I'd like to use this opportunity to point out that Tom Goodwin (a member of the bad group) had a 14-season career line of .268/.332/.339. That is, the good group were better while they were in High-A than Goodwin was for most of his major league career.
For this category, I combined the first extended performance that every player in the study had in Double-A and Triple-A. Repeat or demoted assignments were not included.
Group AB AVG OBP SLG K% Good 7737 .256 .322 .387 18.3 Bad 7565 .253 .301 .336 17.1 Ugly 6416 .232 .297 .330 19.7
If we see nothing else, it's important for understanding the concept of replacement-level to note the tiny ten-point difference in OPS between the bad and ugly groups at these levels. In the levels closest to the Major Leagues, the bad and ugly groups are nearly identical, while arriving at a sub-700 OPS in different ways.
Most amazing to me is the consistency of the bad group in the minors. These are guys who, despite their High-A hiccup, either single or ground out very often. These anti-Three True Outcomes players had walk-to-AB ratios that barely improved in the three categories: 6.4 percent to 6.6 to 6.8. The players added twenty-five points to their Isolated Power and were stealing bases at a 73 percent clip in the higher levels, but overall, their minor league maturation is pretty slight.
The ugly group, which was so much worse than the bad group in Low-A, has caught up with them in some regard, particularly in two areas:
Level BB% K% Low-A 7.1 26.0 High-A 8.9 23.6 Upper 9.2 19.7
The walk number is particularly surprising, and this progression is why we didn't talk about walk rates at the first two levels. The good group has a 9.6 percent walk rate in the higher levels, an improvement from 9.3 percent in High-A and 8.5 percent in Low-A. While the good group had a significant advantage in this regard in Low-A, the ugly group catches up in the end. Both groups dwarf the bad group, which showed steady stagnant progress in their low figure while moving up the ladder. This leads to our final rule of the day:
The Doug Glanville Rule to Identifying Leadoff Busts: Raw speed should not overshadow poor strikeout rates; contact hitters with bad walk rates are just as valuable as slower, swing-and-miss leadoff hitters.