Happy Labor Day! Regularly Scheduled Articles Will Resume on Tuesday, September 2.
May 11, 2006
Doctoring The Numbers
The Draft, Part Ten
We complete our tour of the diamond with a stop at the mound, where we'll compare left-handers to right-handers and starters to relievers, and see what the data shows.
Pos Years 1st Rd 2nd Rd 3rd Rd Overall Busts COL LHP 84-91 - 4.4% + 54.7% +133.4% + 21.5% COL LHP 92-99 - 7.3% + 61.1% + 15.0% + 8.0% COL LHP 84-99 - 5.8% + 57.8% + 82.4% + 15.2% Years Biggest Bargains Biggest Busts 84-91 Jim Abbott, Greg Swindell Drew Hall, Kyle Abbott 92-99 Barry Zito, Randy Wolf B.J. Wallace, Jeff GrangerNote that the two most valuable draft picks from 1984 to 1991 are not Randy Johnson, who was third on the list. Johnson is a future Hall of Famer, but was not a full-time starting pitcher in the major leagues until four years after he was drafted, and didn't become RANDY JOHNSON until 1993. And of course, along the way he was traded by the team that drafted him, the Montreal Expos, essentially for four months of Mark Langston. The point bears repeating: the sooner a draft pick renders his value, the less likely the team that drafted him will have already given him up for pennies on the dollar.
By way of comparison, Jim Abbott had his last good year at age 27, but returned instantaneous value to the Angels. Abbott signed too late to pitch in his draft year of 1988, then made the Angels' rotation the following spring without any minor league experience and won 12 games, the most major league wins of any pitcher in his first pro season in nearly a century.
In consecutive years, the A's made the best draft selection on a college left-hander (Barry Zito in 1999) and the third-best selection (Mark Mulder in 1998). Drafting wisely is still the best way to arbitrage talent.
Overall, college left-handers have yielded good value in both eras, although the value seems to be far more concentrated in the second and third rounds. Whether this is signal or noise is debatable. A quick perusal of college southpaws taken in the first round reveal quite a few guys who were selected because they could throw hard but didn't particularly know how to pitch (Wallace, Granger, Hall). Ryan Mills blew out his arm almost from the moment he signed. Some of the second- and third-rounders who panned out were guys who didn't throw all that hard but changed speeds and pitched to spots, like Denny Neagle, Jarrod Washburn, and Mike Maroth. On the other hand, Johnson threw harder than anyone, and he wasn't picked until the #36 overall selection.
Pos Years 1st Rd 2nd Rd 3rd Rd Overall COL RHP 84-91 + 27.4% + 52.3% + 21.2% + 32.0% COL RHP 92-99 - 23.9% - 1.7% - 55.2% - 22.5% COL RHP 84-99 + 0.9% + 26.1% - 13.4% + 4.5% Years Biggest Bargains Biggest Busts 84-91 Mike Mussina, Jack McDowell Bill Bene, Pat Pacillo 92-99 Jeff Weaver, Jon Lieber Steve Soderstrom, Pete JanickiThat's the long-forgotten third wheel in the Alternatively Spelled William Beans, the one who is neither in the A's front office nor out of the closet. David Schoenfield ranked Bene #53 on his list of the worst draft picks in sports history.
Bene ranks among the two or three wildest minor league pitchers of the last 25 years. In 1989, his first full pro season, Bene had 18 walks in 27 innings. No, wait…he had 18 wild pitches in 27 innings; he walked 56 batters. In 445 career minor league innings, Bene walked 489 batters, hit 31, and threw 140 to the backstop. His main competition for wildest first-round draft pick ever would probably be Jacob Shumate, the Braves' first-round pick out of high school in 1994, who walked 265 batters in 211 innings before retiring. (Jason Neighborgall, selected by the Diamondbacks last year, could one day surpass both of them--he walked 113 batters in 101 college innings, then got his pro career off to a rousing start with 45 walks and 23 wild pitches in 23 innings. But at least Neighborgall was a third-rounder.)
There isn't much of a trend here. The four bargains listed above were successful college pitchers who were known more for their polish than their pure stuff, but you can easily come up with just as many exceptions to that rule. In 1995, the first college right-hander taken was Jonathan Johnson, a highly polished pitcher from Florida State, who barely sniffed the majors after Texas took him with the #7 pick. The Cardinals took advantage of the Rangers' gaffe by taking a comparatively less ready starter with a better arm out of Seton Hall with the #12 pick, got him to start throwing strikes, and ended up with Matt Morris. And while he's outside the scope of our study, Justin Verlander was a so-so college pitcher at Old Dominion despite electric stuff, and the Tigers wisely calculated that they could translate his stuff into results as a pro.
Pos Years 1st Rd 2nd Rd 3rd Rd Overall HS LHP 84-91 - 36.6% +135.3% + 90.5% + 14.7% HS LHP 92-99 - 48.6% - 72.0% - 3.0% - 45.2% HS LHP 84-99 - 40.4% + 55.0% + 21.3% - 8.4% Years Biggest Bargains Biggest Busts 84-91 Tom Glavine, Justin Thompson Brien Taylor, Chris Myers 92-99 C.C. Sabathia, Jeremy Affeldt Doug Million, Geoff GoetzI wouldn't have guessed that Justin Thompson would rank as highly as he did, but he reached the majors quickly and had a pair of very good seasons at the height of the offensive explosion--his 3.02 ERA in 1997 was worth nearly 9 WARP. Steve Avery ranks ahead of him in terms of overall value, but was the #3 overall pick compared to Thompson's #32, so more was expected.
High school left-handers from the 1980s are almost shockingly successful, coming in the era that gave high school pitchers a bad name. As you can see, the value of high school southpaws plunged dramatically in the 1990s, for reasons that are unclear.
And yes, that's Jeremy Affeldt on the list of bargains. If you think listing Affeldt second is bad, consider that fourth on the list of bargains from that era is Jimmy Gobble. And Chris George is fifth. Swear to God. (Rick Ankiel is third.)
Doug Million is excused from approbation; for those of you not familiar with him, he died from an asthmatic attack during Instructional League in 1997, a story which is as tragic and unfathomable now as it was then.
Pos Years 1st Rd 2nd Rd 3rd Rd Overall HS RHP 84-91 - 64.0% - 40.1% - 58.4% - 55.6% HS RHP 92-99 - 7.1% + 1.3% - 17.1% - 6.4% HS RHP 84-99 - 34.9% - 21.4% - 32.7% - 30.7% Years Biggest Bargains Biggest Busts 84-91 Greg Maddux, Roger Pavlik Kurt Miller, Roger Salkeld 92-99 Jeff Suppan, Roy Halladay Kirk Presley, Matt DrewsYes, Roger Pavlik. Of 86 high school pitchers taken between 1984 and 1991, a total of nine of them rendered positive value. Here's all you need to know: of those 86 pitchers, Todd Van Poppel ranked 11th. That means that nearly 90% of high school right-handers taken from 1984 to 1991 were worse draft picks than Todd. Van. Poppel.
It is not a stretch to say that over an 8-year span, Greg Maddux was the only high school right-hander taken in the first three rounds that fulfilled the expectations placed on him on Draft Day.
As we've documented before, high school pitchers were much better values after 1991. As with college right-handers, there does not seem to be a trend towards or away from "stuff" pitchers. Jeff Suppan was about as polished a high school right-hander as you'll ever find; he was in the majors at age 20, and while it took a few years for him to stick, by age 24 he was a decent #3/#4 starter--and he really hasn't improved a whit since then. Roy Halladay, on the other hand, was a scouts' favorite despite terrible strikeout-to-walk ratios in the minors, crashed and burned in 2000, then had his motion completely rebuilt in one of the most impressive reconstructions of a pitcher ever. Among the 10 best picks from this era are Kerry Wood and Matt Clement, as well as Jon Garland and Brian Meadows.
A look at the failures isn't any more instructive; some of the guys weren't particularly good to begin with, and most of them got hurt along the way.
So what can we take from this?
1984 - 1999 1992 - 1999 Pos Overall Pos Overall COL LHP + 15.2% COL LHP + 8.0% COL RHP + 4.5% HS RHP - 6.4% HS LHP - 8.4% COL RHP - 22.5% HS RHP - 30.7% HS LHP - 45.2%From 1984 to 1999 as a whole, there is a small but clear trend favoring left-handed pitchers at both the college and high school ranks. That trend disappears into complete randomness if we just focus on pitchers since 1992, albeit in a smaller sample size. If we break things down by round, the data gets a little more interesting--and a little more confusing.
First Round 1984 - 1999 1992 - 1999 Pos Overall Pos Overall COL RHP + 0.9% HS RHP - 7.1% COL LHP - 5.8% COL LHP - 7.3% HS RHP - 34.9% COL RHP - 23.9% HS LHP - 40.4% HS LHP - 48.6%In the first round, at least, it looks like right-handers have a small advantage for the data as a whole, although in the 1990s the edge only holds at the high school ranks.
Second and Third Round 1984 - 1999 1992 - 1999 Pos Overall Pos Overall COL LHP + 66.5% COL LHP + 45.9% HS LHP + 45.1% HS RHP - 5.5% COL RHP + 12.7% COL RHP - 19.0% HS RHP - 24.8% HS LHP - 41.5%After the first 30 picks, left-handers have an enormous advantage over the course of the entire study, but again, the trend doesn't hold in the 1990s, as high school right-handers gain the upper hand.
I'm leery of coming to any conclusions based on data that is this inconsistent, but I'll try to make inappropriate inferences anyway. At the college level, left-handed pitchers have a persistent edge over right-handers, an edge which primarily manifests itself after the first round. It could be that once teams have gone through the elite collegiate pitchers, they start looking for pitchers who "know how to pitch" despite the lack of top-shelf stuff. As it is far easier for a left-hander to succeed in the majors with average velocity, a left-hander who has a strong collegiate track record and makes the most of his 87-mph fastball is far more likely to maintain that success in the pros than his right-handed counterpart. Washburn was a 2nd-round pick in 1995. Maroth went in the third round in 1998, as did Scott Schoeneweis two years earlier.
At the high school level, I'm reluctant to make any generalizations, for the simple reason that the data swings so violently between eras: between 1984 and 1991, left-handers returned more than twice as much value as right-handers, but between 1992 and 1999 right-handers were about 80% more valuable. The only static data point is that right-handers have a persistent advantage in the first round. I suspect this is because at the high school level, right-handed pitchers tend to have a lot of projection left in them, whereas left-handers are frequently drafted as relatively finished products. Even while they count as "successes" in our system, George and Gobble are good examples of high school lefties who looked like they were ready for the majors before their 21st birthday, and haven't really progressed at all since then.
This may be a case of grasping for conclusions that aren't in the numbers. If Ryan Anderson doesn't get hurt, that first-round advantage for right-handers may completely disappear. The data may be trying to tell us something, but it's not entirely clear what.
Let's see if the data can be any more articulate on the subject of relievers. For this portion of the study, we only looked at college pitchers (high school relievers don't exactly make for high draft picks), and only since 1992. While college relievers were occasionally drafted high before that (notably Gregg Olson, who was taken #4 overall by the Orioles in 1988 and was a huge part of their "Why Not?" season the following year), it was a rarity.
Kevin Goldstein graciously did the legwork for me in coming up with the names of every Top 100 draft pick who was used in relief during his final college season. He gave me a list of 37 pitchers from 1992 to 1999. He might well have missed a few players; it's surprisingly difficult to check the day-by-day game logs for Southwest Texas State's 1997 season. And not every pitcher was drafted with the intention of being used in relief--Darren Dreifort, for instance, was used as a sort of uber-reliever his junior year at Wichita State, but was certainly considered a future major league starter when the Dodgers took him #2 overall in 1993. But it's close enough.
Pos Years 1st Rd 2nd Rd 3rd Rd Overall COL R 92-99 - 32.4% + 35.4% - 7.1% - 18.7% COL S 92-99 - 0.4% + 28.3% - 12.3% + 4.5% Years Biggest Bargains Biggest Busts 92-99 Scott Sullivan, Mike Maroth Rick Greene, Al ShepardAl Shepard is one of the most mystifying first-round picks of recent times; the Orioles took him even though he had a 6.57 ERA his junior season at Nebraska. Kevin is correct when he writes that great college numbers sometimes mean nothing when it comes to a player's major league chances, but terrible college numbers in a player with great stuff is usually a red flag that something's wrong. Speaking of things going wrong, Matt Anderson ranks third on the list of disappointments.
The data here isn't compelling one way or the other; as a whole, relievers returned value below college starters, but that's likely because 31 of the 37 relievers were right-handed. Relievers actually returned more value than college right-handers in the same era. But relievers in the first round did particularly poorly, a result which can be traced to this run of picks from 1992 to 1997:
In total, five college relievers were taken in the top four picks in a six-year span, culminating with Matt Anderson at #1. Looper has returned about 13% more value than expected, Dreifort and Shuey didn't reach expectations but weren't terrible, whereas Gomes and Anderson were unreserved busts. But none of these relievers ever came close to achieving the kind of dominance you would expect from a Top Five pick. It's worth noting that no reliever has been taken in the top five since.
The trend of teams drafting college relievers early in the draft, and the trend of college teams putting their best arms in the bullpen, has accelerated over the last decade, so these results may not be relevant. Both Chad Cordero (#20, 2003) and Huston Street (#40, 2005) figure to join Olson on the list of college relievers who went on to have sterling major league careers. David Bush (#55, 2002) has also proven to be a wise pick, although Bush is interesting in that most teams--not just the Blue Jays--thought of him as a starter in the pros even though he came out of the bullpen in college.
Then again, Ryan Wagner (#14, 2003) and Royce Ring (#18, 2002) have been disappointments, and Bill Bray (#13, 2004) has yet to reach the majors. Craig Hansen (#26, 2005) might live up to the hype, but it's early. Joey Devine (#27, 2005) has so far demonstrated only an impressive knack for giving up grand slams and season-ending home runs.
As much as we'd like to have some hard-and-fast rules about drafting pitchers, there don't seem to be any. The best we can do is make a few suggestions. At the college level, left-handed pitchers tend to be a slightly better value than right-handers, at least in the second and third rounds. At the high school level, right-handers tend to be a slightly better value than left-handers, at least in the first round. And you better have a damn good reason to draft a reliever in the top 5 picks. That's about it.
Next time, we'll start tying up some loose ends by summarizing the "rules" of the draft, as well as looking at whether recent changes in draft strategies have altered the draft equation.