Thursday, November 01, 2007
Twins 2007 Minor League Numbers: Pitchers
Earlier this week I discussed the importance of putting minor-league performances in proper context and introduced a system that attempts to do that by accounting and adjusting for factors like leagues, ballparks, run-scoring environments, ages, and defensive positions. I then used that same system to examine the Twins' minor-league hitters, producing a context-adjusted hitting line for each player and determining which prospects had the best offensive seasons in 2007.
If you missed either of those two entries you'll want to go back and read them, because today I'll be using the same system to examine the Twins' minor-league pitchers. Evaluating pitchers is more difficult than evaluating hitters, because pitchers rely heavily on the defense behind them. Beyond that, pitchers often exit games with runners on base and whether or not those runners score off the pitcher who comes in from the bullpen can have a big impact on ERAs and win-loss records.
All of which is why, rather than looking at more traditional stats, I'm instead focusing on a metric called FIP, which is short for Fielding Independent Pitching. Designed to look like ERA--the MLB average was right around 4.50--FIP measures only things that a pitcher is specifically responsible for, including strikeouts, walks, and home runs. The idea is to remove the impact of the defense behind them, the pitchers who relieve them, and any luck involved.
After calculating FIPs for each of the Twins' minor-league pitchers, I then adjusted them based on the run-scoring environments that they played in. The Twins' minor-league system is filled with extremely pitcher-friendly environments, which makes hitting prospects look worse and pitching prospects look better. In both cases the result is misleading numbers, but by accounting for that and normalizing run-scoring environments to fit the level of offense found in MLB, it paints a clearer picture.
Basically, the adjustments put everyone on an even playing field before evaluating their performance. A tremendous number of Twins minor leaguers posted ERAs in the 3.00s and 4.00s this year, but many of them did so while being below average at preventing runs because their league was such a low run-scoring environment. MLB as a whole had a 4.47 ERA in 2007, while the four full-season leagues in the Twins' minor-league system had ERAs of 3.98, 4.19, 3.96, and 3.78.
You're bound to post nice-looking ERAs when offense is 10-15 percent below MLB, which is why run-scoring adjustments are important. Once that's complete, I made further adjustments for how old each pitcher was relative to the prospects that he was competing against. Age is an important and often overlooked factor in projecting prospect development, and in most cases a 4.00 ERA from a 22-year-old at Double-A is a lot more impressive than a 3.50 ERA from a 28-year-old at Triple-A.
Before I get to the numbers, please note that these are not my rankings of the Twins' top prospects. That annual list will be published later this offseason and includes both multi-year track records and long-term potential as huge factors, whereas the numbers below focus solely on what each player did this year. The point here is to simply determine who had the best 2007 season by putting everyone's numbers through the same context-adjusting system.
100+ INNINGS FIP 70-99 INNINGS FIP 40-69 INNINGS FIPCole DeVries is an example of the various adjustments showing a pitcher in a worse light, because his nice-looking 3.41 ERA at low Single-A gets knocked down a peg with nearly every context adjustment. The Midwest League had a cumulative 3.78 ERA, which means that DeVries' 3.41 ERA there was the equivalent to a 4.03 ERA in a run-scoring environment equal to MLB. Beyond that, FIP shows that he pitched much worse than his ERA suggests.
DeVries had good control, but managed just 108 strikeouts in 148 innings and served up 17 homers in a league where power is rare, which is why his FIP was about 14 percent worse than league average. DeVries pitched worse than his ERA suggests and the pitcher-friendly environment that he played in means that his ERA wasn't as impressive as it looked to begin with, and he was also old for low Single-A as a 22-year-old who played three seasons at the University of Minnesota.
Consider that DeVries was in the same rotation as Tyler Robertson, whose FIP was 39 percent better than the MWL average as a 19-year-old with zero college experience. Robertson dominated a league filled with much older, more experienced players and his 2.48 adjusted FIP ranks second to only Kevin Slowey among pitchers with at least 100 innings. Meanwhile, DeVries goes from a raw ERA of 3.41 to an adjusted FIP of 5.36, which was fourth-worst among 18 pitchers who tossed at least 100 innings.
The best pitching season turned in by a Twins minor leaguer in 2007 came from Bradley Tippett, a 19-year-old Australian right-hander who went 7-1 with a 0.93 ERA and 51-to-4 strikeout-to-walk ratio in 38.2 innings out of the bullpen at rookie-level Elizabethton. Tippett's FIP nearly matched his incredible ERA, he served up just one homer to go along with the fantastic strikeout-to-walk ratio, and he was the same age as the rest of the league. It was nearly a perfect prospect season.
Elizabethton's bullpen also included Michael Tarsi, Michael McCardell, and Spencer Steedley, who posted numbers that were just slightly less dominant than Tippett. However, while Tippett was a teenager, Tarsi was a year older at 20 and both McCardell and Steedley were 22. Rookie-ball numbers should be taken with a grain of salt to begin with because of the small sample of games involved and rookie-ball numbers from 22-year-olds are far less impressive than they initially appear.
Along with Slowey's excellent 2.40 adjusted FIP, Scott Baker (2.75) and Matt Garza (2.86) were also great at Triple-A before combining to post a 4.20 ERA in 293 innings with the Twins. Nick Blackburn's 2.36 raw ERA in 148.2 innings between Double-A and Triple-A might suggest that he should also be included in that group, but his 3.74 adjusted FIP wasn't on the same level as Slowey, Baker, and Garza before he got lit up for 12 runs in 11.2 big-league innings.
Much has been made of the notion that the Twins' minor-league system boasts tons of high-quality pitching depth while being devoid of quality position players. My analysis of the 2007 numbers shows that to have been true this season, because the organization's No. 1 hitter had an adjusted OPS that was about 31 percent above average, while a total of 11 pitchers had an adjusted FIP that was at least that good.
The Twins' system contained 17 pitchers who were at least 20 percent better than average this season, while only four hitters could say the same. On the other hand, the number of players who were simply above average in 2007 was relatively close, with 35 pitchers and 30 hitters. Plus, if you look only at hitters with at least 250 plate appearances and pitchers with at least 250 batters faced, the picture is somewhat different:
FIP% OPS%The organization's top-level pitching is still far more impressive than the top-level hitting, but in terms of above-average performances from players who either batted 250 times or faced 250 batters, there were 23 pitchers and 29 hitters. There are indeed a ton of high-quality pitching prospects in the Twins' minor-league system and they lack the same type of high-quality hitters, but the position-player depth is better than advertised. Still, pitcher-friendly environments and all, the Twins' pitching depth is great.
Once you're done here, check out my latest "Daily Dose" column over at Rotoworld.
Tuesday, October 30, 2007
Twins 2007 Minor League Numbers: Hitters
Yesterday in this space I discussed the importance of putting minor-league performances in proper context based on adjustments for leagues, ballparks, run-scoring environments, ages, and defensive positions. I also laid out my method for arriving at those context-adjusted numbers, so you'll want to read that before going any further with this entry. Today I'll reveal what putting the numbers into context shows about the Twins' position-player prospects and later this week I'll do the same for pitchers.
Before I get to the good stuff, please note that these are not my rankings of the Twins' top prospects. That annual list will be published later this offseason and includes both multi-year track records and long-term potential as huge factors, whereas the numbers below focus solely on what each player did this year. The point here is to simply determine who had the best 2007 season by putting everyone's numbers through the same context-adjusting system.
The first step is taking raw numbers and adjusting them for run-scoring environments, with the idea being that everyone should first be placed on an even playing field before evaluating their performance. For example, former first-round pick Chris Parmelee batted just .239/.313/.414 in 128 games at low Single-A, which looks horrible at first glance. However, the Midwest League as a whole batted just .255/.324/.372 while scoring 8.6 combined runs per game in 2007.
The run-scoring environment that Parmelee played in this season was about 12 percent worse for offense than MLB, which is a big part of his underwhelming numbers. Once you normalize the Midwest League to fit the level of offense found in MLB, then Parmelee's hitting line improves to .251/.325/.471. That's still not great, but it's well above average and clearly superior to his raw numbers while better showing the significant power potential that Parmelee (pictured below) possesses.
Parmelee slugging just .414 this season is misleading, because he played in a league where hitting for power was exceptionally difficult. Once you account for that his adjusted slugging percentage jumps to .471, which shows him in a completely different light. Because the Twins' minor-league system is filled with teams playing in poor run-scoring environments, normalizing their organization-wide hitting numbers causes nearly every position player to receive a Parmelee-like boost.
Even after those adjustments the Twins still boast one of the weakest collections of hitting prospects in baseball, but as a group their position players weren't nearly as inept as the raw numbers suggest. Putting hitting prospects in leagues where the average team doesn't even slug .400 will do a lot to skew the perception of their ability and that's the case throughout the Twins' system. Once you account for run-scoring environments, here are the top adjusted 2007 hitting lines from Twins minor leaguers:
PA AVG OBP SLG OPSPlenty of the organization's hitting prospects had solid seasons at the plate, although five of the top six came from rookie-ball players who batted under 300 times. Even adjusted, those numbers are the product of a small sample of playing time and should also be taken with a grain of salt for another reason that I'll get into momentarily. Among players who batted at least 400 times, Brian Buscher put together the best season with an adjusted hitting line of .316/.388/.540 between Double-A and Triple-A.
The trio of rookie-ball hitters whose seasons ranked above Buscher--Deibinson Romero, Ozzie Lewis, and Ben Revere--show the importance of the next step, which is making further adjustments for age and defensive position. In other words, how does each player's offense compare to other players at his position and how old was each player relative to the prospects he was playing against. If you take the normalized numbers above and apply further adjustments based on age and position, you get:
AGE LEVEL POS OPS%Shown above is a player's age this season, which level(s) he played at, which position he primarily manned, and how his context-adjusted offense compared to "average." For instance, Parmelee was 19 years old, spent the entire season at low Single-A, played primarily right field, and was about 10 percent above average offensively once everything was taken into account. The first list took only offense and run-scoring environment into account, while this list also factored in age and position.
The two rankings are very different and a look at the aforementioned rookie-ball trio of Romero, Lewis, and Revere helps shows why. All three players posted an adjusted OPS above .900 while playing at rookie-ball, but Romero and Revere stay at the top of the second list while Lewis drops from second to 17th. The reason for that is two-fold, with the first being that Lewis split time between left field and designated hitter, while Revere (pictured below) manned center field and Romero played third base.
The further down the "defensive spectrum" that a player moves, the less impressive that his hitting becomes relative to other players at his position. In other words, a .900 OPS is much better coming from a shortstop than a first baseman. Lewis is essentially at the bottom of the defensive spectrum, while Revere is at the top and Romero is in the middle. Toss in their projected defensive value and it's easy to see why Lewis' .900 OPS doesn't look quite as impressive as it did initially.
Beyond that, Revere and Romero both joined the Twins' system without playing college ball and played this season at 19 and 20 years old, respectively. Meanwhile, Lewis played three years at Fresno State before being drafted and was 21 years old this season. The difference between 19, 20, and 21 may not seem like much, but even one year is huge when it comes to projecting player development and evaluating someone's performance in the low minors.
Plenty of 21-years-olds who're drafted out of college thrive at rookie-ball, in large part because they're older and have more high-level experience than most of the players they're competing against. Plus, the Twins have several 21-year-olds already at Double-A. On the surface Romero, Lewis, and Revere had similar seasons offensively, but dig a little deeper and you can see that Lewis' year wasn't on the same level. The same logic also holds true beyond rookie-ball, although perhaps to a lesser extent.
Based purely on offense Garrett Jones had the 14th-best season of any hitter in the Twins' system, but he was also a 26-year-old playing his third season at Triple-A and was limited to first base or an outfield corner defensively. Once you take Jones' normalized numbers and adjust them further for his age and position, he drops to 24th on the list. While guys like Lewis and Jones move down on the second list, up-the-middle defenders who were young for their leagues move up.
In particular, adjusting for age and position pushes Wilson Ramos into the top spot after ranking 12th on the offense-only list. Ramos batted .291/.345/.438 at low Single-A, which works out to an adjusted hitting line of .306/.358/.498. That nice line gets even nicer when you consider that Ramos was a 19-year-old catcher. An .856 OPS from anyone is good, but an .856 OPS from a teenager manning arguably the most important defensive position while playing in a full-season league is fantastic.
Once you account for the run-scoring environment that he played in and adjust for both position and age, Ramos' 2007 season was about 31 percent above average offensively. That's an amazing year for a prospect who many observers think can potentially be a tremendous defensive catcher and Ramos leads the way among a total of 30 Twins minor leaguers who were above average offensively while batting at least 200 times.
Among hitters who came to the plate at least 400 times, Triple-A Rochester's Jose Morales--last seen suffering a gruesome ankle injury during a call-up with the Twins--was best at around 20 percent above average thanks to being a 24-year-old catcher with a .318/.370/.427 adjusted hitting line. Also having strong context-adjusted seasons in 400-plus plate appearances were Trevor Plouffe, Brock Peterson, Joe Benson, Danny Valencia, Matt Tolbert, Parmelee, and Buscher.
Once you're done here, check out my latest "Daily Dose" column over at Rotoworld.
Monday, October 29, 2007
Twins 2007 Minor League Numbers: Intro
Later this offseason I'll be putting together my annual Top 40 Twins Prospects rankings, but before that happens it's important to evaluate how the team's minor leaguers performed in 2007. When it comes to major-league players, most fans recognize the need to take things like defense, home ballpark, league, and run-scoring environment into account when evaluating their performance, but for some reason those same factors are often overlooked in examining minor-league players.
Putting a player's numbers in context is even more important in the minors, because the differences between ballparks, leagues, and run-scoring environments are often far more extreme than in the majors. For an example, take a look at how different the two Triple-A leagues are when it comes to offense. As a whole, the Pacific Coast League batted .279/.346/.437 and saw 10.3 runs scored per game. Meanwhile, the International League batted .262/.332/.395 and saw 8.7 runs scored per game.
Most people wouldn't think twice about saying that a pitcher "had a 4.00 ERA at Triple-A," but "Triple-A" can be vastly different depending on the league. A 4.00 ERA in the Pacific Coast League was fantastic in 2007, checking in at about 15 percent better than average. Meanwhile, that same 4.00 ERA in the International League was actually worse than average in 2007. In fact, a 4.00 ERA in the PCL was roughly equivalent to a 3.30 ERA in the IL, and that's far from the most extreme league-to-league gap.
Beyond leagues, ballparks, and run-scoring environments, there are also factors like age, defense, and multi-year track records. For an example, compare Garrett Jones hitting .280/.334/.473 at Triple-A with Trevor Plouffe batting .274/.326/.410 at Double-A. With a 71-point edge in OPS, Jones' season appears on the surface to be much better. However, look deeper and there's little doubt that Plouffe had the more impressive season and is the superior prospect.
Jones was a first baseman playing his third season at Triple-A and at 26 years old was 1-2 years older than most of his competition. Plouffe was a shortstop getting his first taste of Double-A and at 21 years old was 2-3 years younger than most of his competition. For Jones and Plouffe, putting their numbers in context casts their performances in a completely different light, and the same can be said for players throughout the Twins' minor-league system.
In an effort to create an even playing field for evaluating how Twins minor leaguers performed in 2007, I've created a quick-and-dirty system that normalizes all leagues and run-scoring environments to fit the level of offense in MLB this season (.268/.336/.423 with 9.6 combined runs per game). I've also put in adjustments for a player's defensive position and age relative to the level he played at, so that guys like Jones and Plouffe can be compared properly.
On an organization-wide level much has been made, both here and elsewhere, of the Twins' lack of quality position-player prospects. That's accurate and it's telling that among all the position players in the entire Twins minor-league system who batted at least 350 times, not a single hitter posted an OPS of even .900. However, the system also features extremely pitcher-friendly run-scoring environments at every level, which means that the sub par numbers are much better than they initially appear.
Meanwhile, even more has been made of the Twins' abundance of quality pitching prospects. While also accurate, those same pitcher-friendly run-scoring environments play a big part in the pretty ERAs that were posted throughout the system. Just as the lack of .900-OPS hitters can't be written off entirely to a dearth of position-player talent, the system's never-ending stream of 3.00 ERAs can't be credited entirely to great pitching talent. Over the next few days I'll try to put all of that in context.
Last October, I created a WhatIfSports.com Hardball Dynasty league for readers of this blog. We're on the verge of completing Season 4 of "Gleeman World" and there will be a handful of open franchises this offseason. The league is filled with a bunch of friendly AG.com readers who fill the message board with daily chatter, but it's also extremely competitive. Because of that, any new owners would have to convince me that they're capable of devoting time to maintaining their team on a near-daily basis.
Previous experience with WhatIfSports, and especially Hardball Dynasty, is a substantial plus, but not necessarily required. However, Hardball Dynasty is unique and nothing like fantasy baseball, so you should at least read up on it before deciding if it's for you. If you're interested in claiming a spot and aren't worried about real-life responsibilities getting in the way of running a fake baseball team made up of pretend players, drop me an e-mail.
Once you're done here, check out my latest "Daily Dose" column over at Rotoworld.