Not so long ago, I looked back at the nine years since the Milwaukee Brewers had the best farm system in baseball to see what the impact of having the best farm system in baseball might be, on a 10-year timeline. I wasn’t totally sure what I hoped to get at or if I did get at it, but people seemed to like it and asked me to do the worst farm system of the same year, so I’ll do that now. I still don’t know what I’m hoping to get at, so we’ll see where this goes.

The worst farm system in baseball at the time belonged to the Montreal Expos, according to Baseball America, but the Expos were in a weird place and it’s hard to consider them a representative franchise. So we’re going to look at the no. 29 team instead: the Houston Astros. Since 2004, the Astros have played in a World Series, so how bad could it really be? Really, really, really bad!

28 of 30 prospects remain in the organization, five of 30 on the 25-man roster.

Like the Brewers, the Astros had already cut ties with one of their 30 before the ranking even came out, though they then reestablished ties with him: Rodrigo Rosario (ranked no. 22 in the system) was released but then re-signed with the club that winter. But another pitcher on the list, Jimmy Barrett, didn’t pitch for Houston in 2004 and was released in June.

Barrett’s ranking as the 27th-best prospect in the system gives you a pretty good sense of the Astros' talent at the time. In 2003, he was 22 and pitching in High-A. He struck out 4.9 batters per nine and walked 3.8. He was a 6’ 2” right-hander. In five minor-league seasons, he had one year with an ERA below 4.40. There’s a weird scouting report from 2003 that somehow remains online and describes his “desire for winner” as “minimal.” And he made it into the book, so.

The Astros begin the process of trying to extract some value out of a lousy system, while strengthening the club in the middle of a competitive window, by trading. John Buck (no. 3) goes to Kansas City in the Carlos Beltran move.

It’s very difficult to isolate Buck’s value in this move, as the Astros also send away Octavio Dotel (who had a year and a half left before free agency). After two years as a top-100 prospect, Buck had had a terrible year in Triple-A in 2003, then a good half-year in Triple-A in 2004. Imagine if the Blue Jays traded J.P. Arencibia and Casey Janssen as a package right now. (Janssen is signed for two more years plus a club option.) What percentage of the return would you say gets credited to Arencibia, and what percentage to Janssen? Maybe 30 percent to the catcher? Let’s say 30 percent to the catcher.

So Buck brings back 30 percent of Carlos Beltran, who a) produces 4.0 WARP as an Astro in a half season and b) carries them in the postseason. Beltran leaves after that season and the Astros get two compensation picks: they take Eli Iorg and Tommy Manzella, 30 percent of each of which will be credited to the 2004 farm system.

Through one year, the Brewers system and the Astros system produce about the same amount of value: .8 WARP for the Brewers and .3 WARP for the Astros; or, if you exclude negative contributions (as we will in the rest of this exercise), 1.3 for the Brewers and 1.0 for the Astros. Tack on 1.2 WARP for 30 Percent Beltran and the Astros have done okay so far.

The Astros go 92-70, finish in second place, and lose in the NLCS.

24 of 30 prospects remain in the organization, eight of 30 on the 25-man roster.

Four of the top 30 are let go for no return without contributing to the big-league club: Rosario and D.J. Houlton (26) are both taken by other teams in the minor-league draft, Jesse Carlson (21) is granted free agency; and Derick Grigsby (29) puts his career on hold because of depression, and never pitches again. Grigsby was the team’s first-round pick less than three years earlier. A profile of Grigsby written at the time is pretty disturbing:

On a nearby shelf is a picture of Derick taken in high school. He's smiling broadly, looking happy as can be. A beautiful blonde has her arms around him.

He seems so sad as he speaks that it's hard to imagine he's the same guy.

Has he had thoughts of suicide?

"They always ask me that at the counselor's office," he says. "It hasn't gotten that far. I could never see myself doing that. I don't think it's going to get that bad."

The eight Astros who appeared in the majors contribute 5.7 positive WARP, bolstered by strong performances by Willy Taveras and Jason Lane. That's better than the Brewers’ group, which contributes 5.0 WARP in 2005. The Astros go 89-73, win the Wild Card, and lose in the World Series.

21 of 30 prospects remain in the organization, 12 of 30 on the 25-man roster.

Three Astros—no. 14 Tommy Whiteman who is, in fact, a white man; no. 24 Todd Self who is, in fact, his self;  and no. 28 Tony Pluta who does, in fact, have a tone—are let go. Otherwise, it’s a year mostly of promotions. Top prospect Taylor Buchholz makes his debut, and by the end of the season all nine of the Astros’ top prospects from 2004 (including Buck) have qualified for big-league pensions. Twelve of the top 15 have. That’s a very different question than “how much value has the system produced,” but it’s not a bad record up to this point.

Of the 12, only Taveras is making much of a difference. He produces 2.9 WARP. The rest of the 2004 class produces just 1.1 positive WARP, led by Chris Burke (no. 4). Iorg and Manzella are playing in the minors, too.

The Astros go 82-80, finish second in the NL Central for the third year in a row, but don’t get the Wild Card this time.

12 of 30 prospects remain in the organization, six of 30 on the 25-man roster.

In one offseason, the class of 2004 almost totally disappears. Most prominently, Taylor Buchholz, Willy Taveras, and Jason Hirsh (no. 8) are traded to Colorado for Jason Jennings and Miguel Asencio. Buchholz was the best prospect in the system in 2004, and Taveras has, at 5.4, more major-league WARP than the rest of the system combined, so it’s a real gutting.

Jennings and Asencio, unfortunately, do nothing. Jennings produces 0.0 WARP in his single year with Houston and brings back no compensation when he leaves as a free agent. Asencio plays one year in the minors and is released.

Also released: no. 13 Mike Gallo, who appeared on the big-league squad in 2004, 2005, and 2006 but produced -0.5 WARP; no. 15 Ezequiel Astacio, who also appears in the majors but is below replacement level; no. 18 Cliff Davis, who is out of baseball at age 20; no. 20 Scott Robinson; and no. 23 Charlton Jimerson, after a couple appearances in the big leagues.

Mitch Talbot is traded (along with an unexciting 25-year-old in Double-A named Ben Zobrist) mid-season for Aubrey Huff. Huff produces 0.3 WARP for Houston and leaves as a free agent, with no compensation tied to him. Including Huff and Jennings, the farm system produces a total of 1.7 WARP in 2007. The Brewers farm system produced 19 WARP the same season.

The Astros sign Carlos Lee and Woody Williams, forfeiting draft picks for each. They go 73-89 and finish fourth in the NL Central, behind Milwaukee.

Five of 30 prospects remain in the organization, one of 30 on the 25-man roster.

To put the above in perspective: the Brewers still had 11 of their prospects in 2008, and were getting direct contributions from seven of them.

Matt Albers (no. 9) is traded as part of a big package for Miguel Tejada; his importance in the deal is minor, so we won’t give the system credit for any of Tejada’s production. Jason Lane (no. 2) is traded for cash considerations to San Diego. At 3.3 WARP, he’s the second-most valuable member of the 2003 class, to date. Josh Anderson (no. 18) is traded for Oscar Villarreal, who produces -0.7 WARP in his lone season as an Astro.

Somewhat trickier is the trade for Jose Valverde. That costs Houston Chris Burke (no. 4) and Chad Qualls (no. 7) and also Juan Gutierrez, who isn’t part of the group of 30 we’re looking at. Gutierrez would ultimately close a bunch of games for Arizona, but at the time I think he was just a live arm coming off a bad year at Triple-A. From the Annual the offseason of the trade:

the D'backs would be happy to just get a good year from him in middle relief.

Burke was a post-hype sleeper and Qualls was a rubber arm who had established himself with three pretty good years of relief, so I think we can call them 80 percent of the perceived value going to Arizona in this trade. So the system has produced 80 Percent Valverde.

Fernando Nieve (no. 5) is the only remaining prospect who appears in the majors, and he produces 0.1 WARP in 2008. Eighty Percent Valverde is worth 0.9 WARP. At 1.0 WARP, then, the system badly trails Milwaukee’s 12.7.

The Astros, bolstered by 2004 draft pick Hunter Pence and trade acquisition Michael Bourn, go 86-75 but finish third in the NL Central, behind Wild Card Milwaukee.

Two of 30 prospects remain in the organization, one of 30 on the 25-man roster.

Raymar Diaz (19), Jared Gothreaux (16), and Nieve are waived, released, or some variation of such. The two who remain are:

The Astros also have the remnants of the system in Iorg and Manzella, still playing their way up through the system, and 80 Percent Valverde, who teams up with Paulino to produce a total of 0.8 WARP for the big-league club in 2009.

The Astros go 74-88 and finish fifth in the NL Central, behind Milwaukee.

One of 30 prospects remains in the organization, one of 30 on the 25-man roster.

McKeller disappears, so only Paulino remains. Except that Manzella makes his debut, so he’s in the math now; and Valverde leaves and gives the Astros two compensation picks, which they use on Mike Foltynewicz and Michael Kvasnicka. So the system is nearly dead! But the system is alive anew!

Not that Manzella, or Paulino, or anyone else is doing much. Paulino and Manzella combine for 1.2 WARP, compared to the Brewers’ 11.9 WARP. The Astros go 76-86 and finish in fourth place, a game behind Milwaukee.

Zero of 30 prospects remain in the organization.

Paulino is traded for Clint Barmes, who produces 1.3 WARP—the most the Astros have drawn from the 2004 prospects since 2007—before leaving as a free agent. Tommy Manzella plays and is worth -0.4 WARP.

It’s notable here that the Astros, eight years later, still had one of the worst farm systems in baseball. There is a lot of blame to go around: the Astros didn't get any first-round pick in 2003, 2004, or 2007; they traded prospects for veterans; and, whether it was a drafting thing or a development thing, their top picks in this decade were mostly disappointing. Going back to 1998: Mike Nannini, Mike Rosamond, Robert Stiehl, Burke, Grigsby, Brian Bogusevic, Iorg, Maxwell Sapp, Jason Castro.

Those are causes that can’t be totally separated from the farm system the Astros started with. Without a good farm system, they had to constantly fill holes from outside the organization. The cost of filling those holes is usually young talent, either in swaps of players with six years of club control for players with one or two (or fewer) years of club control; or in free agent compensation picks.

So the Astros never really recovered their farm system until now, nearly a decade after the 2004 group. Here’s how the organization has ranked each year since 2004, with Kevin Goldstein’s rankings replacing BA’s starting in 2007:

  • 2005: 22nd
  • 2006: 20th
  • 2007: 28th
  • 2008: 29th
  • 2009: 30th
  • 2010: 28th
  • 2011: 28th
  • 2012: 26th

In 2012, the Astros don’t get a single at-bat or inning from a member of the 2004 prospect group, or from any player acquired directly or indirectly for a member of the 2004 prospect group. Manzella plays in the minors but is waived and no longer with the organization. Two third-generation players remain in the system: Foltynewicz and Kvasnicka, both drafted as compensation after the departure of Jose Valverde, who was acquired in a trade for the 2004 Astros’ fourth- and seventh-best prospects. Here’s the unwieldy table:

Player 2004 2005 2006 2007 2008 2009 2010 2011 2012
Taylor Buchholz Mi Mi -0.7 x x x x x x
Jason Lane 0.8 2.1 0.2 0.2 x x x x x
John Buck Mi x x x x x x x x
Chris Burke -0.2 -0.1 0.7 -0.5 x x x x x
Fernando Nieve Mi Mi -0.3 Inj 0.1 x x x x
Hector Gimenez Mi Mi 0 Mi x x x x x
Chad Qualls 0.2 0.8 -0.4 0.9 x x x x x
Jason Hirsh Mi Mi -0.3 x x x x x x
Matt Albers Mi Mi 0.1 -0.8 x x x x x
Jimmy Barthmaier Mi Mi Mi Mi x x x x x
Willy Taveras 0 2.5 2.9 x x x x x x
Mitch Talbot Mi Mi Mi x x x x x x
Mike Gallo -0.5 0.1 -0.1 x x x x x x
Tommy Whiteman Mi Mi x x x x x x x
Ezequiel Astacio Mi 0.2 -0.1 x x x x x x
Jared Gothreaux Mi Mi Mi Mi Mi x x x x
Cliff Davis Mi Mi Mi x x x x x x
Josh Anderson Mi Mi Mi 0.3 x x x x x
Raymar Diaz Mi Mi Mi Mi Mi x x x x
Scott Robinson Mi Mi Mi x x x x x x
Jesse Carlson Mi x x x x x x x x
Rodrigo Rosario Mi x x x x x x x x
Charlton Jimerson Mi 0 0.1 x x x x x x
Todd Self Mi -0.1 x x x x x x x
Ryan McKeller Mi Mi Mi Mi Mi Mi x x x
D.J. Houlton Mi x x x x x x x x
Jimmy Barrett x x x x x x x x x
Tony Pluta Inj Mi x x x x x x x
Derick Grigsby Mi x x x x x x x x
Felipe Paulino Mi Mi Mi -0.3 Mi 0.2 1.2 x x
Miguel Asencio x x x Mi x x x x x
Jason Jennings x x x 0 x x x x x
Carlos Beltran 1.2 x x x x x x x x
Eli Iorg x Mi Mi Mi Mi Mi x x x
Tommy Manzella x Mi Mi Mi Mi Mi 0 -0.4 Mi
Jose Valverde x x x x 0.9 0.6 x x x
Aubrey Huff x x x 0.3 x x x x x
Oscar Villarreal x x x x -0.7 x x x x
Clint Barmes x x x x x x x 1.3 x
Mike Foltynewicz x x x x x x Mi Mi Mi
Michael Kvasnicka x x x x x x Mi Mi Mi

Foltynewicz was ranked this year by Jason Parks as the 10th-best prospect in the system (Kvasnicka has stalled in A-Ball) so there is still hope that the Astros can get some value out of that original group of 30. But the difference between the Astros’ system and the Brewers system since 2004 is, so far, stark:

Total WARP (prospects plus return for prospects)
Brewers: 74.9
Astros: 12.4

Total positive WARP (excluding all negative-WARP seasons)
Brewers: 79.0
Astros: 17.9


Of course, the Brewers had to pay a lot more money to their players as they got more service time, so, lastly, here’s the money spent on salaries (estimated) to produce these WARP:
Brewers: $136 million
Astros: $39 million

For $97 million in extra payroll, the Brewers got (to date) 62.5 more wins, in an era in which wins cost (on average) just around $4 million apiece.

The Brewers still have Rickie Weeks and Corey Hart signed, Carlos Gomez (in return for J.J. Hardy) signed, and two draft picks that produced legit prospects in the current system. Which means they could actually still get more value out of their 2004 farm system from 2013 on than the Astros got in total.

These teams’ results might not be typical—there is no typical—but, generally speaking, this is what it means to have the worst farm system in baseball.

Thanks to Ben Lindbergh for research assistance

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
Informative and entertaining. Bravo on coming up with "no. 24 Todd Self who is, in fact, himself".
Outstanding work. Scary stuff for fans of teams with bad farm systems, though.
Completely depends on the organization. Just because the farm is bad now doesn't mean the team will be bad in 10 years. This is the Astros. Like Sam mentions, the worst farm system was actually the Expos. While it took the franchise a few (read: lots) of years to turn into a contender, they did it.
Now in the AL West, Houston could challenge the '62 Mets for losses in a year.
Very, very nice. Should be done with other starting dates and other best/worst pairs.
This has been corrected thanks to @KevinBassStache, who noted that I had Beltran traded a year later than he actually was.
Great two part series, really enjoyed reading it. and timely, with all the prospect rankings coming out. seeing the attrition rate of top-30 prospects after the top 5 gives me some perspective on the likely value of my team's deep but not very star-heavy system.
"For $97 million in extra payroll, the Brewers got (to date) 62.5 more wins, in an era in which wins cost (on average) just around $4 million apiece."

Interesting. How much does it cost to produce a scouting/farm system that produces an extra 62 1/2 wins compared to a really bad system (replacement level farm system?)?
I don't think it really costs all that much more to have a good farm system than a bad one. There is a lot of information out there these days about just about every player. As was pointed out quite some time ago on BP, general managers need to know where they are on the 'success cycle'. The Giants with most of their impact players locked up for the next 3 years at reasonable cost and looking at big pay days after that need to be in 'win now' mode. If they traded Gary Brown and another prospect at the trade deadline for say, Josh Willingham, that would be defensible. If the Astros did it, it wouldn't be. The Astros current problem stems from the fact that they weren't just 'one player away' from winning the WS, but their one near success fooled them into thinking they were.
There is one important problem with this analysis. Prior to the 2007 meltdown, and including some years when they were contenders for the "worst system in baseball" title, the Astros had finished lower than second in their division exactly one time since 1994. It was hardly a "one near success" situation for them. On the contrary, they had been having "near successes" for so long that it wasn't unreasonable -- in isolation -- for them to think that their long record of near successes could become real successes with only minor tweaks. A willingness to literally sell the farm gets much easier to understand against that background.

Note also with that long record of "near successes," they'd been drafting (or wrote off chances to draft) late in the first round for a long time. It has been demonstrated many times, here and elsewhere, that the value of a first-round draft slot declines very rapidly as one gets deeper into the first round. Between 1994 and the 2007 meltdown, they had only one chance to draft in the top ten slots, where the real talent is concentrated. Admittedly, they screwed up that one chance; Chris Burke didn't have the kind of career that a top-of-the-first-round guy really should have. Still, you can't extract value from guys who are gone before you have the opportunity to draft them.
How did they achieve the success of the previous 10 years? Looking at the 96 team. At leat Biggio, Reynolds, and Wagner drafted by Houston. Alou acquired in trade for Astros #1 pick from 96. Bell, Hampton, Nitkowski, Ausmus, Gutierrez and Berry acquired in trades using mainly players they had drafted. Bagwell acquired in trade straight up for Larry Anderson who they had signed as an FA 2 years prior. Not sure how this compares to other successful teams, but looks like they drafted well and were able to trade many of their draft picks for players that contributed.
Meant 98 team that went 102-64 which was in the middle of winning their division 4 out of 5 seasons.
Thanks, I really enjoyed both of these articles. During the times when you're struggling to come up with something to write about, please consider doing more of these. Perhaps, you could next look at some of the more interesting farm systems from years past. Generation K versus Hudson Mulder Zito would be interesting as a comparison of when things go so wrong versus so right. The articles don't necessarily need to lead anywhere; often the education is in the journey, right?
I would absolutely love to see this on the Orioles over time.
Very interesting analysis. Get the coders busy making the article's final table (or something like it) an automated day-to-day (or even week-to-week) feature in the BP site for every MLB team and I do believe you'd have you next big hit.

Great work Sam. Keep it up. This is the kind of critical organizational analysis (with actual number values attached ... WARP, etc. that I believe many subscribers would love to see more of.
Excellent article. Lots of interesting points.

I do think the Astros' 2004 farm system needs to be evaluated with consideration that they continued to draft poorly and neglect the farm system for the rest of Drayton's ownership. McMullen handed off the team with the best farm system in baseball, and the products of that system fueled the success of Drayton's first 10 years. Drayton wanted to put a hard cap on payroll and cut costs everywhere else--like player development--at the same time he wanted the ego trip of signing expensive free agents. The Astros were the ultimate stars and scrubs team throughout Drayton's ownership.

The new ownership and Luhnow are turning things around as quickly as possible. Like Montreal/Washington, they have two #1 picks in a row (and will probably have a third). You don't get there without some dreadful years, again like Montreal/Washington. Astros fans hope the analogy continues. If the Astros in 2018 are where Washington is today, it will be all good. On the other hand, if like Pittsburgh or Kansas City, not so good.
Unfortunately, the Astros won't have 1-1 back to back years with a Stasburg/Harper.
Wait, so you're telling me that badly developing badly drafted players produces results that are bad?

May I send this article through Tachyon Express to the Indians of the previous decade? This may a-splode their heads, but I'm willing to have retroactively assumed that risk.
Sam, that scouting report of Jimmy Barrett from 2003 is actually a player screen from an old version of the PC game Out of the Park Baseball. It's not "real".