Pebble Hunting: What It Means To Have The Best Farm System In Baseball, Part Three

Two years ago, we did a year-by-year retrospective of the Brewers’ 2004 farm system, which had been the best in baseball at that time. Last year, we did the same for the Angels’ 2005 farm system, which had been the best in baseball at that time. The idea was to see how valuable such a farm system is over time, and, more interestingly, how long the value of such a farm system holds. Does it open a team’s competitive window for three years? Five? Ten? Beyond?

This year’s retrospective is the 2006 Diamondbacks, and hoodoggy is it a fun one. Way more fun than the other two. Both the Brewers and the Angels got a ton of value out of their systems—about 75 WARP for Milwaukee at the time we wrote the piece, about 105 for the Angels–but in different ways: The Brewers got star performances from four of their top seven prospects (Weeks, Fielder, Hardy, Hart) and little else. The Angels, meanwhile, got most of their value from the middle and bottom of their system: Prospects ranked 3rd (Aybar), 5th (Morales), 7th (Santana), 8th (Kendrick), 11th (Trumbo), 14th (Saunders), 16th (Izturis), 19th (Jepsen) and 29th (Napoli). Each got a little value from second-generation profits (trades and such), but more than 90 percent of the production came directly from the 30 players on each team’s list.

The Diamondbacks’ system, on the other hand, makes for one giant ridiculous trade tree. The top 12 prospects (and 14 of the top 15) were eventually traded, some shortly after these rankings came out and some this offseason. If our first two teams showed the value of 30 good players, the Diamondbacks show the value of 30 good properties. Will they top the Brewers and Angels’ systems? The world will finally know!

Year One (2006)
28 of 30 prospects remain.

A good reminder of the huge talent gap between the top of a system and a guy in the mid- or late-20s: There’s usually a player on these lists who gets released even before he plays another game. In this case, it was no. 25 Josh Kroeger, a corner outfielder who had been bonkers at Triple-A in 2004 but hit just .261/.316/.422 at the same level as a 22-year-old in 2005. The Phillies claimed him, but he never did join Glenn Dishman, Rafael Quirico and Bronson Heflin in the exclusive Guys Who Played For The Phillies At Least Once club.

Also leaving: no. 24 Jason Bulger, traded to the Angels (where he produced .1 WARP in four seasons) for Alberto Callaspo. Ten of the prospects appeared in the majors for Arizona, including no. 2 prospect Stephen Drew and no. 4 Carlos Quentin, who were each awesome. The group produces 4.2 WARP—or, if you exclude negative values, which I will going forward unless noted, 4.9 WARP. In Josh Byrnes’ first year as GM, the Diamondbacks win 76 games.

Year Two (2007)
24 of 30 prospects remain.

The Diamondbacks entered the season as PECOTA’s favorites in the NL West, if not so in the world at large: If you were in Vegas that spring you could get the Diamondbacks at 40 to 1 to win the World Series; to put that in perspective, that’s what you’d get if you bet on the Rangers today. “Even though the team projection comes off as optimistic at first glance, not many of the individual forecasts do,” Nate Silver wrote. He continued,

You probably have to go back to the 1994 Indians to find a team with this much position-player talent coming up through the system together. We have the Diamondbacks projected higher than most other outlets, but that's because they're a young club, and as I mentioned in Monday's chat, I think the other forecasting systems have some work to do before they catch up with the quality of Clay's translations.

It's worth mentioning that PECOTA is a self-correcting forecasting system—it works by means of a two-stage process. First it creates a baseline forecast for each player: both the 'test player' (the guy whose stats we are projecting) and the comparables. Then it sees how the comparable players performed against that baseline, and makes adjustments to the test player's forecast accordingly. If the DTs were systematically overestimating young talent, for example, then PECOTA would catch that by noticing that the comparables were failing to live up to their baselines, and would downgrade the test player's projection as a result. Being able to hedge against its own assumptions is a big part of PECOTA's accuracy.

That’s not all relevant to this exercise we’re doing today, but it’s interesting! The Diamondbacks won 90 games, so score one for PECOTA, except that, actually, they were outscored (they overperformed pythag by 11 games) and the core-five prospects who PECOTA (and, in a “sticking to my guns” note, Silver) was so bullish on largely flopped:

You know who was a big contributor on that team? Mark Reynolds, who wasn’t on the top 30 prospects a year earlier. Kellen Raab (no. 29) was on the top 30 ahead of Reynolds, but by 2007 Raab was out of affiliated ball. He was released after a brutal 2006 season in High-A and spent the Diamondbacks’ division-champions season pitching in independent ball. Then he retired.

Other Top 30 guys who left the system that year: no. 18 Matt Chico and no. 9 Garrett Mock, who were traded to Washington for Livan Hernandez; and no. 26 Alberto Gonzalez, who was traded to New York for Randy Johnson and a package of stale Attorney General jokes. Chico and Mock were the entire package headed out, so they’ll be credited with all of Livo’s contributions going forward; Gonzalez was just a quarter of the package sent east for Johnson, and with a slightly higher profile than the others we’ll credit him with a third of Johnson’s future value.

In all, the top 30 directly produced 6.4 WARP, led by no. 11 Micah Owings. Livan Hernandez produced virtually no value; Randy Johnson did, and the Top 30 gets credit for about a half win from him. In all: 7 WARP. Nothing staggering, but enough to get Arizona to the postseason.

Year Three (2008)
15 of 30 remain.

So we’re 24 months later and half the system is gone. Carlos Quentin (no. 4) was traded for Chris Carter. The other Chris Carter (no. 12) was traded for Emiliano Fruto. And then the first Chris Carter, who was actually the Diamondbacks’ second Chris Carter (but the first Chris Carter mentioned in this paragraph), was traded, along with no. 6 Carlos Gonzalez and no. 14 Greg Smith for Dan Haren. There were other, non-top-30 guys in that trade: Brett Anderson, Dana Eveland and Aaron Cunningham; measuring the mood around that trade, I think we can give the Top 30 guys credit for 62 percent of the Dan Haren acquisition.

Dustin Nippert (no. 7) had established himself by now as a bad major-league reliever, and he was traded for a 24-year-old High-A reliever named Jose Marte. No. 16 Enrique Gonzalez had made 18 starts for the 2006 Baby ‘Backs, but they weren’t good, and the 27 he made in Triple-A in 2007 were worse, so he was waived. You might remember him from his later work with the—no, wait, you don’t remember him. No. 17 Jon Zeringue was released; he never made the majors for the Diamondbacks (or anyone else), though he did go over to the A’s system and immediately doubled his walk rate; kind of looks like Youkilis and everything, so you can imagine being an A’s blogger in the summer of 2007 and tagging this post something like Billy Beane Does It Again.

Brian Barden (no. 20) was waived after batting 12 times for Arizona. He would go on to win the NL Rookie of the Month award in April 2009 with the Cardinals. He hit .385/.432/.641 that month. He hit .162/.219/.206 the rest of his career. So cross “one month” off your list of “big enough sample sizes probably.”

No. 27 Jereme Milons was released after failing to clear High-A, and the Elmer Dessens trade tree was officially rotting firewood.

Owings was traded mid-year for Adam Dunn, with assurances from Byrnes that at the very least they would get two draft picks when Dunn left as a free agent that winter. Dunn was worth about a win down the stretch. Then, strangely, he left as a free agent without the Diamondbacks offering him arbitration, and thus his trail ends there.

Billy Buckner, having been acquired for Alberto Callaspo, who’d been acquired for Jason Bulger, appears in the majors and provides the first third-generation contribution from the Top 30.

The Top 30 produced 11.5 WARP directly this year, almost entirely from Drew, Jackson and Young, all of whom were above-average regulars. No. 1 Justin Upton, having debuted the previous season, played regularly for the first time but didn’t add much as a 20-year-old. In all, the top 30 adds 16.1 WARP, but the Diamondbacks win 82 games and miss the playoffs.

Year Four (2009)
Eight of 30 remain.

Just 36 months later and nearly three-quarters of the Diamondbacks’ top 30 prospects are out of the system. Further, seven of the players they had used their top 30 to acquire were gone, including three players who had been traded for and subsequently traded. And while there are only eight of the top 30 remaining, there are 15 players who were either in or descended from the top 30. One is Jon Rauch, who was acquired for no. 28 Emilio Bonifacio. One is Brandon Allen, acquired for no. 15 Tony Pena. They’ve also got Billy Buckner now, acquired for Callaspo, who had been acquired for Bulger. Buckner gets ranked sixth-best prospect in the system, higher than Bulger ever was.

It’s in Year Four that the releasing of the flops becomes more common: No. 21 Brandon Medders is granted free agency after three years of replacement level relief, and no. 22 Jamie D’Antona is granted free agency after 19 plate appearances with the team that drafted him in the second round. He spent two years hitting like, oh, about like Mark Trumbo in Japan. Yakyu Night Owl still wishes him happy birthday.

No. 23 Cesar Nicolas, no. 19 A.J. Shappi, and no. 13 Matt Green are all released. None plays affiliated ball again. Is Shappi an Herbalife salesman now? It’s a good question that we’ll probably never know the answer to.

The Diamondbacks win 70 games and finish in last place. Without the contributions of that farm system, and particularly Justin Upton, who breaks out as a star, they’d have lost 100, easily:

First generation: 11.9 WARP
Second generation: 3.0 WARP
Third generation: .3
Total: 15.2.

Year Five (2010)
Six of 30 remain.

The six include no. 3 Conor Jackson, who brings back Sam Demel in a midseason trade with the A’s.

No. 30 Agustin Murillo, one of two remaining players in the system without big-league experience, gets released. In 2008, Arizona had loaned him to a team in Monterrey, Mexico; he was named MVP. After returning to the states in 2009, he went to Mexico for good in 2010. He’s still there, and in 2014 he had his best season, hitting .340/.417/.583 and, at age 32, playing his second game at shortstop. Tony Pena (a fellow top 30 Diamondback) is a teammate.

Dan Haren gets traded midseason for Joe Saunders and some pitching prospects, so 62 percent of whatever Joe Saunders adds will be credited to the Top 30. Spoiler: Joe Saunders doesn’t add.

Billy Buckner gets traded for Dontrelle Willis. Our first fourth-generation product!

First generation: 14.6 WARP
Second generation: 2.0 WARP
Third generation: 0
Fourth generation: 0

The Diamondbacks win 65 games. Josh Byrnes is fired.

Year Six (2011)
Five of 30 remain.

But an additional nine players who were acquired for members of the 30 remain. Those nine include three of their top eight prospects. Also Brad Ziegler, acquired from the A’s for Brandon Allen. Something like 11 of the players we’ve mentioned so far in this article went on to play for the A’s.

And, finally, the core prospects themselves all produce at the same time: Upton, Young and Montero all have (by some measures) the best years of their career, and Drew is at least average. Those four and Triple-A pitcher Matt Torra are the only ones left from that class, and Torra will be traded for cash midseason. So just the four. And they’re all really good. The plan worked. The Diamondbacks win 94 games and make the playoffs.

First generation: 14.1 WARP
Second generation: 0.3 WARP
Third generation: 0.2 WARP
Fourth generation: 0.0 WARP

Year Seven (2012)
Four of 30 remain.

Two of the four are still good; Young’s career begins its cliff dive, and Drew is injured and awful and gets traded mid-year (to the A’s, obviously). But besides the two are now two new contributors: Patrick Corbin (via Dan Haren via Carlos Gonzalez et al) and Wade Miley (via Livan Hernandez’s compensation pick via Garrett Mock and Matt Chico). They throw 300 innings; Miley is probably the club’s best pitcher and makes the All-Star team. Miley has thrown more career innings at that point than Mock or Chico did before their retirements.

First generation: 9.6 WARP
Second generation: 0.0 WARP
Third generation: 4.2 WARP
Fourth generation: 0.0 WARP

The Diamondbacks finish in third place, 81-81.

Year Eight (2013)
One of 30 remains.

Chris Young is traded, to the A’s, for Cliff Pennington. Justin Upton is traded, with Chris Johnson, for Martin Prado, Randall Delgado, Zeke Spruill, Brandon Drury and Nick Ahmed. (I’ll credit Chris Johnson with no more than 10 percent of the acquisition, so Upton is 90 percent.) Montero remains, as do 11 players acquired for members of the top 30.

Corbin joins Miley at the top of the Diamondbacks rotation, as the two of them lead the team in innings. Ziegler is the club’s best reliever and ends up with 13 saves in the summer. Prado and Pennington are both useful. Skaggs is the club’s best prospect, better than Archie Bradley. Eight years later you could argue that, with the exception of Paul Goldschmidt, most of the Diamondbacks’ most valuable assets can be credited to the top 30.

First generation: 0.9 WARP
Second generation: 2.0 WARP
Third generation: 2.7 WARP
Fourth generation: 0.0 WARP

But the Diamondbacks win 81.

Year Nine (2014)
One of 30 remains.

Montero is still here, though it's getting less clear that he’s worth the contract extension that keeps him on the team.

First generation: 1.8 WARP
Second generation: 2.3 WARP
Third generation: 0.7 WARP
Fourth generation: 0.0 WARP

Year Ten (2015)
None of the 30 remains.

We could end the accounting now and say that the Diamondbacks got 99 wins out of their top 30, nearly as many as the Angels (105) and quite a bit more than the Brewers (74). But there’s a lot of story left here. After trading Prado last summer, and Montero, Miley and and Spruill this winter, they have 16 players who are remnants of that farm system:

• Patrick Corbin (Haren via Gonzalez et al)
• Brandon Drury (Upton)
• Nick Ahmed (Upton)
• Randall Delgado (Upton)
• Sean Jamieson (Drew)
• Cliff Pennington (Young)
• Zack Godly (Montero)
• Jeferson Mejia (Montero)
• Peter O’Brien (Prado via Upton)
• Myles Smith (Spruill via Upton)
• Brad Ziegler (Allen via Pena)
• A.J. Schugel (Skaggs via Haren via Gonzalez et al)
• Mark Trumbo (Skaggs via Haren via Gonzalez et al)
• Raymel Flores (Miley via Hernandez via Chico/Mock)
• Rubby De La Rosa (Miley via Hernandez via Chico/Mock)
• Allen Webster (Miley via Hernandez via Chico/Mock)

Half of those guys are awful, and as a group it’s obviously nothing like Justin Upton, Stephen Drew, Carlos Quentin and so on. But after 10 years the Diamondbacks have managed to string out the original group’s value, and the remainders include nos. 1, 8 and 10 on the club’s 25-and-under rankings. If a couple pieces hit, this Farm Tree could keep going for another generation or two.

What’s interesting about the three teams we’ve looked at so far: All three got substantial value out of their cores, yes. The Brewers clumped their value into a few years; the Diamondbacks, via trades and so forth, managed to get a smooth 10 or 15 win boost every year for a decade. But none of the three teams won a World Series. Twenty-seven seasons (or 30 if you count the Brewers and Angels’ seasons since my articles) and none of the teams won a World Series, or even made a World Series. The Brewers and Diamondbacks only made the postseason twice each. Each year we do this, we get a surer sense of what a club with the best farm system in baseball can expect: Around 100 wins above replacement, most of it at sub-market prices, peaking three to seven years after the rankings come out, but some remnants of that value lasting into the next decade. (The Angels, for instance, just added Andrew Heaney for Howie Kendrick, a member of their 2005 top 30. So their tree is still alive.)

But we also have a surer sense of what that club can’t expect: A dynasty, a run of dominance, or even a single World Series appearance. The best farm system in the world will get you 15 or 20 extra wins in a good year. Don’t knock that, obviously, but just as we tend to be disappointed even by the prospects who turn out very, very good (like some of the guys on the Diamondbacks list), we probably tend to be disappointed even by the farm systems that turned out to be very, very good. Two GMs were fired for not doing more with these 30 Arizona prospects.