Fun with WAR: The best and worst Rockies rosters, plus one somewhere in between

It’s December, and the Rockies are on the cusp of playoff contention. I say this not based on the untrue cliché that before the first game, every team has the same record and an equal chance. Instead, I say this based on the team’s composite WAR projections.

Dan Szymborski recently rolled out the ZiPS projections for the 2015 Rockies. The projected WAR for the starting position players, the expected rotation, and the bullpen is about 30, which is the rough threshold for a spot in the playoffs. According to a recent Jeff Sullivan article on FanGraphs, since the Second Wild Card era — the past two seasons — the 20 team seasons resulting in a playoff berth have averaged a pre-season projection of about 37 WAR. Fourteen of the 20 were projected to have at least 35, and 19 of the 20 at least 30. Just one team, the 2013 Indians, was projected to have fewer than 30 WAR.

So should we start chilling the champagne? Of course not! The correlation between projected WAR and actual wins is still small and subject to the whims of what makes baseball fun: unpredictability. Or, as Sullivan puts it, "projected WAR + breaks = Actual WAR. Actual WAR + luck = Actual record."

Counting the Rockies’ projected WAR on my fingers spurred a couple of questions in my mind. If we could pick player seasons from the Rockies' history, what combination of players would make the best and worst teams? That should tell us something about exceeding and falling short of expectations, as projections shy away from the extremes. I then wondered about player comparisons. ZiPS provides the top historical comparison for each player. For instance, for the last three years, Troy Tulowitzki’s top comparison has cruelly been Cal Ripken Jr.

Rather than those types of comparisons, however, I want to create a team of comparisons from the Rockies’ past. In other words, if every projection ends up exactly right in terms of WAR, what is the most comparable player season in Rockies history for that position? As Rockies fans, being able to evoke another Rockies player might help us gain a better grasp of what the projection suggests. Does that collection of players resemble anything like a playoff team? The answer, a few paragraphs below, won’t shock you (it is no). I used Baseball Reference’s WAR to create my roster monsters. For brevity, I left off relievers and focused on position players and a five-man rotation.

The best possible Rockies team is a terror from top to bottom. The infield is excellent on offense and defense. Chris Iannetta turned in the best catcher season in Rockies history before being traded after the 2012 season. His 3.2 bWAR rested on the strength of his defense and .372 OBP. Good to great defenders complete the infield, Kaz Matsui’s 2007 and Vinny Castilla’s 1998, and Todd Helton’s 2000 representing the former, and Troy Tulowitzki’s 2007 signifying the latter.

Defense can only go so far. This is the best possible Rockies’ infield because of bats. Helton’s transcendent 2000 is clearly the best, but Tulo’s 2007 isn’t too far behind. You’ll note that the Rockies have never had a truly great second baseman — indeed, the best second base season of all time for the Rockies is more good than great. The Rockies have had several quality third baseman over the years. The best single season was Castilla’s best offensive season, when he finished with a 127 OPS+.

The outfield is fun. The corners have a couple more outstanding seasons, including Larry Walker’s MVP 1997, and an Ellis Burks season that produced a 149 OPS+. And then in center field we have Juan Pierre’s 2001, a year in which Pierre had an 89 OPS+, but a .378 OBP, and led the league with 46 stolen bases.

C: Chris Iannetta 2011, 3.2 rWAR

1B: Todd Helton 2000, 8.9

2B: Kaz Matzui 2007, 3.4

3B: Vinny Castilla 1998, 5.6

SS: Troy Tulowitzki 2007, 6.8

LF: Ellis Burks 1996, 7.9

CF: Juan Pierre 2001, 3.1

RF: Larry Walker 1997, 9.8

The starting rotation for this all-time team is outstanding. The five best pitcher seasons in Rockies history each exceed five bWAR, with Ubaldo Jimenez’s 2010 being by a fair margin the best of the bunch. Four of the five top pitching seasons have come since the installation of the humidor in 2002, the only exception being Pedro Astacio’s 1999.

P1: Ubaldo Jimenez 2010, 7.9

P2: Pedro Astacio 1999, 5.9

P3: Jhoulys Chacin 2013, 5.8

P4: Joe Kennedy 2004, 5.6

P5: Jason Jennings 2006, 5.0

And now let’s turn to the ugly. The worst possible team in Rockies history begins with Kirt Manwaring’s 1997. Manwaring’s wrongness begins with "Kirt," and for the Rockies it culminated in well below replacement level performance powered by a 38 OPS+, the awfulness of which was even mitigated by plus defense.

The rest of the infield hovers around replacement level. Interestingly, the worst second base season in Rockies history is by a very slim margin the best of the worst possible roster in the infield. The Rockies have never had a great second baseman, but it appears that they’ve also never had a horrible one. The same can be said of first base. Ten years after Todd Helton produced the best first base season in Rockies history, he was responsible for the worst one.

The left side of the infield is unbecoming. Neifi Perez played 162 games in 1998 and finished the season a full win below replacement level. Not-third-baseman Jordan Pacheco produced the worst third base season in 2012. Pacheco’s .309 batting average was empty (but sixth in the National League! Drew Goodmen is thinking), and any positive it contributed was eliminated by his defense.

The outfield is fun here, too. Willy Taveras and Brad Hawpe joined forces in 2008 to make a single minor leaguer. More interesting is Dante Bichette’s 1999. He hit .298/.354/.541 with 34 home runs, but Bichette was just two percent better than league average offensively with a 102 OPS+, and his defense was abominable. By Baseball Reference’s reckoning, he was worth negative four wins on defense that year.

C: Kirt Manwaring 1997, -2.5

1B: Todd Helton 2010, 0.2

2B: Mike Lansing 1998, 0.3

3B: Jordan Pacheco 2012, -0.4

SS: Neifi Perez 1998, -1.1

LF: Dante Bichette 1999, -2.3

CF: Willy Taveras 2008, 0.0

RF: Brad Hawpe 2008, 0.0

The worst possible starting rotation is full of terrible seasons, just like the best rotation is full of great ones. There are simply many more pitcher seasons to choose from than, for example, second base seasons. The range here is from -1.5 bWAR to -1.9. And like Todd Helton, Joe Kennedy has the privilege of being on both lists. Any one of these pitching performances would challenge to be the worst in all of baseball in any given year.

SP1: Mike Hampton 2002, -1.9

SP2: Joe Kennedy 2005, -1.7

SP3: Denny Stark 2004, -1.6

SP4: Esmil Rogers 2011, -1.6

SP5: Bryan Rekar 1996, -1.5

No team would ever be projected to be as good or as bad as these roster monsters. The first finished with a bWAR of 78.9, and the second has one of -14.1. In other words, the first team might be expected to finish the season with 129 wins (WAR plus about 50 wins), while the second would languish with about 36. That means the first team would have a shot to be the best regular season team ever, and the second team the worst of all time, and both by fair margins.

It’s a lark of an exercise, but there’s still a lesson. Namely, no two teams have more than two players from a single year. This makes sense, as the marks are necessarily on the extremes. The players cited above either exceeded or fell far short of what any projection system might forecast. It’s easy to buoy optimism by assuming that a projection will be exceeded, but it also might go the other way. Both the very bad and the very good are rare.

Expectedly, ZiPS doesn’t project for rarity. We can dig back into past player performances to find similar types of seasons as the 2015 forecast in order to gain a grasp of what the projections mean.

Standing in for Wilin Rosario, we have 1996 Jeff Reed. Playing first base in place of Justin Morneau is 1995 Andres Galarraga. At the keystone, 2004 Aaron Miles stands in for DJ LeMahieu. Manning the left side of the infield in place of Nolan Arenado and Troy Tulowitzki is 1996 Vinny Castilla and 2013 Troy Tulowitzki.

From left to right in the outfield, 2012 Carlos Gonzalez is our projected 2015 Corey Dickerson, 2010 Dexter Fowler (a switch-hitter) is the platoon of Drew Stubbs and Charlie Blackmon, and in right field 2013 Michael Cuddyer is a hopefully mostly healthy Carlos Gonzalez.

The almost sure to be incorrect projected starting rotation of Tyler Matzek, Jorge De La Rosa, Jordan Lyles, Jhoulys Chacin, and Eddie Butler find their comparisons in 1993 Armando Reynoso, 1994 David Nied, 2002 Denny Stark, 2010 Jorge De La Rosa, and 2000 Kevin Jarvis.

C: Rosario 1.1 ZiPS; Jeff Reed 1996, 1.0

1B: Justin Morneau 1.1: Andres Galarraga 1995, 1.1

2B: DJ LeMahieu 1.2; Aaron Miles 2004, 1.2

3B: Nolan Arenado 3.2; Vinny Castilla 1996, 3.2

SS: Troy Tulowitzki 5.2; Troy Tulowitzki 2013, 5.3

LF: Corey Dickerson 1.7; Carlos Gonzalez 2012, 1.6

CF: Drew Stubbs/Charlie Blackmon 1.8; Dexter Fowler 2010, 1.7

RF: Carlos Gonzalez 1.8; Michael Cuddyer 2013, 1.8

P1: Tyler Matzek 3.1: Armando Reynoso 1993, 3.1

P2: Jorge De La Rosa 1.9; David Nied 1994, 1.9

P3: Jordan Lyles 1.7; Denny Stark 2002, 1.7

P4: Jhoulys Chacin 1.6; Jorge De La Rosa 2010, 1.6

P5: Eddie Butler 0.6; Kevin Jarvis 2000, 0.6

This two-pronged and somewhat circuitous thought experiment leads us back to where we started. The Rockies are on the cusp of playoff contention with about 30 projected WAR heading into the season. If the 2015 iteration of the actual Rockies appears more hopeful than the flawed duplicate of a roster monster I created, it probably has something to do with recency bias. We believe in the possibilities of what Jorge De La Rosa can do while simultaneously marginalizing David Nied as our first-but-not-forgotten regret. But you know what? Maybe De La Rosa — as well as other members of the 2015 team—has a rare season in him and, unpredictably, exceeds the projections.

Luck and the types of breaks that don’t involve bones will have to play a part. We know that. We accept that. And that’s why we watch.