DENVER -- Twenty years of data screamed to Rockies general manager Dan O'Dowd that pitching at an altitude of 5,183 feet was different from pitching anywhere else in the Majors, especially for starters.
Until this year, O'Dowd treated the calls for change as background noise. The pile of stats and injuries grew under the watch of a general manager before him, four managers before current skipper Jim Tracy and five pitching coaches before Bob Apodaca -- by far the most successful -- grew tired and left his post on June 26.
Many trends worsened. Starting pitchers' stats threatened all-time futility records. Coors Field, normalized somewhat in 2002, when the club began storing baseballs in a humidor, turned on pitchers at rates not seen in a decade. Even the booing and public calls for his job couldn't drown out the numbers-based message O'Dowd was hearing.
The common practice of having five starters take turns trying to throw 100-120 pitches a game -- the norm in the Majors for more than four decades -- wasn't working here. So O'Dowd has trashed it.
"If we can be blamed for anything from a front-office standpoint, and I know we're being blamed for a lot, we should be blamed for the fact we didn't attempt to do something like this a lot sooner," O'Dowd said.
In June, the Rockies hatched what O'Dowd calls the "paired pitching plan," with a rotation of four starters under a stricter pitch limit, plus three other pitchers who take turns as a "piggy-back" to the starter. The starter's pitch limit started at 75, with the companion pitcher throwing up to 50, but there is flexibility in the starter's allotment.
The goals of receiving better outings from starters and reducing injuries are based on trying to stem the franchise's long-term issue of inconsistency. The Rockies have succeeded on occasion. But in other years, rarely are they close enough to contention to make a shrewd move, get lucky and slip into the postseason.
2012 Rockies starters
"In reality, every 10 years, a payroll model of our organization gets to the playoffs roughly 2.7 times," said O'Dowd, who has made two postseason trips in 13 years. "Our peaks and valleys have just been excruciating.
"If you really dig in and study the history of the franchise, you're going to see that so many things are reflective of struggles over 20 years as relates to starting pitching."
The paired pitching plan hasn't slowed this year's losing, and the club is on pace for its first 100-loss season. Going into Tuesday's game against the Cardinals, the Rockies -- 37-63, in last place in the National League West -- have a lower winning percentage under the new system (.343) than the old (.385).
But this year ceased to be the priority once the Rockies turned the season over to research and development and a plan that hadn't been tried. Four-man rotations of long ago were not pitch-count oriented. If anything, the Rockies' priority was finding out what could go wrong. There was plenty.
The Rockies started the experiment at the beginning of a stretch of 20 games in 20 days -- timing that left them with no choice but to work through problems.
"In April, you normally have a lot of off-days, and in September, you have a lot of guys," O'Dowd said. "So you're going to be managing the months of June, but more importantly July and August, with workload."
2012 Rockies starters
Walks per nine innings
Under five-man rotation
Under new model
Also, in the beginning, those piggy-back innings were being handled by guys who had pitched their way out of the rotation, meaning the worst performers on staff were handling middle innings of winnable games. With planned-upon starters Jorge De La Rosa, Jhoulys Chacin and Juan Nicasio missing most or all of this season with injuries, and veteran acquisition Jeremy Guthrie having struggled before being traded to Kansas City, the starting staff was destined to struggle no matter the system.
Tracy worried at first about the plan forcing him into double-switches early in games and hamstringing his ability to match hitters with pitchers later. But he has seen some success, especially with veteran lefty Jeff Francis and at times with rookie Christian Friedrich.
The Rockies see some stats as pointing toward efficiency and injury prevention -- two massive problems this year and over the history of the franchise. Walks per nine innings are down, from 3.76 under the five-man rotation to 3.17 under the new system. Pitches per innings also are down, from 17.6 to 16.6. The Rockies also have a winning record when their starter pitches at least five innings and gives up five or fewer runs, and when the starter goes just four and yields four or fewer. They are one game under .500 when the starter goes four and gives up five runs. That's hardly dominating, but it is competitive with a decent offense.
"How much better can it get?" Tracy said. "The only way you're going to find out is if you keep venturing into the water."
Another adjustment is that Rockies pitchers have reduced throwing between starts, which was estimated at 200 throws of soft-toss, bullpen and flat-ground work.
2012 Rockies starters
Pitches per inning
Under five-man rotation
Under new model
The only health issue occurred with rookie lefty Drew Pomeranz, who had his latest start moved back because of left arm soreness -- but Pomeranz hasn't pitched on three days' rest. Francis, who excelled under the previous system, only to become one of the club's many injury casualties (left shoulder problems marred his 2008 season and cost him all of '09), has adjusted best. The team is 8-2 in games Francis has started since he rejoined the Rockies in June.
"I feel the same," Francis said. "I've been happy with my recovery. I'll just continue to cut down on throwing in between, trying to do my best to stay sharp."
Pitchers could end up throwing more in game action, making as many as 40 starts per season as opposed to 31-32 under a five-man system. The Rockies' reduction of work between starts is only part of the strategy for reducing fatigue.
The Rockies believe that stressful innings lead to injury. One notable stat is a pitcher's success when facing a batting order for a third time. The Major League batting average of opponents against starters rises from .250 the first time through the order to .266 the second time, then to .277 the third. At Coors Field, those numbers are .273, .285 and .303.
Two games last week illustrate the Rockies' strategy.
The club called up lefty Edwar Cabrera from Triple-A Colorado Springs to start on Tuesday against the D-backs. Cabrera reached 81 pitches with the lineup coming up for a third time with one out in the fourth.
Tracy was left with a choice. He could hope that Cabrera, having thrown just 45 strikes, would defy the stats, or he could turn to Josh Roenicke, who was fresh and has been the team's best reliever in this role. Tracy went with Roenicke, who gave up two runs en route to a 6-2 loss.
When starter goes at least ...
2012 record (through July 18)
Five innings and allows three or fewer runs
Four innings and allows three or fewer runs
Five innings and allows four or fewer runs
Four innings and allows four or fewer runs
Five innings and allows five or fewer runs
Four innings and allows five or fewer runs
In Francis' start on Wednesday, by the time the order came up a third time, the left-hander was one out into the fifth at 76 pitches -- a full inning deeper and five pitches fewer than Cabrera the previous night -- and not once did he throw more than 20 pitches in an inning.
Francis would pitch until yielding Miguel Montero's homer to open the seventh, at which point he was replaced by Adam Ottavino. The righty successfully preserved the Rockies' 4-2 lead.
"What we're beginning to find out is that when you do it correctly and you do it efficiently, more times than not, the guy that goes out there to start the game takes you to the place where you want to be," Tracy said.
Said Francis: "I think we've done better -- we've executed better as a starting staff. We haven't won as much as we'd like to. We've made some adjustments. We still have a lot to do and a lot to prove. We still think we can do a lot better. We're going along with it and embracing it."
Why would the Rockies make such a radical adjustment just five seasons after a World Series appearance and three after the best pitching year in club history?
Just once since the expansion season of 1993 have Rockies pitchers posted an ERA lower than the NL average, and just twice have they held opposing hitters to a lower batting average than the NL norm. In all but three seasons, the club fell below the league average in walks and hits per innings pitched, strikeouts per nine innings and opponents' on-base plus slugging percentage.
Also, injuries and performance declines have been predictable. According to the Rockies:
The club has had 49 pitchers throw 150 or more innings in a season. The season after, 27 have spent at least 15 days on the disabled list -- 55.1 percent, compared to the Major League average of 35.7 percent. That included 13 Rockies who spent at least 60 days on the DL.
"We've made some adjustments. We still have a lot to do and a lot to prove. We still think we can do a lot better. We're going along with it and embracing it."
-- Rockies lefty Jeff Francis
The Rockies have had 13 pitchers post back-to-back 150-inning seasons, and 10 of them spent at least 15 days on the DL the third year -- a 76.9 percent rate, compared to the Major League average of 28.6. Four of the Rockies' 13 such pitchers spent 60 or more days on the DL.
Staying effective over time, even without injury, has proven difficult. The only pitchers in club history to throw three or more consecutive years of 150 or more innings and achieve an ERA below 4.75 are Aaron Cook (2006-09) and Ubaldo Jimenez (2008-10). The Major League per-team average over that period is eight, and in the NL West, the average is 10.
In 2009, the Rockies' best season for pitching, five pitchers exceeded 150 innings as the team went to the playoffs. Jimenez (218 innings in '09) had an incredible first half in '10 but has declined precipitously since. Jason Marquis (216), De La Rosa (185) and Cook (158) spent time on the DL the following year with pitching-related injuries. De La Rosa had a left finger injury in 2010, blew out his left elbow last year and is hoping to pitch again before this season is done. Jason Hammel (176 2/3), now with the Orioles, has inconsistent statistics but at least has kept his arm healthy.
The nagging question is whether the problem is strictly in drafting, developing and acquiring the right pitchers. But according to the Rockies' research, some of the best pitchers of the last two decades have elevated numbers at Coors Field.
Among active pitchers, 2011 NL Cy Young Award winner Clayton Kershaw of the Dodgers has a Coors Field ERA 3.02 runs above his career figure, Giants two-time NL Cy Young Award winner Tim Lincecum is 0.84 above his career mark, and Lincecum's teammate Matt Cain has an ERA 0.81 runs higher in Colorado. Retired pitchers, many of whom had significant pre-humidor innings, didn't fare much better. Curt Schilling's Colorado ERA was 2.45 runs above his career mark, Greg Maddux was 2.11 higher and Randy Johnson was 0.77 higher. Tom Glavine, arguably the best at Coors Field among great pitchers of the past two decades, was 0.23 runs higher.
"I believe it's going to work. But I know that we're at a point in time we've had so many peaks and valleys, until we can come up with a sustained formula for success as relates to the venue we play in, we're going to continue to have years that are completely inconsistent."
-- Rockies GM Dan O'Dowd
Would any of those pitchers have figured out Coors Field based on their pedigree, or would they have suffered the same pattern of injuries and declining performance? The change in plans illustrates the Rockies' theory.
But the system is not set in stone.
Rockies co-pitching coach Bo McLaughlin -- who runs the starters, while Jim Wright runs the bullpen -- has experienced a creative, flexible system. A first-round pick of the Astros in 1975, McLaughlin found himself at Triple-A Memphis in 1976 -- his first full pro year.
"I was in a five-man rotation, but I pitched every fourth day and everyone else pitched around me," said McLaughlin, who threw 147 innings at Memphis, then 79 innings in the Majors that year. "If we find one guy that really dominates and can pitch every fourth day, why not pitch him every fourth day?"
There are challenges, such as developing talent and identifying trade and free-agency targets who fit. But the Rockies -- believing they have a better chance at finding four starters than five, and that they can break in young starters as piggy-back guys -- are committed to finding something other than the old way.
"Is this going to work?" O'Dowd said. "I don't know. I believe it's going to work. But I know that we're at a point in time we've had so many peaks and valleys, until we can come up with a sustained formula for success as relates to the venue we play in, we're going to continue to have years that are completely inconsistent.
"So instead of complaining about it, we have to become an organization of adaptability. The people in uniform and the people in the front office have to embrace the fact that we are going to have to constantly adapt to the environmental, climatic conditions that exist in playing at Coors Field."