Auditing the 2010 Colorado Rockies: Home/Road Splits

Going into the 2010 season, Rockies fans were expecting a lot out of this team. On paper, they were the most talented team in the NL West, they had no obvious holes and boasted superior depth. Several pundits even picked for them to win the World Series before the season. I certainly bought in to Colorado's chances to win the West, going so far as to guarantee a title for them. At the very least, I said, the Rockies should beat the preseason O/U win line set by Vegas at 84. Obviously, none of those things occurred. I've taken it upon myself to understand why.

While there have certainly been some great memories to take away from the 2010 season (thanks PIP), this 83-79 season by and large has to be looked at as a failure from a baseball standpoint (though as I'll touch on in the future, it was a successful season financially) given the massive expectations entering 2010.

In future installments, I'll go position by position and look at the expectations going into the season for that spot. Then I will look at the results, trying to analyze where the disconnect between the performance of those positions and their expectations coming in stemmed from. Finally, I will briefly touch on some potential solutions for those issues and what the future of each position looks like. I will repeat this process for the pitching staff  in a different installment, then I will evaluate the performance of the front office and coaching staff based on my findings.

But before I do that, I feel I need to discuss a prominent topic that most certainly shaped the poor outcome of the 2010 season, the Rockies' poor performance on the road. My methods when evaluating performance are primarily based in statistics (the majority of which can be easily referenced at Fangraphs or Baseball Reference), but I've done my best to take potential psychological effects into account. In some cases, it's difficult to separate psychology from physics. Nowhere is this effect more stark than in the home/road splits of the Rockies.

I'll get this out of the way right now: nobody can explain why the Rockies' road troubles have reached the extent they have. Due to this fact, the purpose of this article is not to answer that question, but merely to provide some discussion points. There's a lot of great research out there (notably this Fanpost at SBN blog Athletics Nation and this ESPN Insider piece by Ben Jedlovec of Baseball Information Solutions) that helps explain why the Rockies' hitters are so effective at Coors Field.

Unfortunately, nobody has yet been able to discover why the Rockies (through September 18th, before Operation Shutdown commenced) had scored 6.35 runs per 27 outs at home but only a pathetic 3.46 runs per 27 outs on the road (from the ESPN article). Colorado's pitching splits make sense: the Rockies allowed 4.51 R/27 at home and a slightly smaller 4.22 R/27 on the road. It's strange that the offense is so much more affected. Furthermore, it's strange that this year Colorado's H/R offensive difference is much more pronounced than it was in the two previous years (1.84 times more runs at home in 2010 vs. 1.46 times more runs in 2009 and only 1.30 times more runs in 2008) while the pitching difference remained fairly constant (1.07 in 2010, 1.07 in 2009, 0.98 in 2008).

I wish I could tell you why Colorado hit only .226 on the road vs. .298 at home (.303 vs. .368 OBP, .351 vs. .498 SLG, etc). Basically, it was like the Rockies were Willy Taveras on the road and good Brad Hawpe at home. Some of that can be accounted for by the fact that Coors is such a great place to hit in and that the Rockies have really learned to maximize that advantage. Some of it can't. For example, the Rockies' sOPS+ at home was 130 while on the road it was 85. Why is this significant? In order to understand that, it's important to understand what sOPS+ is.

Most of you know OPS (on base % plus slugging %). sOPS+ differs from standard OPS in two ways. First of all, the + added to the end of the statistic means that the statistic is adjusted to include the Baseball-Reference three year home park factor, which for Coors Field is 115 for both pitchers and hitters (explained here) and is normalized on a scale where a score of 100 is league average.

Troy Tulowitzki had a 138 OPS+ this year while Clint Barmes had a 67 OPS+. This tells us that Tulowitzki was 38% better than a league average hitter (and quite a bit better than the average shortstop) while Barmes was 33% worse than league average. In other words, OPS+ is a great stat for comparing how the Rockies would do in a neutral context. 

What is even better is taking into account the player's position when evaluating a player, because obviously Tulowitzki is providing premium offensive production at a position where average offensive output (as measured by OPS+) is lower than 100, meaning that when compared to only shortstops, Tulowitzki's offense is much more impressive while Barmes' poor offense is more acceptable when compared to only second basemen. I'll go into that more next week when delving into each position, but for now it's fine to look at OPS+ on a team level.

The second difference of sOPS+ from OPS is the s, which stands for split (any split, like for a lefty hitter facing a lefty pitcher). sOPS+ tells us how a performance is relative to the league average for that split. For example, with sOPS+ a lefty facing a lefty will be measured only against other lefties.

To answer why sOPS+ is so significant to Colorado's H/R woes: sOPS+ shows that, even after adjusting for park factors and also in comparison to all hitters' home and road performance, the Rockies are 30% better than an average team at home (130 sOPS+) and 15% worse on the road (85 sOPS+). This isn't a new problem either: this Fanpost by rururuland84 from late August shows that only once in team history have the Rockies even been an above average offensive road team relative to MLB (1997, when they had a 104 sOPS+) but have been above 100 every year at home, again even after park factors are neutralized.

This year's 130 sOPS+ was the highest since the implementation of the humidor in 2003, but the team's 85 sOPS+ is worse than all but the woeful 2005 club's 82 sOPS+ in the Humidor Era. Carlos Gonzalez had an incredible sOPS+ of 204 at home (104% better than average and 74% better than his teammates!) while he only posted a 118 sOPS+ on the road (only 33% better than his teammates). Obviously the Rockies are maximizing their home-field advantage, but something is keeping the Rockies, despite their high talent level, from being an above-average offensive team on the road.

It's a big problem, and I'm sure it results from a number of factors beyond anyone's control. Obviously, the ball moves differently on the road than at home and doesn't travel as far, but no single factor explains this phenomenon. It's something that needs to be taken into account when discussing every Rockies hitter--by and large they're not as good as their home offensive numbers seem to indicate nor are they as poor as their road numbers. It will affect debates for All-Star selections, MVPs, and the Hall of Fame.

I just wish somebody could explain it to me, because if there is something Colorado could do to improve its performance on the road, they'd be perennial playoff contenders with their home-field advantage (due to their hitters' ability to maximize the benefits of the home park and their pitchers' ability to limit its advantages for opoosing hitters, it is the best in MLB). 

More research into this phenomenon is needed, because more than almost any other factor (I'll get to that later), the Rockies' putrid road performance this year was the cause of their failure to live up to expectations.