Regarding the Colorado Rockies road performance, the saying generally goes “Eh, they don’t hit.” Ironically, the pitching performs about the same at home and on the road when adjusted for park factors. The offense, however, does not seem might not share that same kind of fortune. Matthew Gross does a fine job of outlining the differences in home/road splits in terms of wRC+, wondering if the Rockies are at a perpetual competitive disadvantage or is the wRC+ formula itself an issue. In contrast, our own Eric Garcia McKinley used wOBA which suggested that home-road differences had more to do with the quality of the players on the team than the altitude. So is it altitude or talent?
There are multiple theories as to how altitude could affect the offense. One has to do with a hangover effect as the Rockies leave altitude which affects the hitting as Rockies hitters are unaccustomed to curveballs and other slow pitches with movement on the road that the altitude flattens out at home. Others, as far back as Jeff Cirillo, suggest that it’s not that the Rockies do poorly on the road but that they do so well at Coors Field that it makes the road offense look bad. Another possibility, specifically with 2014, is that the Rockies contact-oriented approach “rewards” singles at Coors Field by turning them into doubles while turning them into dying quails on the road. There’s
Then, there is the counterargument, expressed by Jonah Keri and at the second Rockies Bloggers Panel, that talent plays anywhere and the Rockies just don’t field talented teams. Just in 2014 they had to cross their fingers on Justin Morneau and acquire Brandon Barnes and Drew Stubbs instead of having an internal solution that they were confident about. Compounding this problem is that the Rockies Single-A and Triple-A teams, similar to the major league team, play as hitters parks. So, the true talent of Rockies hitters are continuously masked, looking better than they are, until they have to play a game in San Francisco.
So, there are multiple possibilities and it is quite likely many of the above reasons contribute. Meanwhile, as Eric noted, random variation can muddy the waters. Thus, the question remains. And since I like questions, I decided to take a stab at it.
For the record, I’m not a statistician. I know enough about the numbers to form a question, do a little research, then throw my stuff out to the masses hoping someone with true analytical chops can take it to the next level to either prove or debunk. My question was this “If altitude affects pitches for whatever reason, would we see a difference in contact rates at Coors Field and away from Coors Field?” Then, since I’m not a statistician, I decided to tap with some of the tools provided by ESPN’s Stats and Info Department. The neat thing about this tool is there are a tons of clicky things to click. Want to look at heat maps on 3-0 counts? We got that. Want to figure out how well lefthanded hitters on the Oakland Athletics do with runners in scoring position against pitches on the outside corner? They got that too. The data itself goes back to 2009 which, in terms of my question, happily keeps some of the early pre-humidor quirks out of the equation.
Now, I do know enough about statistics to note that there is some selection bias. After all, we are looking at Rockies hitters since 2009 so people who have gathered a lot of at bats such as Troy Tulowitzki will factor more into the data than, say, Justin Morneau. Seth Smith with 1169 at bats matters more than Drew Stubbs and his 140 at bats. You get the point. I also included pitchers, partially because I wasn’t sure how to remove them and partially because if they keep swinging and missing at home, they should keep swinging and missing on the road too. Selection bias is important to at least acknowledge because there is always the chance that someone like Carlos Gonzalez gets a tummy ache unless he’s playing at altitude and that tummy ache affects his swing. Considering the rash of weird injuries the Rockies have had, I wouldn’t be completely surprised if a Rockies regular had a quirk like that. There’s also the slight chance that pitches in the strike zone are calibrated slightly differently between Coors Field and other parks, so I’ll try to stay in the strike zone as much as possible in this discussion.
So, first we are going to look at the contact rate. This is merely the ability to put a bat on a ball, even if the ball goes foul. Going back to my question, if altitude does affect pitches, we should see a difference in contact rates.
We see a difference. Even for pitches right down the middle of the plate, the contact rate drops from 89% at home to 86% on the road. Each segment of the strike zone, in fact, has a worse contact rate on the road by at least a few percentage points with the sole exception of the area that is low and away to right handed hitters, also known as low and in to Todd Helton and Carlos Gonzalez. Interestingly, pitches that are low in the zone tend to have similar contact rates whether the Rockies hitter is at home or away. Furthermore, Rockies hitters tend to make contact better on pitches beneath the strike zone on the road then they do at home. Also, as a general statement, it seems Rockies hitters tend to do worse on the road on pitches belt high and above. We’ll get back to this later.
Do we care about swing rate? I would think that whether players swing at pitches more or less at home than on the road, they should still make contact at a similar rate. Nonetheless, we’ll take a look.
If anything, Rockies players tend to swing at more pitches at home with the sole exception of pitches that tease Charlie Blackmon‘s beard or Troy Tulowitzki’s chin. Going back to the contact chart, though Rockies tend to make contacts on pitches below the belt on the road better than they do at home, they also swing less at pitches below the belt on the road.
Overall though, the decreased contact rate on the road does not appear to be caused by swinging at more pitches on the road. In other words, the Rockies hitters are less aggressive on pitches away from Coors Field even if those pitches are on the strike zone. More on that one later too.
Ok, so to briefly recap, it looks like Rockies hitters are better at hitting pitches below the belt when away from Coors Field and worse at hitting pitches above the belt when away from Coors. In general, their overall contact rate is down. Earlier, we talked a bit about the theory that altitude may be flattening pitches out and, conversely, pitches move more away from Coors Field which would cause our beloved brethren to swing and miss more.
From here, we go to pitch type and get clicky. First, we’ll talk a bit about those curveballs and the general category of “Soft” pitches which generally average around 82 mph.
Remember that part where we said we’d get back to that later? Oh wait I said it twice. Well, I was referring to the first time. Anyway, looking at the soft chart we see… gosh… looks like the Rockies do about the same away from Coors Field on slow pitches as they do at home. And if the pitch drops below the strike zone they make contact more often than they do at home.
Let’s flip to curveballs for a second, where flattened pitches go to die in the bleachers…
Um… well… this is counterintuitive from the given wisdom that Rockies hitters are more susceptible to curveballs on the road. And this, dear reader, is where we progress from contact rate and go to straight batting average which gives us some indication of the quality of contact.
Turns out that though the Rockies do make contact on those soft pitches away from Coors Field, they don’t put the ball in play well. On “Soft” pitches, the Rockies batting average dips from .268 at home to .219 on the road. Look at how that red part in the “Home” heat map gets so blue once the Rockies are on the road. Curveballs in particular, dip from a .265 batting average at home to .204 on the road.
So why did we go through all those exercises using contact rates when we could’ve just used batting average to show the poor performance on the road? The point was to show that while the Rockies were not getting base hits off slow pitches, they were at least making a contact at a similar rate as they were at home. It turns out, though, that the Rockies aren’t just getting burned on slow pitches…
The Rockies batting average on hard pitches, which are pitches that average 91 miles per hour, is .312 at home and .252 away. However, unlike soft pitches which have similar contact rates at home and away, the Rockies are definitely swinging and completely missing on the road more. Remember that second part where I said “More on that later”? That’s where this comes into play where the higher a hard pitch gets, especially belt-high and above, the less likely a Rockies hitter is able to make contact.
And when Rockies hitters do make contact on a fastball, they are less likely to get a hit on the road. The difference between batting average for each area of the strike zone between home and away is dramatic, even if that pitch is over the heart of the plate.
The ironic part is that, in hindsight, the inability for Rockies hitters to hit fastballs on the road does a good job at explaining why the Rockies perform so poorly away from Coors. After all, not every pitcher throws a curveball and many relievers don’t throw a “soft” pitch. However, many opposing pitchers have fastballs. The key takeaway is that fastballs away from Coors Field are devastating to Rockies hitters.
The following charts help us take a look at just how horrid the Rockies do against fastballs on the road.
The x-axis shows the velocity of the pitches from 84 mph to 98 mph and the y-axis shows the contact percentage (blue line), in-play batting average (green line) and the line drive rate (aquaish-kinda line). We will use in play batting average, which is the hits divided by the number of balls in play as an indicator of “decent” contact. We will use line drive rate as a metric for “solid” contact though the numbers are also similar for baseball card stats like Slugging Percentage (SLG) and more obscure stuff like Isolated Slugging Percentage (ISO).
For simplicity sake, look for 90 mph on the x-axis. In other words, a standard major league fastball. Also, it’s easier to peek at since it’s on the “90″ line. On the graph on the left, representing the Rockies hitters performance at home, you can see that Rockies hitters make contact at around 88%, an in-play batting average of .460 and hit a line drive at a 23% of the time. On the road, Rockies hitters only hit that same generic 90 mph fastball 82% of the time, an in-play batting average of .400 and hit a line drive a mere 18% of the time. Or, as another way to phrase the numbers, the Rockies hitters on the road make contact on fastballs 8% less, get a hit on a ball in play 12% less and make solid contact 20% less often. Even from a general perspective where you compare the lines on the graph, the home graph’s lines on the left are higher than the away graph’s lines on the right.
Let’s discuss talent for a second. I got clicky but rather than spam you with more graphs, we’ll just look at Tulowitzki who shows a similar discrepancy in home and road performance on fastballs as Helton and Gonzalez.
There are fewer pitches being tracked when looking at an individual player instead of a whole team so the lines get more jiggy with it. Nonetheless, the story is similar in that Tulowitzki performs worse on the road from a contact, in play and line drive perspective. If talent was the sole reason for why the Rockies are bad on the road, I would expect talented hitters like Tulowitzki, Helton and Gonzalez would be the best at compensating for that. However, back to our generic 90 mph fastball at Coors, Tulowitzki has a contact rate of 92%, an in-play batting average of .390 and a line drive rate of 21%. On the road his contact rate dips to 90%, in-play average to .340 and his line drive rate to 14%. Now, as I said, the lines are pretty wobbly with only 84 pitches thrown at 90 mph on the road to Tulowitzki. If you look at 89 or 91 mph instead, his line drive rate is around 19%. In any event, it is clear when comparing the lines from the two graphs, just as we did with the Rockies home and away graphs, that there is some dip in performance on the road. Absent of some particular tummy ache or other explanation on why Tulowitzki is worse on the road, if talent is a factor in home versus road performance, than Tulowitzki’s talent doesn’t seem to be counteracting whatever it is that fastballs are doing on the road. A less talented hitter would fare even worse.
The next question is “Why do the Rockies perform worse on fastballs on the road?” I can only hypothesize that, while altitude may flatten out the movement on soft pitches at home, it may also flatten out fastballs at home. But that also, in my mind at least, implies that an average major league fastball on the road has a ton of movement. In other words, we would expect most pitchers with flat fastballs in Coors Field, like a Wilton Lopez, to throw pitches with Greg Maddux-esque movement and sink on the road in order to account for such a huge difference. While we might expect something like that for a generic curveball since curveballs by their very nature are supposed to move, I’m not sure I can make that kind of claim that most major league pitchers fastballs move extremely away from Coors and flatten out extremely at Coors.
There may be other issues of altitude to consider including humidity and general brightness of the sun or even the supposedly soggy balls at Coors which makes it easier to hit at Coors and makes it difficult for players to adjust away from Colorado. That’s just grasping at straws though.
I fear this kind of research is left to better minds than mine. However, the inability to hit fastballs on the road may just be one more reason why the Rockies don’t hit on the road. Perhaps that concept will give someone somewhere a place to start looking and maybe even offer up a solution. If so, I look forward to reading it.