The last few years, we’ve seen the Colorado Rockies do well in the early part of the season only for a summer swoon to ensue. Since I think of warmth when I think of summer, I wondered whether the weather affected baseball in Denver. After quite a bit of research, it turns out the weather does seem to have an effect. Some of the results, such as teams scoring more runs at Coors Field as the weather heats up, were expected. But I came across other differences that I didn’t quite expect. This is the first of a series of articles about weather effects at Coors Field.
This is a popcorn study. Baseball and weather are messy enough separately, much less together. Further complicating analysis is that only game time data is publicly available. Even if we had weather data available to the minute, it’d be very hard to match weather conditions with the moment of a single batted ball. This article will focus on temperature only—we’ll set aside wind, precipitation and other factors. Enjoy your popcorn and this article with a flavorful number of salt grains.
Let’s dive into the data set.* I built a Retrosheet database that contains almost all the box score weather readings at Coors Field from 2007, when Earth Networks installed their anemometer at Coors Field, through the end of the 2016 season.** I discarded October games since there are so few. This gives us a data set of 807 games at Coors Field.
The chart below shows the average temperature and Rockies winning percentage by month and further divided by whether it was a day or a night game.
Through those 807 home games, the Rockies had a .540 winning percentage. One thing that seems fair to say is that, within a given month and over the course of the year, the average game time temperature for day and night games were within a degree or two of each other. With that in mind, I’m comfortable with no longer splitting apart day and night games. You may be comfortable getting another bag of popcorn.
From April to May, the average temperature increases from 58.5 degrees to 64.4 degrees, a six degree change. However, June spikes up to 78.1 degrees, fourteen degrees warmer than an average day in May. The heat peaks in July, but it’s still only a four degree increase from June. August and September are relatively mild with only minor dips in temperature. If we were to loosely group things by temperature, June through September are pretty similar. By those groupings, the Rockies have a .458 winning percentage in April and May. Once the weather heats up, they become a better home team and have a .561 winning percentage for June through September.
The Rockies home winning percentage fluctuates wildly from one month to another for day games; they lose most frequently during May day games with a .412 winning percentage but spike up to a .660 winning percentage in September. By contrast, the Rockies winning percentage at night is pretty stable, hovering around .520 with the noticeable exception of July, when it peaks at a .615 clip. Overall, the Rockies clearly do better at Coors Field from July 1st on, regardless of game time or weather. Temperature during day games probably behave differently than temperature during night games, but at least the game start temperatures are similar.
Because there’s little difference between a single degree plus or minus in temperature, I smoothed out the curves and bucketed the data. This means that in the 65 degree bucket, the data for 64, 65 and 66 degrees is included. For the 66 degree bucket, the data for 65, 66 and 67 is included, and so on. That stops the graph from spiking from a blowout game at a certain temperature and flattens it into something closer to a curve. I’ve also dropped out single games at the extremes for temperature since there were no similar games to bucket it. That way a single game time temperature that had a thirty run game doesn’t affect the graph. That eliminates seven games, giving us 800 Rockies home games.
The graph below shows temperature buckets along the x-axis and average runs scored per game (RPG) along the y-axis. Most of the data is in the 70 to 85 degree range, dropping to as little as four games in a bucket at temperatures lower than 45 degrees. Since there are more games, the data in the middle of the graph is somewhat more reliable than at the extremes.
In terms of the most runs scored per game at Coors Field, we have a spike at 43 degrees that reaches 13.5 RPG, but that’s only over four games. More telling is at 47 degrees where, over 14 games, the average RPG is 13.1. We then have a steep dive in run scoring per game, bottoming out at 8.3 RPG at 51 degrees over 14 games. From that low, the RPG steadily increases as the temperature climbs. For every ten degrees above 50 degrees, expect an extra run to be scored at Coors Field.
Let’s turn to the split of those runs between the Rockies at Coors Field and teams that visit Coors Field. This graph looks a little more noisy, so there are a few more points.
Note that the graph trends upwards, suggesting that both the Rockies and opposing teams tend to score more runs as the temperature increases. Also observe that while the average runs scored by visiting teams increases, it’s not as drastic a change as it is for the Rockies. From 55 degrees to 68 degrees, there are many instances where the Rockies actually get outscored by their opponents. Once the temperature hits 69 though, the Rockies start to party, as they outscore their opponents at virtually every temperature. At 69 degrees, the Rockies score 5.6 RPG while their opponents score 5.2 RPG. Nice.
As the temperature gets warmer than 69 degrees, the gap between the Rockies run scoring and their opponents often favors the Rockies and gets more extreme once it is warmer than 80 degrees.
There’s another curiosity that happens once the temperature hits that 69 degree range. The graph below shows the temperature buckets in terms of home and away runs scored at Coors Field at 69 degrees or greater.
The Rockies trendline, indicated by the blue dotted line, slopes upwards quicker than the away team’s orange-dotted trendline. The Rockies score about half a run more for each ten degrees of temperature, while visiting teams score about a tenth of a run. One explanation might be that the skillsets of Rockies hitters, particularly flyball/home run hitters, can take better advantage of the temperature. It could also be that any Rockies groundball pitchers are less affected by the change in temperature because grounders are less affected than balls in the air.
Nonetheless, there is a difference between how many runs the Rockies score at hotter temperatures than their opponents do. That added run differential translates into the Rockies winning more games in hotter weather. That correlation is evident in the subsequent graph, which shows the Rockies winning percentage by temperature.
When the weather heats up, so does the Rockies win/loss percentage. Between 55 and 68 degrees, though, the Rockies often lose.
That all leaves us with a few takeaways:
- Rockies have a better home record from June to September than from April to May.
- Rockies home record is more stable, month-to-month, during night games.
- For every ten degrees that the game time weather at Coors Field is above 50 degrees, expect an average of one more run scored per game.
- Generally, as the temperature gets warmer, runs at Coors Field increase, but extremely cold environments tend to preclude high scoring games.
- From about 55 degress to about 68 degrees, Rockies opponents tend to score as many runs as the Rockies, and the Rockies tend to lose.
- When it’s 69 degrees or warmer, the Rockies tend to outscore their opponents by a significant margin.
Overall, I’m surprised the Rockies do so well at Coors Field when the weather heats up. Whatever reason there is for the Rockies “summer swoon” myth, it’s not because of the temperature at Coors Field.
So, why is there a dip in the Rockies winning percentage in that temperature range of the mid fifties to high sixties? Also, at the lower end of the temperature scale, did you wonder why run scoring is so high when the weather is cold out? Why do the Rockies outscore their opponents so much in hotter weather? To answer these questions, we need to look at another aspect of weather. Next time, we’ll discuss how wind might affect scoring at Coors Field. Have your popcorn, as well as your salt, ready for that one.
Part two of this series addresses the effects of wind.
*The information used here was obtained free of charge from and is copyrighted by Retrosheet. Interested parties may contact Retrosheet at 20 Sunset Rd., Newark, DE 19711.
**Denver weather readings and forecasts on the news and radio often come from Denver International Airport which is over twenty miles away from Coors Field. The temperature displayed on the scoreboard is also from DIA. As meteorologist Chris Spears (@ChrisCBS4) from CBS4Denver told me, “The weather is different from Colfax to Blake Street, but when you get down to the nitty gritty unless a cold front or some large scale feature is moving through and stalls out, the weather is "similar" across the whole metro area.” The best fit for real-time regularly updating data is actually the weather station at the Denver Museum of Nature and Science located about three miles away from Coors Field, though it is on the other side of a bunch of wind-obstructing skyscrapers.