When Do We Play?
There’s an odd problem with our baseball statistical record. The problem is that nobody has kept good records of when games start.
We have some records, but what we have seems spotty. I can tell you, for example, that this random Chicago Cubs game from 1960 started at 1:30 PM:
However, I can’t tell you anything about what time this random 1950 Cubs game started:
And, yes, this is a problem.
Who Cares?
Who cares, right? This is a bunch of trivia and gibberish, right?
Not really. This matters for 3 basic reasons.
The time a game starts influences how well batters can see the ball. This is obvious if you’ve ever watched a late afternoon or twilight game on television. Most broadcasters to this day will comment on how hard it is to see the ball when half the stadium is in the shade.
The pace of play was directly influenced by the time the game started, especially in the days before stadiums had lights. This is because the umpire had a strong incentive to get the game completed before darkness fell.
The time games started in the days before lights was much later than you think.
Let’s look at a simple example. We’re going to look at the New York Giants, starting in 1887 and moving up to 1957. We’ll jump forward 10 years at a time, and will choose a random weekday home game in July. Let’s see if we can figure out when each game started!
Please note that none of this information is easily accessible on Baseball Reference. All of this comes from searching through newspapers on newspapers.com. Naturally, this research is possible because teams advertised their games in the local newspapers.
1887
We’ll look at Tuesday, July 12, 1887. The Giants hosted Pittsburgh that day, losing 15-6.
The Sun on that day printed only 4 pages. However, I was still able to easily locate an advertisement for that day’s game:
The game started at 4 PM.
That’s right: 4 PM. Surprising, isn’t it? We don’t see many games start at 4 PM these days.
1897
Now we hop up a decade.
The game we’ll look at now is the July 30, 1897 game between the Giants and the Brooklyn Bridegrooms. I had to choose this date because the Giants played on the road almost every single day during that month.
The Sun had expanded to 8 pages by 1897. However, some things didn’t change:
That’s right — this game started at 4 PM as well.
1907
Up to 1907.
This time we’ll look at the July 8, 1907 game between the Cardinals and Giants.
Once again, the Giants played at 4 PM. Interestingly enough, Brooklyn started its game at 4 PM that afternoon, too. And, if you were curious, the Highlanders also usually started their games at 4 PM on weekdays throughout this era.
1917
Let’s look at the Tuesday, July 17, 1917 game between the Cubs and Giants.
I had to switch to the New York Tribune for this one, since The Sun was incomplete:
By 1917, the starting time had gone back to 3:30 PM, which still seems a bit late by modern standards.
1927
This time I’ve chosen the July 19, 1927 game between the Cubs and Giants — also played on a Tuesday:
I’ve had to switch to the Daily News here since the other New York newspapers dry up after 1923 or so (well, except for The New York Times, but ProQuest is frustrating to use).
Anyway, you can see that single games still started at 3:30, and that single entry doubleheaders started at 1:30 and were advertised in advance. In contrast, doubleheaders requiring separate entry fees usually had a morning and an afternoon game.
1937
Now we’ll look at the Wednesday, July 14, 1937 game between the Pirates and Giants:
This is an interesting one:
The time of game moved up ever so slightly, standing now at 3:15 PM — 45 minutes earlier than it had been 50 years before.
1947
The Cubs visited the Giants on July 16, 1947.
Things had changed by this time. Night baseball came to the Polo Grounds on May 24, 1940. Among other things, this appears to have pushed up the standard starting time for day games:
That’s right: the game now started at 2:30 PM. I’m not certain what the reason for the change was. Further research is necessary.
1957
Finally, we’ll take a look at a random game in July 1957, the last year the Giants played in New York.
This time, we’ll look at this July 11, 1957 Cardinals - Giants game.
As you can see, the time of game was pushed all the way up to 1:30 PM by 1957, which is a lot more in line with what we see these days.
Why?
But why the change?
Honestly, I don’t know. It will take some more investigation to figure out what is going on here.
I know from my preliminary research that this is not unique to New York. The starting times in other cities follow similar patterns, particularly after World War II. My guess is that there was probably a series of league decisions that caused this shift. It might have something to do with wartime curfew and blackout rules — or it might have something to do with trying to avoid the costs of unexpectedly turning on the lights as much as possible.
I’ll present more of those findings as this blog moves on.
Implications
There are a few notable implications of all this that historians, sabermetricans, and random internet commentators all seem to be unaware of:
The poor hitting of the deadball era may have been at least somewhat attributable to games being played so late in the afternoon.
Similarly, it is possible that pitchers in the early days could throw more complete games because batters near the end of the game had a very hard time seeing the ball.
Teams during the era before lights were not stupid, and tried to time the start of games to allow businessmen and adults who could pay to attend.
For some strange reason, teams decided to start day games earlier after lights were installed — maybe to reduce the chance that the lights might need to be turned on?
Games were likely played so quickly in the old days out of necessity — and umpires were likely under a lot of pressure to keep the game moving and avoid the hassles of darkness.
The idea that baseball teams were run as some sort of civic charity in front of mostly empty stands back in the “good old days” is completely false. Teams wanted to attract paying customers, not hoodlum kids who would sneak in.
It is likely not good enough for us to separate statistical splits into “day” and “night” alone. I’m reasonably confident that we would see different offensive statistics if we separated the 4 PM starts from the 2 PM starts and the 11 AM starts.
More to come!
It would be interesting to see a stat comparing the effect on the performances from different start times. Of slight interest is the editorial by Henry Chadwick published in the 08 Spaulding guide where he takes note of a movement in England where they set the clock 20 min forward on each of the 4 Sundays in April to gain more daylight hours...I kind of like that idea to daylight savings instead of what we in the U.S. eventually settled on. Anyway, enjoy your thoughts...If you were to do a statistical look I think you would have to find the time of sunset for exactly where the game was played as there was no uniform daylight savings time across the nation and states and localities including cities were free to adopt it or not ...don't think it was until Nixons term that it became more national...keep up the good work and roll on...