Posted by Andy on February 2, 2010
Yesterday I wrote about winning 1-0 and some of the limits in the Play Index in searching for such games (such as the 1954 limit.) Loyal reader DavidRF pointed out the Situational Reports tool on the site, something I am embarrassed to admit I never knew about before.
What a cool toy! The first thing I did was figure out how the margin of victory has broken down since 1900. (A quick reminder that below where I talk about 1-run margins of victory, I'm not referring to only 1-0 games like yesterday but rather all games with a 1-run differential.)
Here are the percentages of games won by a particular number of runs since 1900:
As you can see, for most of baseball history, 1-run victories have been the most common. That's not too surprising--given that both teams start with zero runs and one run is the smallest margin that can decide a game, it makes sense that it is the most common outcome. I'm curious to see what type of 1-run margin is the most common but I'd assume it's probably 2-1 or 3-2. (I'll post that tomorrow, OK?) There have been just a few times when all victories of 5 runs or more were more frequent than 1-run wins: in the early 1900s, some of the 1930s, 1948, and, ho hum, the Steroids Era. I think we're beating a dead horse on that one, but suffice it to say--scoring lots of runs leads to big wins sometimes.
In some periods, run-scoring was so low that 2-run victories actually became more frequent than all wins of 5 or more runs. That happened in 1917 and 1968. The low overall scoring in 1968 caused the mound to be lowered, giving batters a little assistance against pitchers.
I find it interesting how consistent 2, 3, and 4-run victories have been. For the last 100+years, you can bank on the fact that 4-run victories would comprise 11 percent of all games, plus or minus a small fraction. Three-run games have been nailed at 15% and 2-run games have been fairly consistent around 18%. I did not show the broken-out data above, but victories of exactly 5 runs have been very consistent at 8-9% and 6-run wins have been right at 6%. I would imagine that even larger wins show more variation due to being increasingly rare events.