Stats in the Wild

On April 21, 2012 Phil Humber threw a perfect game.  A few days later this was followed up by a no-hitter by Jered Weaver.  Then nearly an entire month passed before we saw another no-hitter thrown by Johan Santana, the first in Mets history.  7 days later, Kevin Millowd and 5 other guys I’ve never heard of combined for one of the more bizarre no-hitters in history. Then less than a week after that Matt Cain pitches the second perfect game of the season.  For those scoring at home, that’s five no-hitters so far this season, two of which were perfect games.  And it’s only June 15.  The seems like a lot.  (NPR agreed and they discussed “Why Are Baseball’s No-Hitters No Longer Rare?” with Stefan Fatsis.)

So, I wrote some R code and scraped box scores for all games from 1918 through 2012 that are available on baseball-reference.com.  I pulled out all of the perfect, no-hitter, and one hitter games and plotted than by year.  (Note: My graph doesn’t exactly match with the official MLB record of no-hitters.)

The graph shows two plots.  The top plot consists of the raw counts of perfect games, no-hitters, and one-hit games by year.  However, since fewer games were played in the early part of the 1900s, it’s not a fiar comparison  to compare 1920 to 2012.  The bottom graph attempts to correct this.  Using the rate at which the rare pitching feats were occurring, the bottom half of the graph displays how many of the games there would have been over a 162 games season with 30 teams at the rate for each year.

You can notice at the far left of the graph their were the most one-hit games in any year over the time period plotted.  This was considered the end of the dead ball era and you can see as the 1920s rolled around many fewer low hit games were pitched over the decade.  You can also see in the 1990s and 2000s that there is a big drop in these low hit games, probably as a result of the steroids era.

Now it might seem like we are entering a golden age of pitching as.  In the last three years, their have been 5 perfect games (Not even counting Armando Galaraga).  A total of only 22 have ever been thrown and only 20 of those were in the modern era.  At this rate of no-hitters and perfect games so far this season, we are on pace for over 10 no-hitters and about 5 perfect games, both of which would be unprecedented.  The most no-hitters in a season, according to MLB, is 7, which occurred in both 1990 and 1991.  The most perfect games in a season is 2, which happened in 2010 and 2012.

What strikes me as most interesting about the graph is that if you compare the number of one-hitters projected this season, it is high, but certainly not the most ever.  8 seasons have produced more one-hitters of better than this season is projected to produce when controlling to season length and the number of teams.  What seems to be happening is that the rate of these great pitching outings actually being converted to no-hitters or perfect games seems to be at a very high rate.  I don’t have a reason why this might be happening, but let me throw this idea out there: Do you think that if a game gets to the ninth inning and the pitcher has a no-hitter or perfect game going that umpires are reluctant to make a call that would jeopardize the pitcher’s feat lest they be remembered in the same way as Jim Joyce?  This might explain at least some part the high rate of converting a near no-hitter or perfect game to an actual no-hitter or perfect game this year.  Of course, it’s possible that we’ve seen all of the no-hitters and perfect games that we’ll see this year, and when we look back, it will just be another regular year with an anomalous few months to start the season.

UPDATE: Jason Hammel just added another 1-hitter tonight taking a no-hitter into the 7th inning where it was broken up by Jason Heyward.

UPDATE: Dickey just added another 1-hitter.  I’ve updated the graph to reflect this here.

Cheers.