Offensive Support In Perfect Games

I caught the last two innings of Felix Hernandez’s perfect game last week, which was quite fun not only because of the tension of the perfect game, but also because the game itself was still on the line, and mattered quite a bit to the Tampa Bay Rays. The Mariners wound up with a 1-0 win, a score which seemed rather common for perfect games.
And indeed it is: in the 23 official perfect games in MLB history, the final score has been 1-0 more often than any other outcome. Here’s the summary:
1-0 7 times
2-0 4 times
3-0 2 times
4-0 5 times
5-0 2 times
6-0 2 times
10-0 1 time
This is a total of 73 runs in 23 games, or just under 3.2 runs per game. While 23 games is not a big sample size, these data are consistent with a hypothesis that some other environmental factors may tend to depress offense on days when there’s a perfect game.
I noted this on a thread at Tom Tango’s blog, and he asked about seeing OBP and SLG numbers. I found that Baseball Almanac conveniently has a page listing all the perfect games, with links to boxscores for each game before this season. So it wasn’t too hard to get at raw data to try to figure this out. Using their site, and Baseball Reference for the games this season, I extracted all the perfect game pitcher’s team’s batting data into a spreadsheet.
Alas, the old boxscores are quite sparse, so I didn’t find the raw data to compute OBP, but I could compute batting average and slugging percentage. I’ve copied the raw data I used into a spreadsheet. The raw numbers were a .225 average (162-719) and .312 slugging percentage (224 TB). These numbers are markedly below average. The Baseball-Databank database has MLB data from 1871 through 2010, and the cumulative batting average is .2625, and slugging percentage .3825, so in perfect games batting average dropped nearly 14%, and slugging percentage declined even more, about 18.5%. The overall MLB team scored 4.555 runs per game, so run scoring dropped even more than average or slugging, which makes intuitive sense.
Since the Baseball Almanac also had links to boxscores for Harvey Haddix’s game (he pitched 12 perfect innings before giving up a hit and losing the game in the 13th) and Pedro Martinez’s 9 perfect inning game against the Dodgers (the Expos didn’t score until the top of the 10th, and Martinez was pulled after giving up a hit to lead of the bottom of the inning), I also reran numbers adding those two games. This does introduce bias (obviously neither offense scored in the first 9 innings, or else the game would have ended), but they didn’t make much difference, as the totals were now .224 and .301 in 800 AB. Haddix’s Pirates had a dozen hits, all singles, while the Expos had just 5, also all singles.
Some other amusing trivia from scanning through perfect game boxscores:

  • Catfish Hunter drove in 3 runs himself, going 3-4 with a double, in his 4-0 win.
  • Those were the only hits ever by an American League pitcher in a perfect game.
  • 6 NL pitchers had a hit in their gems: Lee Richmond, John Ward, Jim Bunning, Dennis Martinez, and Matt Cain each had one hit apiece.
  • Bernie Williams went 3-3 with a HR in David Well’s gem, while Jose Canseco hit a pair of solo HR to back Kenny Rogers.
  • Matt Cain’s Giants won 10-0, with every starter except Angel Pagan getting at least 1 hit. The Giants went 15-37 with 2 HR and 3 2B, a .405 Avg and .689 Slg. Removing that game from the total (admittedly cherry picking) would reduce the cumulative numbers to .216 and .293 (and now just 2.9 runs/g).
  • Lou Johnson’s double off Bob Hendley was the only hit in Sandy Koufax’s perfecto. Johnson scored the only run.
  • Cy Young was the only player on his team not to get a hit in his perfect game.
  • Richmond and Ward, the two 19th century pitchers, are also the only ones (other than AL pitchers during the DH era) not to bat 9th. Richmond hit 2nd, and Ward 6th.

Leave a comment

Your email address will not be published. Required fields are marked *