What's with all the perfect games?

[Jeff Lichtman]

A perfect game is one of the rarest achievments in baseball. There have been over 180,000 games in MLB since the National League was founded in 1876, and only 21 of them have been perfect. So doesn't it seem odd that there have been four of them in the past three years? They are:

• Mark Buehle, White Sox vs. Rays, July 23, 2009
• Dallas Braden, A's vs. Rays, May 9, 2010
• Roy Halladay, Phillies vs. Marlins, May 29, 2010
• Philip Humber, White Sox vs. Mariners, April 21, 2012

There was also Armando Gallaraga's near-perfect game for the Tigers against the Indians on June 2, 2010. What should have been the 27th out was ruled a base hit on a blown call by the first base umpire. That would have been five perfect games in three years, and three in a month.

I found a reference on-line that said 172975 games were played in the NL and AL from 1876 through 2008. With 17 perfect games in that time period, there were 10175 games per perfect game. Since then, there have been about 2449 games per season (including playoff games). Assuming that there will be the same number of games this season, and that there won't be any more perfect games in 2012, that comes to 2449 games per perfect game. So, perfect games from 2009 through 2012 will have been more than four times as frequent as in all of preceding major league history.

Could this just be random clustering? Or has something changed in Major League Baseball that has resulted in more perfect games?

[furt]

Originally Posted by Jeff Lichtman

I found a reference on-line that said 172975 games were played in the NL and AL from 1876 through 2008. With 17 perfect games in that time period, there were 10175 games per perfect game. Since then, there have been about 2449 games per season (including playoff games). Assuming that there will be the same number of games this season, and that there won't be any more perfect games in 2012, that comes to 2449 games per perfect game. So, perfect games from 2009 through 2012 will have been more than four times as frequent as in all of preceding major league history.

Lumping all of history from 1876-2008 together like that conceals a lot. There were no perfect games for 30 years from 1922-1956 ... not coincidentally, the 20s and 30s were Baseball's biggest era for offense. There were 3 in the 1960s, but none in the 1970s. And then we've had 12 since 1981.

Or has something changed in Major League Baseball that has resulted in more perfect games?

My theory: hitter's strategies have changed. In the past, when faced with a pitcher with dominating stuff, hitters were inclined to choke up, shorten their swings, and focus on trying to just put the ball in play instead of striking out. That isn't the case nowadays; most hitters, managers, GMs and baseball analysts don't see strikeouts as an especially bad thing, and generally encourage hitters to take their normal swings all of the time.

In any event, strikeouts per game tracks pretty well with the periods where there have been a lot of perfects.

[Cliffy]

In addition, enhanced (physical) training and attention to players as athletes (and, of course, the possibility of juicing) stands to make any given pitcher better in terms of speed and possibly control. It helps the offense too, of course, but by uneducated guess is that it would enhance power more than average.

--Cliffy

[critter42]

Originally Posted by Cliffy

In addition, enhanced (physical) training and attention to players as athletes (and, of course, the possibility of juicing) stands to make any given pitcher better in terms of speed and possibly control. It helps the offense too, of course, but by uneducated guess is that it would enhance power more than average.

--Cliffy

There's also the possibility of the reverse - that the LACK of juicing on the hitters' part has made the pitchers' job easier - note that the rash of perfect games started AFTER the steroids scandal really broke.

[Maserschmidt]

I wonder if expansion increases the likelihood, i.e. if the league has generally crappier-hitting teams at the bottom.

Actually, you could probably analyze that if you wanted to pull the data.

[RickJay]

Originally Posted by critter42

There's also the possibility of the reverse - that the LACK of juicing on the hitters' part has made the pitchers' job easier - note that the rash of perfect games started AFTER the steroids scandal really broke.

But there wasn't a glut of perfect games back in the day, before steroids.

It's just a fluke that we've had four (five, really) of them in three years. They've never been evenly spaced, anyway, this is just a bit more flukey than it was before.

Of course, everything else being equal, it should be a bit (a TINY bit) easier to pitch a perfect game now just because there are fewer errors. An error ruins a perfect game, but errors are much more unusual than they used to be. In 1922, when Charlie Robertson (no superstar) pitched a perfect game, the average AL team made about 190 errors a season. Today the average AL teams makes half that many errors. So there's less a chance your perfect game will be ruined by some idiot clanging one off his glove.

Originally Posted by Maserschmidt

I wonder if expansion increases the likelihood, i.e. if the league has generally crappier-hitting teams at the bottom.

Actually, you could probably analyze that if you wanted to pull the data.

Just looking at a list of perfect games it doesn't appear to be the case, at least not convincingly. Since the beginining of expansion in 1961, here are the perfectos:

1964 - Bunning, vs. Mets
1965 - Koufax, vs. Cubs
1968 - Hunter, Vs. A's
1981 - Barker, vs. Blue Jays
1984 - Witt, vs. Rangers
1988 - Browning, vs. Dodgers
1991 - Martinez, vs. Dodgers
1994 - Rogers, vs. Angels
1998 - Wells, vs. Twins
1999 - Cone, vs. Expos
2004 - Johnson, vs. Braves
2009 - Buerhle, vs. Rays
2010 - Braden, vs. Rays
2010 - Halladay, vs. Marlins
2011* - Galarraga, vs. Indians
2012 - Humber, vs. Mariners

There's no obvious connection between expansion and perfectos. The only recently-expanded teams to be victimized are the Mets in 1964 and Jays in 1981, and the Blue Jays one is a stretch - five years should be enough time to come up with a few good hitters. You had two perfect games in the two seasons following the 1998 expansion, but the 1969 super expansion was immediately followed by a 12-year drought in perfect games, and of course we now have a glut with no recent expansion.

Of course expansion makes it likelier a perfect game will happen just because there are more games. This season has twice as many games scheduled as the 1960 season. But you'd be hard pressed to convince me the level of skill and talent is lower now.

[Ulf the Unwashed]

One thing that shouldn't be ignored is the cumulative effect of slight changes in offense.

In 2003 the aggregate MLB on-base percentage was .332. So all other things being equal*, a pitcher had a keep-off-base-percentage of .668--that is, the average pitcher would retire the average batter 66.8% of the time.

So to find the theoretical odds of pitching a perfect game in 2003, you'd need to take .332 to the 27th power. Which is ~ 1.86 x 10 to the -5.

In 2010 the aggregate MLB OBP had declined to .325. Doesn't sound like much of a change, but it increases the pitcher's keep-off-base-% to .675. Raising that to the 27th power gives about 2.46 x 10 to the -5. So dropping the average OBP by just seven points makes a perfect game about 1/3 more likely.

Things are way more complicated than that, of course (see note below). But viewed in that way perfect games would certainly have been more likely in 2010 than in 2003, because that little difference in the ability to get on base becomes seriously magnified as we move through the lineup two and then three times.

*Disclaimer: First, yes, I know nothing is ever equal.

In particular: This analysis leaves out reached-base-on-error, so the chances of reaching base are slightly higher than given; that doesn't invalidate the main point, which is that the harder it is to reach base, if only by a little, the easier it is to pitch a perfect game, sometimes by a lot.

Also in particular: No lineup is made up only of guys who all have the exact same OBP. The further away you are from that perfectly balanced lineup, the harder it is to pitch the perfect game. If you had a guy with an OBP of 1.000 and two guys with an OBP of .000 (same number of PAs), the aggregate OBP would be .333, but a perfect game against 'em would be impossible. Also, a lot depends on the pitcher, each of whom has his own KOBP, and the connection between OBP vs a lefty vs OBP vs a righty--there are dozens of factors.

But I'm just talking about theory here. If we could design a couple of dice, 1000-sided dice, one with 332 faces shaded red (a .332 OBP) and one with 325 faces shaded red (a .325 OBP), and rolled 'em 27 times in a row and recorded what came up, we'd see that we could get 27 straight non-reds much more often with the 2nd die than with the first.

[Bootis]

Another factor may be better gloves, better field conditions (I'd bet bad hop hits were more common in the first 60 years or so) and better fielders in general, with MLB players being chosen from a much, much bigger pool of potential prospects now than in the past

[Intelligently Designed]

There might be a few outlying factors, but it is essentially a fluke. To pitch a perfect game, you need to have oustanding stuff on a particularly day, but you also need luck, and sometimes lots of it. As in, for example, the umpire ruling that the last Seattle batter against Humber had swung on that 3-2 pitch, or DeWayne Wise's fabulous catch in the 9th inning to maintain Buehrle's perfecto.

Let's also note that 4 of the last 5 perfect games (I include Galarraga's) have occurred in the American League, where it should be tougher because of the DH. So it's mostly random.