Lightning struck twice yesterday, for the second time in the last three years, as Matt Cain pitched a perfect game against the Houston Astros less than two months after Philip Humber accomplished the feat. Just two years ago, Roy Halladay and Dallas Braden thew perfect games just 20 days apart. Prior to that, there had been only two occasions when perfect games were even thrown within a season of each other, and you had to go back to 1880 for two in the same season.

Cain’s perfect game was the 22nd in MLB history, and half of them have occurred since 1988. Over 20% of them (5) have happened in the last four seasons.

So why the dramatic increase in perfect games in recent years? Well, they are still extremely rare events, and they still require a lot of things to go right. We still may not see another one for several years. It probably won’t be 34 years, though, which was the span between Charlie Robertson (1922) and Don Larsen in the World Series (1956). There are factors that at least increase the environment conducive to a perfect game now versus in the past: opportunity, better fielding, and more strikeouts by pitchers.

The first is pretty self-explanatory and obvious. There are more teams, more pitchers, and more games. With the additional 8 games and almost double the teams compared to before 1960, a perfect game is twice as likely.

I also looked at the overall league pitching numbers in the AL for every tenth year, starting in 1901, and continuing to last year (via baseball-reference’s league encyclopedia). Here are the WHIP (Walks + Hits per Inning Pitched), Strikeouts and Walks per 9 Innings, BABIP (Batting Average on Balls In Play) and Errors per 9 innings for each of those years.

The WHIP was highest during that stretch where there was a perfect game drought. BABIP has fluctuated but is higher than in the past, so it’s not that the balls in play are converted more frequently. The biggest difference, though, is that fewer balls are put in play. The strikeout rate is almost double what it was during that stretch from 1922 to 1956. It began climbing in the 1960′s, when the pitching friendly rules went into effect, and we got two high strikeout perfect games from Jim Bunning and Sandy Koufax.

The other difference is that the error rates are also much lower now, and you can’t have the ingredients for a perfect game if a fielder commits an error. The next time an old-timer talks about how these fielders just don’t have the fundamentals of the old timers, you might want to show him the error trends over time.

In the last decade, we’ve seen three of the six perfect games thrown by Randy Johnson, Roy Halladay, and now Matt Cain. All three had at least 11 strikeouts, with Cain’s 14 being the most. Philip Humber also had 9 strikeouts in his perfect game. None of the perfect games before Bunning’s in 1964 reached 9 K’s. With fewer balls in play, the pitcher has a better–still small–but better chance at perfection. Over half the batters Cain faced did not hit the ball into the field. His chances at perfection were more than 30 times a pitcher with 7 strikeouts over 9 innings, assuming similar BABIP rates.

Cain threw a perfect game, and if you were going to line up the candidates, he would have been in the top 10 right now, unless you were superstitious about his historic lack of run support or prior failed efforts at no-hitters. He is also pitching at a time when the environment is a little more favorable, and perfection isn’t as unattainable, thanks to power pitching and better fielding.

[photo via US Presswire]