Hypothesis Disproved
Written March 2, 2009

Scientists frequently make guesses, then conduct experiments to find out whether those guesses are correct.  For example, maybe apple trees will produce more fruit if they’re pruned in the fall.  The only way to know for sure is to gather some data.  Go to the orchard, prune half the trees, and record their production next year as compared to the unpruned trees.

The same applies to sports.  For example, Pittsburgh Penguins captain Sidney Crosby is a big asset to his team.  Therefore, a reasonable hypothesis would be that the Penguins play better when Sid’s in the lineup than when he’s out with an injury.

However, the data suggests that they actually play better without him.  I’m talking about facts like these:  (1)  Last season as of March 11, in 49 games with Crosby, the Pens power play had only a 17.7% success rate, but in 21 games without him, they improved to 26.8%.  (2)  The sample size is smaller, but this season as of March 2, Crosby has missed four games, and the team has won all four of them.  Hmmm.

For another example, take the pace of baseball.  In 1960, Bill Mazeroski won Game 7 of the World Series for the Pirates with a home run.  Despite a final score of 10-9, the game took only two hours and 36 minutes to play.  But in this century, few postseason games are finished before midnight.  An exception was the deciding Game 5 of the 2008 World Series; because of rain, part of it was played on Monday and part on Wednesday.  The Phillies won 4-3 in a total time of three hours and 28 minutes.

What has slowed baseball over the last half-century?  Some factors are obvious, including increased reliance on the bullpen (such as stopping the game after every batter to bring in a new relief pitcher).  The number of TV commercials has increased.

However, I hypothesized that the problem is more fundamental.  To me, the game is actually being played more slowly and deliberately nowadays.  Batters step out after every pitch to adjust their gloves.  Pitchers agonize over which pitch they want to throw, then throw over to first base instead.  They’re ignoring Ray Miller’s dictum:  work fast, change speeds, throw strikes.

To find out whether I was right, I went to the video tape.  This winter, Major League Baseball’s new cable channel reran some classic telecasts, starting with Don Larsen’s perfect game in the 1956 World Series and  including the Series-clinching win by the Chicago White Sox in 2005.  Those two games are comparable because both were low-scoring pitcher's battles, 2-0 and 1-0 respectively.

I played back the final portion of each, including the innings when runs scored, and I marked down the time that each pitch was thrown.  Then I subtracted that from the time that the next pitch was thrown.  I averaged the results by category.

The old game and the recent one produced rather similar results.  Therefore, the data do not completely support my hypothesis.

Here are the numbers, first for 1956 and then, after the slash, for 2005.

When merely throwing the ball back and forth with the catcher, pitchers took only :20 / :22 between pitches.

If the batter hit a foul, replacing the baseball required an additional :13 / :11.

(A recently-enacted rule requires that, with bases empty, the pitcher must deliver the ball within :12, but that clock doesn’t start until he has the ball and the batter is ready.)

If there was at least one runner on base, the pitcher delayed an extra :06 / :05.  (I expected this number to be much greater, but in both games I chose, once the pitcher started throwing to the new batter he paid little attention to the runner.)

The average batter saw 3.6 / 3.9 pitches in a plate appearance.  A new plate appearance began 2:46 / 2:48 after the previous one began.

However, there were a couple of significant differences.  One was the batter change time.  That’s how long it took for a batter to see the last pitch of his plate appearance, to hit it to center field or strike out or whatever, and to run to his base, then for the defense to reset and for the next batter to come to the plate and see a pitch.

Following the last pitch of an at-bat, it took :42 / :48 before the first pitch was delivered to the next batter.  With runners on base, this increased to :61 / :72.  As there were more base runners in the 2005 game, the overall average for all batters was :46 / :67, with the modern players taking 21 seconds longer.

Every half-inning requires at least two of these batter changes, so this difference would lengthen a modern low-scoring game by at least 13 minutes.  The effect would be greater for higher-scoring games.

The other difference, of course, is television.  After the last pitch of a half-inning, it took 2:01 / 3:01 until the first pitch of the next half-inning.  That would make today’s longer breaks responsible for about 17 additional minutes per game.  (In 1956, Gillette was the only sponsor and aired only one commercial per break.  In 2005, the breaks as originally broadcast might have been even longer than they were in the re-edited version I timed.)

Conclusion:  Many factors contribute to longer games.  Not the least is that today, few pitchers can get the ball over the plate as consistently as Don Larsen and Sal Maglie did on that October afternoon in 1956.  And the “batter up!” procedure requires more time than it used to.  But the routine stuff, pitching balls and strikes, seems to move along at the same pace as always.