Run differential far from a perfect predictor

Among the shrines at which I no longer worship, we find run differential.

• Cast your Esurance All-Star ballot for #ASGWorthy players

I lost my faith in this statistical denomination in 2007. This was when the D-backs ended the regular season with a run differential of minus-20. They should have finished in what used to be called the second division -- or the Pacific Coast League, for that matter -- with a crummy run differential like that.

But the 2007 D-backs apparently were not aware that they were breaking fundamental mathematical laws. They went 90-72. They won the National League West. Later, they advanced to the NL Championship Series. Their manager, Bob Melvin, won the NL Manager of the Year Award. Those D-backs won a bunch of stuff -- with a run differential of minus-20.

Video: The Diamondbacks clinch a spot in the 2007 Postseason

But in the eternal search for certitude, I occasionally return to this statistic in the hope that it will live up to its advance billing and actually predict success. I have checked in again this season, waiting until the summer solstice had passed, so that the sample size would be substantial enough.

And I found once again that run differential does predict success -- once in a while. Yes, run differential works, except in those cases in which it doesn't.

For instance, the best run-differential team in the American League is Toronto, at plus-80. The Blue Jays are merely the fourth-best team in the AL East, a division notable for not containing anything like a dominant team.

Run-differential conclusion: Wow, this isn't particularly close.

Video: TOR@TB: Encarnacion smashes a 9th-inning solo homer

The worst run-differential team in the AL is the White Sox, at minus-75. OK, the White Sox are in last place in the AL Central. While they don't have the league's worst record, they're in the neighborhood.

Run-differential conclusion: As in horseshoes, this close can count.

But, but, but ... here are the Oakland Athletics, last in the AL West with the second-worst record in the league. But they have a run-differential of plus-37, which should make them the fifth-best team in the AL.

Run-differential conclusion: We can only determine that run differential does not apply to teams managed by Bob Melvin. How else to explain run-differential predictor missing so badly on the 2007 D-backs and the '15 A's?

Video: OAK@TEX: Phegley lifts a solo shot to left-center

Meanwhile, in the Senior Circuit: The best run-differential team in the NL is the Cardinals, at plus-78. They have the best record in the league. They have the best record in the Majors.

Run differential conclusion: You can't miss on the Redbirds. Run differential didn't miss on them, either.

Video: STL@MIA: Heyward drills a solo homer to right-center

The NL's worst run-differential team is the Phillies, at a truly unhealthy minus-114. They also have the worst record in the Majors.

Run-differential conclusion: Another one in the "W" column for the University of Run Differential. But again, how could it miss on a team that is -- to put it kindly -- this relentlessly inadequate?

So, as summer begins, run differential once again presents us with a mixed bag of results that include some precise indicators and some complete misses. As a reliable predictor of baseball events, it is not quite as good as simply studying the standings.

Or, as Adam Wainwright once suggested, be sure to let me know when the pennant is decided by WHIP. Run differential is a tool, but too often, it comes up well short of the two fundamental measurements of success and failure: wins and losses.

We'll check back during the postseason, when people will tell you that run differential is a better predictor of playoff success than win-loss records.

Right. In 2014, which of the five AL postseason teams which had the worst regular-season run differential? The Kansas City Royals. Of the five NL postseason teams, which had the second-worst regular-season run differential? The San Francisco Giants.

Oops. A less than stellar October for run differential. Maybe next time.