Olney: The rising price teams pay by going to the bullpen early

Just back from the IL, Brandon Kintzler was immediately given a one-run Cubs lead he couldn't protect. Charles LeClaire/USA Today Sports

WILLIAMSPORT, Pa. -- Yu Darvish had thrown seven dominant innings Thursday night in Philadelphia, requiring just 92 pitches, and Chicago Cubs manager Joe Maddon took him out with a more than reasonable expectation that his bullpen would protect a five-run lead and get the last six outs.

It didn't happen that way, of course. After an error and three more pitching changes, six hits and only four outs, Bryce Harper blasted a walk-off grand slam, giving the Phillies a 7-5 win over the Cubs.

The move that Maddon made has become something of conventional wisdom. It used to be that managers would feel comfortable leaving an overpowering pitcher on the mound for 120 pitches, but that line of demarcation drifted downward to 110 pitches, then 100. Now managers start to look to pull the parachute on a starter if he hits 85 to 90 pitches, once he begins to pass through the opposing lineup a third time.

Because the numbers have suggested that, generally, it's better to use a fresh reliever instead of a starter nearing 100 pitches, teams generally are using their bullpen for more outs and expecting less of their starters.

But the numbers are changing, and bullpen performance is worsening. According to league-wide numbers dug out by Paul Hembekides, starters (4.52) and relievers (4.52) own the exact same ERA. The last time league-wide bullpen ERA was not lower than that of starters was 1969. The gap between starters and relievers is shrinking quickly -- nearly half a run over the past four seasons.