Too good to start, too talented for the 'pen

This is the rule about relief pitchers, even the best ones: They are failed starters.

Maybe they lack usable secondary pitches. Maybe they break down under the strain of 100-pitch outings. Maybe they're too emotionally wound.

If they could start, they would. The White Sox, back in the mid '70s, had a couple of young hard throwers who had had success in relief. Paul Richards, who had been a genius manager in the 1950s as a builder of pitching staffs, was back in the dugout after a couple of decades as a general manager, and he moved Terry Forster and Goose Gossage into the rotation, because they were his best arms and his best arms belonged in the rotation.

As a relief pitcher in 1975,
Gossage struck out 130 batters
in 141.2 innings; as a starter the
following year, 135 Ks in 224
innings.

Forster went 2-12, 4.37; Gossage went 9-17, 3.94. And that was that for them as starting pitchers. Forster is remembered now, if at all, for David Letterman's gibes about his weight — the fat tub of goo — but he was a very good reliever for more than a decade. Gossage is now in the Hall of Fame, and if you're going to put relief pitchers in Cooperstown, Gossage is the place to start — but he was a failure as a starter.

That's the rule. One can find exceptions here and there — Dave Righetti, Joakim Soria – but the rule holds. Mariano Rivera became a reliever after having Tommy John surgery. Joe Nathan moved to the pen after getting hurt. The Twins tried hard, really hard, to find a starting pitcher in LaTroy Hawkins'  talent and never found it. (Baseball history, at least the part before pitching staffs got split between starters and relievers, is riddled with guys like the Hawk —  check out Si Johnson and Milt Gaston as examples.  Great arms, bad teams, lousy results.)

These guys thrive in the bullpen because it's less demanding. They can succeed without a change up, or simply keep their joints attached,  for 15 pitches at a time.

But here's a odd trend developing, and Joba Chamberlain (referenced in my previous post) is a harbinger of it.

Chamberlain was a revelation in 2007. The big guy burst onto the scene with 24 relief innings, 0.38 ERA. The front office imposed "the Joba rules"  — usage limits — to keep manager Joe Torre from burning the kid out. He was going to be their next star, the rotation anchor of their future. But first, they wanted to win this year, and they needed a big arm in the pen.

Things have gone down hill from there. Chamberlain's ERA has gone from 0.38 to 2.62 to 4.75 to 4.40. General manager Brian Cashman this winter essentially wrote him off as a starter, and the Yankees don't trust him as a late-inning option right now either.

Consider the Texas Rangers and Neftali Feliz. Or the Cincinnati Reds and Ardolis Chapman. Or the White Sox and Chris Sale. Three young power arms, all stuffed into bullpen roles to help the here-and-now rather than develop as starters in the minors.

From this perspective, it's silly. They're be far more valuable as successful starters than as successful relievers, because 200-plus innings is more valuable than 60 or fewer. But as their teams saw it last season, and most likely this season as well, their immediate impact is too tempting to delay the gratification.

And they may be right. Chapman may not be able to develop his secondary pitches. Sale's delivery may be too risky for heavy use. The Rangers aren't about to ship their Rookie-of-the Year closer back to Triple A to master a change up, and as defending AL champs aren't likely to opt for on-the-job training for Feliz as a starter. Even Chamberlain — as controversial as the use limits put on him have been, he did have a history of arm problems in college.

The Twins are not immune to this thinking either. Remember: They resisted using Johan Santana in the rotation because Ron Gardenhire so loved having that extra power arm in the bullpen. What a wasted resource that would have been.

There are enough failed starters around with which to build bullpens. Let the most talented ones fail as starters first.