When it became evident that David Price would be pitching on Sunday and thus would be unable to pitch in the All-Star Game, John Farrell had a menu of options from which to choose a replacement. The fans voted in Chris Sale from a group of five Final Vote candidates, and the other four—all starting pitchers—were not pitching Sunday and very much available.

Garrett Richards would have made the most sense with a half-season that a traditionalist would love—a great record, ERA and team performance—and the underlying stats to go with it.

But Farrell didn’t go with Richards or any of the four. He didn't go for a starting pitcher to replace Price at all. Like when he picked Koji Uehara to replace the injured Masahiro Tanaka, Farrell chose a relief pitcher, Fernando Rodney. The American League team will now have seven starters and six relievers active instead of the original nine and four.

How one views this move depends a lot on how one views not only the purpose of the All-Star Game, which we’ll get back to in a second, but also on how one views who is having a better year. One could argue that, with a 1.98 ERA and a 2.38 FIP thanks to a much improved walk rate, Rodney is performing the job he’s been asked to do just as admirably as Richards has. But, as we know, starting pitching is a more important skill and the pitchers who are starting could probably do just as well or better if they were only asked to throw in one inning sprints.

Mostly, though, if you acknowledge that the starting pitcher is the candidate more deserving of recognition, the decision becomes about what the goal is. If the goal is to recognize the outstanding players of the first half, Rodney was probably the wrong call. But if the goal for the American League manager is to field the best American League All-Star team, it’s not so clear that this wasn’t the right call.

What Farrell did was take a pitcher who’s used to doing exactly what he’ll be asked to do in the All-Star Game: go to the bullpen, warm up, come in, probably to start an inning but maybe mid-inning, go all out for 15 or so pitches and then ice up.

Since 1988—and we pick 1988 as a tip of the hat to new BP boss Sam Miller, who believes modern baseball history begins then—there have been 382 instances of a pitcher relieving in the All-Star Game. In 211 of those instances, the pitcher was actually a starter for his regular team, and in the other 171, he was accustomed to the job as reliever.

The relievers have done better. Their ERA is 0.30 points lower, 3.66 vs. 3.96. Their walk rates have been nearly identical. And the biggest gap has been in the strikeout rates. As much as we’d like to think the starters who are tasked with shoving it up there for an inning can turn it on, the relievers’ strikeout rates in All-Star relief are still 30 percent higher.

All-Star Game relief appearances

Statistic Starters Relievers
Appearances 211 171
Innings 231 2/3 147 1/3
ERA 3.96 3.66
BB/9 2.4 2.4
K/9 6.4 8.3

There are a few alternative explanations for why the data may look like this beyond relievers being able to handle relief more expertly, although not all of them pull the same direction.

First of all, actual relievers tend to (although don’t always) work later in games and thus don’t face as many starting position players as the starters do in what are usually early-game relief appearances. This can help the relievers, but some years, the position player reserves picked by the players and manager are better than the fans’ selections.

The fact that the best starting pitcher (or the maybe-best starting pitcher) is excluded from this because he started the All-Star Game may widen the gap, although it’s not like starters have been stellar of late either.

And then there are selection issues that may work both ways. Relievers can have an easier time fluking their way to an All-Star stat line by compiling 35 or 40 good innings, so maybe they don’t belong as much. However, the fact that they’re usually closers means they’ve had real success in the role for longer than just three months, whereas a starter doesn’t need to have that track record.

All said, there are lots of small factors that go into this, but none that appear to overshadow the idea that it might not be so silly after all for Farrell to want to load his staff with relievers. The last five years were unique among five-year periods in that it’s the first with more relievers used in All-Star Game relief than starters used in All-Star Game relief.

All-Star Game relief appearances

Years Starters used Relievers used Reliever percentage
1989–1993 38 28 42%
1994–1998 42 23 35%
1999–2003 38 36 49%
2004–2008 46 35 43%
2009–2013 38 44 54%

It wasn’t just Farrell going this direction. National League manager Mike Matheny had more pitchers to replace than his AL counterpart, with four Sunday starters who couldn’t go and one injury. He went starter-for-starter in a few cases: Tim Hudson replacing Madison Bumgarner, Henderson Alvarez replacing the injured Jordan Zimmermann and Alfredo Simon replacing Johnny Cueto. But Matheny, like Farrell, is plus-2 on relievers, as Tyler Clippard fills in for Julio Teheran and Huston Street for teammate Tyson Ross.

There is certainly less hassle in using a reliever. That he’s used to the routine is just part of it; he’s also not going to be on short rest, nor will he have to conserve for his next outing. And, if it’s worth anything, he’s less valuable to his team if you put him in and he happens to get injured.

He also might be better at the job, and for all those reasons, it looks like the surge of the last five years is continuing.

Thank you for reading

This is a free article. If you enjoyed it, consider subscribing to Baseball Prospectus. Subscriptions support ongoing public baseball research and analysis in an increasingly proprietary environment.

Subscribe now
You need to be logged in to comment. Login or Subscribe
"First of all, actual relievers tend to (although don’t always) work later in games and thus don’t face as many starting position players as the starters do in what are usually early-game relief appearances. This can help the relievers, but some years, the position player reserves picked by the players and manager are better than the fans’ selections."

I'd also note that because they tend to pitch later, relievers are more likely to face platoons.
It seems to me that the 13th pitcher to be added to the roster probably shouldn't be scheduled to pitch during the first 9 innings and the starter (Richards in this case) is probably more useful if the game goes into extra innings.

I realize the Richards v Rodney decision was mostly a lead-in to your topic, not the real topic, but I thought I'd mention this perspective.
Notice in your chart above that the number of starters used hasn't really changed. It's just the number of relievers. More pitchers are being used, period, and the guys making up the gap are the relievers.
The All-Star game is supposed to be something special. It used to be special to see NL and AL players face each other but that literally happens every day now. It's also kind of special to see great starters come in for an inning or two, but clearly that is on the wane. And people wonder why ratings for the All Star game drop every years.