Have relievers really gotten better?

At some point along the way, the pitch count made its way into the conversation. And then it made its way onto the scoreboard. And then it made its way into the rationale for actual pitching changes. And then it made its way onto the television screen. And it made its way into the strategy of the game.

The starting pitcher used to come out when he "€œlooked tired" or he "€œhad nothing left"€ or he "€œgave up seven runs in one inning."€ Then came the advent of the 100-pitch timer. Once he crossed the century mark, it was a pretty good bet that the starter was on his way out the door — or at least off the mound. It seemed like everyone in baseball adopted this strategy. A few weeks ago, the Indians even removed Corey Kluber after eight innings and 18 strikeouts. And 113 pitches. It caused a commotion. Then again, we’€™ve seen Tim Lincecum and Johan Santana left in for more than 140 pitches to try to finish off no-hitters. Like it or not, fretting over the 100-pitch mark is here to stay.

And then came the counter-strategy. If the starter is going to come out after 100 pitches, let’€™s take a few extra pitches early in the game. It might cost us an extra strike, but if we can get the starter out by the sixth inning, the other team will have to go to its rather suspect bullpen. We can do our damage against the soft underbelly of the ‘€˜pen. Remember back in the days when the Red Sox and Yankees would get together on a Sunday night, take pitch after pitch and turn a nine-inning game into a nine-hour game? OK, well, some things never change.

Then came the counter-move to the counter-strategy. The evidence points out that letting pitchers go much beyond 110 pitches really is detrimental to their health. But what to do when the other teams kept purposefully dragging out at-bats longer than they used to? It was time to fortify the bullpen.

Last year, the Royals made it to the World Series, and a big part of their success was the relief triumvirate of Kelvin Herrera, Wade Davis and Greg Holland. All three men threw gas, were lights-out, and allowed Ned Yost to "€œshorten"€ a game to six innings. But taking nothing away from the Royals bullpen, there was a point when people started looking around and realizing that a lot of teams seemed to have these fire-balling, lights-out guys who could throw 96. It used to be that teams were lucky to have one. Now everyone had three or four. It seemed like one reason that scoring was down was that a manager now had the luxury of only needing his starter to go five or six innings, and he didn’€™t have to rely on suspect arms in the bullpen to make a lead stick. Now, it didn’€™t matter as much if the other team tried to spoil a bunch of pitches by fouling them off.

Is that really what happened?

Warning! Gory mathematical details ahead!

It’€™s no secret that offense has been in a spiral downward for the past few years, reaching levels not seen since the Bicentennial. In 2014, run scoring fell to 4.07 runs per game, down more than a run since the beginning of the millennium. It’€™s hard to tell whether we live in an era where the pitchers are just really good or the hitters are just really bad.

Maybe the answer isn’€™t that relievers have gotten better. Maybe it’€™s just that pitchers have gotten better. I took a look at data from 2003-2014. I looked at how many runs (earned or unearned) starting pitchers gave up per nine innings and found this:

We can see that starters have gotten stingier with giving up runs over time, going from a high around 4.75 in 2007 to a low of 3.85 in 2014. That’€™s a .90 drop in 7 years. Pretty impressive. Have relievers shown the same pattern?

One thing that we need to think about with relievers is that the quality of the relief pitching that a team is going to get is going to depend on how deep into a game their starter goes. If he’€™s out in the sixth inning, a manager is going to have to use guys further up the pecking order to make sure he doesn’€™t blow out the arms of his ace setup man and closer. Also, to make sure that we were getting a look at the good relievers, and not the mop-up guys, I looked at bullpen RA/9 for games in which, when the bullpen was activated, the score was within three runs either way.

Here’€™s the bullpen RA/9 for games in which the starter finished the sixth inning, but did not come out for the seventh:

And the bullpen RA/9 for games in which the starter finished the seventh, but did not come out for the eighth: 

Most of the other graphs have the same basic shape. The rates for mid-inning changes are a little higher, because a common reason for a reliever to come in mid-inning is that there are already runners on base, which will lead to more runs being scored after the bullpen comes in. This method puts starters and relievers on the same footing.

Looking at these graphs, we see that bullpens, writ large have improved a lot over the past few years. If you look at the top of the graph and the bottom of the graph, we see that the range of their improvement has been about 1.1 to 1.2 runs of RA/9. That’€™s certainly nothing to sneeze at, but let’€™s keep it in context. Starters have seen a drop almost as big. Almost.

Still, the drop for relievers has been 20 to 30 percent bigger than that for starters. It seems that the new crop of relievers really is doing a better job. While they aren’€™t entirely responsible for the dip in offense, there’€™s evidence to suggest that the guy in the bullpen now is better than the guy in the bullpen was 10 years ago.

And soon the counter-gambit to the counter-move to the counter-strategy …

Now that we know that teams have better bullpens at their disposal than they used to, what’€™s the counter-move? One that’s been suggested is that hitters should stop being so patient and just attack what the starter gives them. In fact, there’€™s evidence that one reason for the recent rise in strikeouts is that hitters are being too patient. If the batter comes up saying, "€œI won’€™t swing until you throw me a strike,"€ the pitcher is likely to say, "€œOK, here"€ and we end up in an 0-1 count. If bullpens really are getting better, maybe it’€™s worth it to go after the No. 4 starter and make sure that it’€™s already 7-2 by the fifth inning so that the All-Star bullpen doesn’€™t matter.

But the other thing that might have to happen is that if there are going to be bullpens filled with guys who throw gas is to start looking for hitters who can handle velocity. Hitting is a complex neurological process. It involves pattern recognition, reaction time, eye tracking, muscle coordination, and about a half-dozen other neurological systems, all coordinating with each other within a half second. Some people are good at quick reactions, which you need to hit velocity. Some are good at tracking movement. Some can pick up on a pattern of what a pitcher is throwing and anticipate it. And these are all separate skills and some hitters are better at some of them than others. Major-league hitters are probably good in at least one of these areas. You can be good at all of them, and if you are, you’€™re probably on your way to the Hall of Fame.

The next counter-gambit might just be finding guys who are good at hitting the kinds of pitches that are out there in the ecosystem. If it’€™s all 96-mph heat, let’€™s find some good fastball hitters!

But then there will be a counter-ploy to the counter-gambit to the counter move to the counter-strategy …