Reliever domination is at a historic level

Around twenty years ago a threshold was crossed for good with regards to the way innings are distributed amongst pitchers. Up until 1995 relievers had always been responsible for throwing fewer than a third of all the innings in a season, but from that season on they were handling at least a third. The figure below shows that crossing this mark was not something that came about from a drastic change. Rather, the proportion of innings thrown by relievers has been consistently increasing over time and this is having a considerable impact on offensive output.

For most of the last twenty years, getting into the bullpen early was a core tenet of successful offenses. Seeing a lot of pitches early in the game drove up starters' pitch counts, leading to their earlier removal and meaning a chance to feast on the often-inferior pitchers that were used as relievers. But with the way bullpens are constructed now - usually with seven or eight guys who  throw hard, are usually well rested and rarely throw more than an inning (only 20 or so pitches per outing) - it seems reasonable to question this approach. That soft nougaty center of the pitching staff may not be there to knock around anymore. We know that run scoring in general is down, but to what extent is this a result of weaker performance against relievers?

To take a stab at responding to this question I will trot out my trusty friend the Baseball-Reference.com play index, look at league splits against relievers and compare the performance in that split to overall performance using tOPS+. This is similar to what I have done in a couple of other posts this year. For those unfamiliar, a tOPS+ of 100 is the baseline. On offense, which I will be looking at here, higher numbers mean the team is better in the given situation than in general and lower numbers mean they have been worse than in general. I am interested in the league split for when batters are facing a relief pitcher. A tOPS+ over 100 shows that hitters are faring better than normal against relievers, and a result under 100 shows they are faring worse. The absolute number is one thing, but the difference across seasons is also important.

In the figure below you can see the tOPS+ results for the 1975 - 2015 seasons. These seasons are shown because they are the seasons for which data are available:

As is evident, batters have tended to perform worse against relievers than they do in general, but this is especially true for the last five years. At a time when relievers are throwing the largest percentage of innings they ever have, batters are really having difficulty facing them. Here are the ten seasons for which tOPS+ is lowest (including ties):

	Year	OPS v. RP	Total OPS	tOPS+
1	1982	.687	.713	93
2	2007	.730	.758	93
3	2012	.696	.724	93
4	2001	.734	.759	94
5	2006	.746	.768	94
6	2011	.696	.720	94
7	2013	.690	.714	94
8	2015	.690	.713	94
9	1989	.676	.695	95
10	1998	.734	.755	95
11	2004	.743	.763	95
12	2008	.731	.749	95
13	2009	.729	.751	95
14	2014	.682	.700	95

Each of the last five seasons (2011-2015) appear on this 'leaderboard'. In fact, save 2010, in which batters had decent success against relievers (97 tOPS+), each of the last ten seasons are found on the wrong end of this performance spectrum. I must note that the 2015 sample is limited as it is not a complete season, but thus far the performance is right in line with what we have seen of late. It is tied for the second lowest mark since 1975. We really are in an interesting strategic point in the history of the game. Relievers are being used more often, for shorter and shorter stints, which is perhaps contributing to them throwing harder and harder, and it is dampening offensive output.

Year	Avg Outs per Appearance	Avg. RP Fastball Velocity
2006	3.25	-
2007	3.18	91.7
2008	3.21	91.6
2009	3.16	92.0
2010	3.07	92.4
2011	3.07	92.6
2012	3.04	92.9
2013	3.13	92.9
2014	3.03	93.0
2015	3.07	93.1

*Velocity data are for fourseam fastballs as recorded by PITCHf/x and listed on FanGraphs

Given this information it should not be all that surprising to find that the teams that strategically invested in constructing their bullpens over the last couple of seasons (Royals, Astros, Yankees, Blue Jays) are reaping rewards.

The Royals have their three-headed monster, which is now somewhat of a four-headed monster with Ryan Madson. Royals' relievers' RE24: 64.97 (best in the game).

The Astros signed relievers Luke Gregerson and Pat Neshek to multi-year deals before many thought they should be investing in building their bullpen. Astros' relievers' RE24: 43.05 (third best in the game).

The Yankees signed Andrew Miller to go with Dellin Betances and a suite of other failed starters from their system. Yankees' relievers' RE24: 10.57 (13th best in the game).

The Blue Jays are another example. Much was made about the Blue Jays' acquisitions of David Price and Troy Tulowitzki at the trade deadline, which were no doubt huge pickups, but getting LaTroy Hawkins and Mark Lowe were also very important for the success of the team. Having David Price allowed them to move flame-throwing youngster Aaron Sanchez to the 'pen, in front of other flame-throwing youngster Roberto Osuna; making for a nice little end-of-game duo. Slotting Hawkins and Lowe in with Brett Cecil, Aaron Loup and Bo Schultz, results in the rounding out of a formidable group. They went from having a middle of the pack bullpen (6.80 RE24, 3.99 runs allowed (RA) in the first half) to a group that has been top five since the trade deadline (6.03 RE24, 1.96 RA). Obviously small sample caveats must be applied here, but, with their moves, the Jays have effectively shortened the game.

None of this is meant to suggest that building a strong bullpen is necessarily easy now. Relievers, because they throw fewer innings individually per season, are difficult to predict from year-to-year. Strong relievers can seemingly come out of nowhere. For instance, nobody really knew who Roberto Osuna was 6 months ago but now he's closing games for a team that looks to be playoff bound.

The message here is that relievers are responsible for the largest proportion of innings they ever have been and are dominating hitters at a historic level. This raises more questions about strategy and what teams will need to do to adjust. They can get in line for the next lights-out reliever and continue to the trend, start working on collecting hitters that are capable of chipping away at the relievers, and/or implement more 'bullpen games' in which starting pitchers, particularly back-end rotation guys, are limited to 15-20 total batters faced per start. This approach has worked fairly well for the Tampa Bay Rays this year. It avoids pitchers being penalized by working a third time through the order, and can reduce the opponents' ability to stack the lineup in such a way that they consistently have a platoon advantage. If implemented well it could be really something for National League teams, as they could often avoid having their pitchers' hit. Teams could further this idea by starting the game with a 1-inning reliever; a pitching role Bryan Grosnick dubbed an Opener. In any case, it seems there is room to get ahead of the curve with on-field strategy as it pertains to this era of reliever dominance.

All told, we are two-thirds of the way through the season and to date batters have floundered against relief pitching. However, given that we are into August and there are a number of teams already planning for 2016, we are likely to see more relief appearances from players further down the depth chart, so the performance level can rise over the next couple of months. Yet, this is undoubtedly something that could have been said in each of the last few seasons and we saw how things ended. I for one welcome our new reliever overlords; perhaps you should too.

. . .

Chris Teeter is a featured writer and editor at Beyond the Box Score. He is also a contributor at BP Boston. You can follow him on Twitter at @c_mcgeets.