How Well Has The Opener Worked?
The now-famous opener strategy was introduced to MLB on May 19, 2018 when back end reliever Sergio Romo was used as the starting pitcher. Since then, the Rays have used openers and bullpen games over 100 times. Looking ahead to the 2020 season, the Rays have an overwhelming possibility of using a traditional 5 man rotation. So, in the opener’s prime, how’d it do?
In a pure win-loss perspective, the Rays have done slightly better without a regular starter than with one. Since 5/19/18, they’ve gone 62-42 (0.596 W%, 97 win pace) in a game started by a reliever and 106-77 (0.579 W%, 94 win pace) in normal games. Included in the starter category are shortened starts due to performance or injury, such as when Trevor Richards pitched multiple innings after Tyler Glasnow in September to cover an injury rehab shortened outing.
A 94 win floor with good starting pitching is good for annual playoff contention, but the bullpen games are the ones that pushed the Rays over the edge. Consider if underdeveloped Ryan Yarbrough, Yonny Chirinos, and Jalen Beeks, the primary “bulk guys,” started every game instead of facing the bottom half of the lineup to start. The intangible effect of overexposure surely would’ve led to more losses in the short term and the long term; the team would face an early deficit more often than if Romo, Ryne Stanek, Diego Castillo, etc. pitched the first inning, and the player would suffer a back and forth cycle from AAA instead of sticking in the big leagues as an effective pitcher. (In this piece, openers refer to pitchers who are normally relievers, as bulk guys refer to the long men in the rotation. All stats are with Rays, 5/19/2018 to present, including playoffs.)
A strategy created from necessity, the opener has served its purpose in successfully covering holes in the rotation. Considering the contracts given to top starting pitchers these days (Gerrit Cole, Stephen Strasburg, David Price…), the Rays better be able to develop their own if they want to win.
Now, a look at the individuals involved in the opener scheme: Over the last two years, the main openers were Stanek, Castillo, Hunter Wood, and Andrew Kittredge. Stanek and Wood are no longer on the roster, and Kittredge didn’t make the playoff rosters.
In real production stats, there is no correlation of performance to whether the opener is starting or relieving a game. Stanek had been significantly better as a starter than a reliever, Kittredge was better as a reliever, and Castillo and Wood were about the same at both aside from Wood’s astronomical 1.76 WHIP as a starter. However, every one of the four has experienced a generally substantial decrease in FIP and xFIP when serving as the opener. As a unit, their combined FIP when starting is 3.07 compared to 4.35 in relief. Bulk guys aside, Kevin Cash was getting the most out of his openers before even reaching the main part of the strategy. The team’s philosophy here is that pitching in the first inning is much lower in leverage than the 7th, 8th or 9th and that that is what helps young relievers grow as players.
Out of Chirinos, Yarbrough, and Beeks, they have generally been better as starters than relievers with Beeks being the exception, as he isn’t quite ready to be a regular member of the rotation. Chirinos had an extremely similar 3.71 ERA and 4.11 FIP as a starter compared to 3.72 and 4.09 when bulking. Yarbrough had a 3.83/1.21/3.99 ERA/WHIP/FIP as a reliever as opposed to 4.33/1.03/3.57 as a starter. The ERA was worse, but the WHIP was much better, indicating improvement from his more regular bulking days, which is reflected in a lower FIP. Beeks, starting more out of necessity than anything else, posted a 10.95/6.94/4.99 ERA/FIP/xFIP compared to 3.76/4.19/4.66 in his normal role as a long reliever or bulk guy. The three combined put up a 4.01 FIP when starting compared to 4.09 when following an opener or serving a rare middle relief purpose. The difference is minimal but there’s an intuitive reason why: the bulk role is meant to favor the pitcher, as he would only have to face the bottom of the lineup a third time, and when the bulk guy is starting it’s because he’s made improvements that make him worthy of winning a game on his own even without the added relief of skipping the top four hitters the first time around.
Because FIP is on the same scale as ERA (but independent of defense), we can use it to see how effective the average opener/bulk guy duo has been. As previously stated, openers had a combined 3.07 FIP alongside the 4.09 FIP of the bulk guys. In a generic game where the opener pitches one inning and following pitcher goes five, we can give a rough estimate of how many runs they might give up by pretending it’s ERA—FIP isolates the pitcher’s individual success without defense involved. By doing that, the pair gives up a combined 2.61 runs in 6 innings (3.92 ERA). In 2019, that’s most comparable to Madison Bumgarner, who had both a 3.90 ERA and FIP. Not bad for a would-be empty rotation spot.