Avoiding the disaster start

One key to winning is not immediately giving the game away. Who knew?

After giving the Royals some decent seasons from 2012-2014, Jeremy Guthrie was not good in 2015. He pitched to a 5.95 ERA (5.61 FIP) and led the league in home runs allowed, despite pitching in spacious Kauffman Stadium and losing his rotation spot after just 24 starts. Yet the Royals managed to go 14-10 in those starts, a .583 winning percentage that was just about the same as the team’s record on the season.

Now that was a good team with an above-average offense and an outstanding bullpen, so he had support. But was there a reason the Royals were able to win with Guthrie on the mound, even though he rarely pitched particularly well? Yes. Despite rarely winning the game by dominating the other team, Guthrie rarely gave the game away with an outing I have labeled the “disaster start.”

The Disaster Start

For the purpose of categorizing each start, a “disaster start” is defined as an outing where the starting pitcher goes less than five innings and gives up more runs than he has innings pitched; or gives up at least eight runs, regardless of innings pitched. Guthrie only had three such starts in 2015, a total already matched this year by both Jordan Lyles (who was supposed to be a Guthrie-type presence) and Brady Singer.

(I should note that this is a hindsight-only, purely results-based stat, not process-based or predictive, so Brad Keller’s last outing where he allowed 8 walks in 3.2 innings does not qualify, because he only allowed three runs. I also used runs, rather than earned runs.)

Teams rarely win when a starter has a disaster start. I don’t have the research tools to run this on a league-wide level (I’ve done the three seasons featured here by hand), but the Royals over the past two seasons have had 37 disaster starts and are 2-35 in those games. Even the 2015 Royals were only 1-16 when their starter turned in a disaster start.

Categorizing other starts

Some time ago, the statistic “quality start” gained popularity, defined as a pitcher completing at least six innings and giving up three or fewer runs. With starters not going as deep into games, I thought it would be good to redefine these parameters and also add a category to correspond to “disaster start” on the other end of the spectrum.

Here are the categories and definitions:

Disaster start: less than 5 IP, runs allowed > innings pitched OR at least 8 runs allowed

Weak start: any non-disaster start that does not qualify as a strong start

Strong start: Innings pitched exceeds runs allowed by at least three

Outstanding start: Innings pitched exceeds runs allowed by at least five

Looking at starts this way tells a story that’s different from what combined stats like ERA or FIP or K-BB% can tell. Those speak to pitchers’ overall effectiveness and can help you see where their performance is headed. Looking at the starts in categories tells you how often the starter is making it possible to win and how often he is (more or less) winning the game on his own. Over the past two seasons, the Royals are 2-35 in disaster starts and 37-10 in outstanding starts.

Last year, the Royals with the most disaster starts were Brad Keller and Kris Bubic, with six each. Keller was the most likely to have a disaster start (27.3%), followed closely by Jon Heasley (23.8%) and Bubic (22.2%).

Brady Singer led the team in outstanding starts with ten, followed by Zack Greinke and Brad Keller with eight each. Singer was the most likely to turn in a gem (41.1% of his starts were outstanding starts), followed by Keller (36.4%) and Greinke (30.8%). The 2022 Royals went 33-7 when their starters had outstanding starts.

As you can see, Keller had very little mediocrity last year. He was really good or really bad. In 14 of his 22 starts, it was either “outstanding” or “disaster.”

Here are last year’s starters:

Last year’s team had a disaster start 28 times, 17.3% of their starts. This year’s staff has been even worse, with 9 disasters in 45 starts, 20% of their starts. Each of their top five starters (Greinke, Lyles, Keller, Singer, and Ryan Yarbrough) has at least one, and Lyles and Singer have three apiece.

As you can see from last year’s chart above, Singer had no such starts last year. And Lyles has already matched his 2022 total of three disaster starts.

The Royals are 0-9 in disaster starts. They have a winning record when the starter provides a strong start or outstanding start. Here is this year’s chart:

When the Royals used an opener, I combined the opener with the bulk innings guy to call it a “start.” This year, that was Taylor Clarke/Yarbrough, Josh Taylor/Max Castillo, and Carlos Hernandez/Mike Mayers.

Over the past two years, the opener/bulk strategy has produced no disaster starts, two strong starts, one outstanding start (Gabe Speier and Joel Payamps last year), and only one weak start. So it could be that using this strategy could help the Royals avoid disaster more often.

Conclusions

The Royals’ offense and bullpen has been coming around. This month, the starting pitching has too often given the game away before the offense or bullpen could make the difference. For the 2022 Royals, it took an outstanding start to for the team to do better than a 100-loss pace (they were only 12-22 after “strong starts,” compared to 7-7 so far this year). This year’s team should be able to maintain a decent record if the starters would just not give the game away.

But Singer and Lyles have already doubled their combined number of Disaster Starts from last season. For Lyles, that tells us that he’s not doing the one thing he was brought in to do: provide innings and keep the game close. As for Singer, I am hopeful that those disasters are behind him and we can see him pitch with more consistency.

As the Royals work on developing pitching, we often recall that the 2015 Royals were not known for their starting pitching. But the starters were really good about getting the game to the bullpen and allowing their offense to keep scoring while the bullpen held the other team down. I will finish this article by posting the same chart for the 2015 Royals: