Does starting pitching make or break a team?

I tip my hat to Jim McLennan who inspired this piece with his round table question, “The D-backs have used 16 different starters this year. What does that tell you?”

Does the number of starters matter? Is there a characteristic of pitching that separates winning teams from losing teams? Let’s look at a few possibilities.

Did teams that used a lot of starters lose more games? Through 31 August, the Diamondbacks used 16 starters. Only 3 teams used more pitchers through 31 August: the Dodgers, the Marlins, and the Mets. The Marlins are in last place in the NL East and the Dodgers are challenging the Giants for first place in the NL West. Using a lot of starters did not necessarily mean a team lost a lot of games.

Do quality starts explain which teams won more games? Let’s look at the percentage of each team’s wins that were quality starts and compare it to the percentage of games each team won. The following scatterplot shows the comparison for all 30 teams. Unexpectedly, there was no significant relationship.

2021 season. Games through 31 August. Data from Baseball Reference.

Because quality starts did not explain team winning percentages, let’s look at relief pitching. The previous graph was modified to mark with purple the 13 teams with the best relief pitchers based on Wins Above Average (WAA).

2021 season. Games through 31 August. Data from Baseball Reference.

Two observations:

• Ten teams won 54% or more of their games: 9 were marked purple as having the best relief pitching. The Astros were the lone exception – another reason to wonder about the Astros.
• The 13 teams with the best relief pitching all won more than 50% of their games.

Much to my surprise, there was a strong link. Although that strong link does not necessarily mean causation, it means that successful teams have above average bullpens (with rare exceptions like the Astros).

Did it make a difference if the starter left the game with his team ahead? The following scatterplot shows the comparison of each team’s winning percentage and the team’s percentage of games that the starter left the game with the possibility to be credited with the win. There was a significant relationship.

2021 season. Games through 31 August. Data from Baseball Reference.

Finally, the data showed something that might be expected. Being ahead after 4, 5, 6, 7 or 8 innings increased the odds of winning the game. That was certainly correct, but odds don’t tell the whole story.

Perhaps the managers used their better bullpen pitchers when their team was ahead. On the other hand, managers rely heavily on other considerations, such as matchups, how rested is each bullpen pitcher, and roles of each bullpen pitcher.

Perhaps the bullpen pitchers pitched better knowing their team was ahead. Possible explanations for better pitching are they expected a good outcome (winning the game) or they pitched boldy with boosted confidence.

Runs scored by the pitching team’s offense were important in determining whether the starter left the game with the possibility of being credited with a win. The exception was when a starter pitched a shutout. This season through 3 September, starters shutout other teams in half a percent of their starts, which was not enough to reduce the importance of scoring runs.

The Diamondback starters were better. They shutout their opponents in 2.9% of their starts (which was 4 starts). Zac Gallen left the game after 5 innings, while Luke Weaver, Madison Bumgarner, and Tyler Gilbert left the game after 7 innings. The Diamondbacks won all four games.

Summary. My conclusion is that more important than quality start percentage, is whether the team was ahead when their starter left the game, and whether the bullpen was above average.

Addressing the roundtable question, it was relatively unimportant how many starters the Diamondbacks used because starting pitching does not break or make a team. The bullpen and the offense were equally if not more important in determining team winning percentage.