Monday, January 28, 2013

Opening Day Starters

AP Photo/Al Behrman
Given a healthy rotation, every starter on a team will pitch about once every five games and all the starters will make approximately the same amount of starts. We tend to put a lot of thought into who a team's opening day starter is, often denoting them as the "Ace" of the staff, but from a statistical perspective, it doesn't really matter who starts the first game. If a team follows the rotation order perfectly it could make a difference of a an extra game around the all-star break or the end of the season, but there are likely to be more external factors which affect how often each starter takes the mound.

Nevertheless there is a certain amount of prestige in being named the opening day starter. It's the first big draw of the season and teams want to put out the best possible team out on the field. There are many reasons which one could be named the opening day starter, not all of them having to do with performance. Often the longest tenured member of the staff will get the opening day start as a token thank you, and to be the face of the franchise for the public.

If you were to ask me a year ago who the Jays opening day starter would be in 2013 I would have said Romero. He was one of the best pitchers on the team and had the best history with the franchise. Things can change quickly in a year though. Romero had a terrible season, ending in off-season shoulder injury, and the Jays pulled off a couple massive trades with the Mets and the Marlins to restock the rotation.

I don't think there's a default opening day starter on the Jays team, and really any member would be a reasonable choice. Romero has tenure, Morrow was the best Jays pitcher last year, Johnson and Buehrle have both piled up opening day starts as staff aces, and Dickey is coming off a Cy Young season. So it occurred to me that if Morrow or Dickey were to get the start, four out of the five members of the Jays rotation would be opening day starters at some point in their careers. This seemed like a lot and off the top of my head only the 2011 Phillies in recent memory would be able to match that feat. The Phillies actually topped it with five: Halladay, Lee, Oswalt, Blanton, and Contreras out of the pen. Hamels has yet to start an opening game.

With five previous opening day starters on the team I thought about what the most a team has ever had. To beat five, you probably need to have at least one starter coming out of the bullpen. 

This is a little more likely than you might think. For every Halladay and Verlander being put out by a team, there is a team just rolling the dice and hoping that at least one of their pitchers will still be in the majors by the end of the season. The last two seasons alone have seen the likes of Bruce Chen, Jeremy Guthrie, Kevin Correia, Luke Hochevar, Mike Pelfrey, and Tim Stauffer start opening day. It isn't just bad pitchers either but also players who have aged beyond their usefulness as starters. Carl Pavano, Ryan Dempster, Brett Myers, Fausto Carmona, and Livan Hernandez have all started opening day in the last two years and many of those have either already ended up in the bullpen or are clearly destined to show up there soon.

If we go back a decade to 2002 we see opening day starters, Chan Ho Park, Jeff Weaver, Jon Liever, Kevin Jarvis, Livan Hernandez, Ron Villone, and Tanyon Sturtze, who all ended up serving bullpen time later in their careers. So to make it simple, how does a team get more than 5 opening day starters on their roster? By filling their bullpen with ex-starters who either got too old, too injured, or were really never good enough to be starters in the first place.

What kind of a team loads up with players like this though? The obvious answer is great teams. They load up their rotation with at least four of these guys, and fill in their bullpen with a few more. But does this entirely make sense? If a team has a real ace they end up starting opening day for a long time. Half of Jack Morris' Hall of Fame case is his 14 opening day starts. So how do these guys all end up on one team. One possibility is that they aren't top quality starters any more and their original team has let them go. Players like that can be collected by anyone, particularly poor teams that could overpay slightly and offer the pitcher one last chance at glory. Let's see how well reality matches up.

Before we look into generalities let's take the teams with the most opening day starters as case studies. The most opening day starters on a team is eight, and the record is held by three teams.
  • The 1998 Red Sox are the most recent example. That team had Pedro Martinez, Tim Wakefield, and Bret Saberhagen as starters all season, Pete Schourek as a spot starter, and Dennis Eckersley, Greg Swindell, Carlos Reyes in a bullpen anchored by closer Tom Gordon. So how did this collection come together? Pedro was making his first of eight straight opening day starts while rotation mates Wakefield and Saberhagen had come to Boston a few years earlier to revive their careers. Tom Gordon started the season before, but after splitting duties in 1997, this was to be his first full year as closer. Eckersley was in the final year of his career and Swindell had moved to the bullpen in 1996. Schourek and Reyes were both midseason acquisitions to help the Red Sox pennant push. It didn't help as the Red Sox finished 92-70 and lost to the Indians in the ALDS.
  • The 1963 Cleveland Indians had Mudcat Grant, Dick Donovan, Jack Kralick, and Pedro Ramos in their rotation with former Opening Day starters Early Wynn, Jerry Walker, Jim Perry, and Gary Bell in the bullpen. Unlike the '98 Red Sox, this Indians team wasn't very good and were in the middle of a stretch of average finishes, including 79 wins this season. They were looking for any starter who could stick, which explains how five of their eight opening day starters had made their starts for the Indians. Grant made the 1963 start, Donovan made the 1962 start, Perry made the 1961 start, and Gary Bell made the 1959 and 1960 starts before being moved to the bullpen. Perry was also traded for 1962 Twins starter Jack Kralick in May. The Twins also provided Ramos who was traded to the Indians before the season. Early Wynn was 43 and the previous 6-time opening day starter moved to the bullpen for this, his final season. Jerry Walker previously started with the Orioles but was traded to the Indians and ended up in the bullpen after his 1962 season ended with a 5.90 ERA.
  • The 1947 Pittsburgh Pirates are the final team with eight opening day starters on their roster. Their rotation had Kirby Higbe, Fritz Ostermueller, Tiny Bonham, and Preacher Roe. Roger Wolff and Rip Sewell made spot starts while Jim Bagby and Hugh Mulcahy came out of the bullpen. The Pirates were not a good team (62-92) and looking for pitchers, which explains why the four guys I mentioned in their rotation were all 31 or older. In fact their opening day starter in 1947 (and 1943) was the 40 year old Sewell who spent half the season in the bullpen. Higbe and Mulcahy were both prior starters with the Phillies, although in '47 Higbe started 30 games while Mulcahy played in only 2 and was released in May. Fritz Ostermueller started opening day for the Pirates the previous two seasons, while Preacher Roe started in '44 but after a season with a 5.14 ERA in '46 didn't deserve the opening day start, which he continued to justify with a 5.52 ERA in '47. Tiny Bonham was the opening day starter for the 1943 Yankees and joined the Pirates in a trade in the offseason. Roger Wolff and Jim Bagby were purchased during and before the season respectively and both put up ERAs above 4.70. For both, 1947 would be their final season in the majors. Most of these players were washed up by this season which helps to explain the Pirates last place finish in the National League.
It's clear that opening day starters can come together in many ways. The three examples with the most feature three wildly different teams all with very different reasons for gathering the starters, and all having the starters at very different points in their careers. We can look to see if there is a relationship between the amount of opening day starters a team has and their success. The following graph shows the amount of wins teams have (scaled to a 162 game season) compared to the number of opening day starters they have.

Wins given the number of Opening Day starters on teams' rosters
There is a definite upwards trend in terms of number of wins. A team with only one opening day starter wins on average only 74.6 wins while a team with five opening day starters wins on average 87.4 games. Above five starters the trend is less obvious, for one for the various reasons that a team may acquire so many starters as illustrated by the example above, but also due to the small sample size. The following table shows how many teams fit into each bin for the number of opening day starters. The amount of data available clearly becomes more limited with six or more starters.

Number of Opening Day StartersNumber of Teams
8
3
7
14
6
41
5
133
4
203
3
298
2
280
1
111

It's definitely important not to mix up correlation with causation. Although having more Opening Day starters may make a team better, especially if they are acquired in their prime, a team often has more Opening Day starters because they are already a good team and looking for an extra player or two to push them over. I think this is more of an interest piece than looking for any real cause-and-effect algorithm for wins. The way teams are constructed and the changing roles of players as they age and change teams provides some insight into the history of baseball. The amount of data for the higher number starter sets should also increase now given the greater number of teams and more liberal use of relievers. In studying all the teams since 1901, 22.4% of the teams with 6 or more opening day starters on their roster occurred in 2000 or later.

No comments:

Post a Comment