By Chris Weikel
The penalty for using a pitcher for the third time through the order has been well-documented in recent years, influencing team strategies and redefining the starting pitcher’s role as we know it. The gist of this phenomenon is that pitcher production nosedives from a combination of fatigue and hitters becoming accustomed to his offerings and sequencing. As a result, starting pitchers have a shorter and shorter leash as the bullpen has a greater responsibility, leading to 13-man pens. Despite how much this trend has transformed the game, it’s interesting to note that in the public sphere there has been little investigation of what type of pitchers are most and least affected by the third-time-through-the-order penalty. If a certain type of pitcher can go deeper, this information would create a competitive edge for teams. The goal of my analysis is to identify that type of pitcher and help give teams that edge.
The first factor I will analyze is pitcher style. While each pitcher is unique and has a mixture of traits, the industry for decades has tended to classify players as either finesse or power. While there is no exact measurement for style, I have created my own set of qualifications to separate players. The first statistic I used to identify a player as a finesse pitcher was Baseball Prospectus’ command score. This metric takes advantage of BP’s called strikes above average stat, originally intended for catcher framing. They use this to isolate how often a pitcher steals strikes at the corners and penalizes players who miss by wide margins (More information can be found here). Overall, it works as a way to score players who nibble on the edge, instead of attacking hitters with pure stuff, a critically important aspect of a finesse pitcher. Below, I ran a simple linear regression of the command score and OPS drop-off from first time through the order to third. My sample for this data included pitchers who have reached the third time through the order 20 times during the last two seasons, as I wanted to capture the current baseball environment.
Unfortunately, there is almost no correlation between the two variables with an R squared at basically zero and a pretty flat slope. But this is not the only aspect of a finesse pitcher. The second variable I examined was fastball velocity. One of the common aspects of a finesse pitcher is the lack of top-end velocity, forcing them to live on the edges. Below is the linear regression.
Again, there is little to no correlation. There is more here than there was for command, but there is still nothing statistically significant.
Then I decided to combine both skills (command and velocity) as a measurement for finesse vs power. I divided the players into two groups for comparison. For my Finesse group, I used players who were above the 3rd quartile of command and below the 40 percentile in velocity as it’s not as tight of a requirement and I wanted to enlarge my sample. For the Power group, I used the same cutoffs, but below Q1 for command and above 60th percentile for velocity. Here is the list this generated, command first then power.
The sample size isn’t huge but when analyzed, it did result in some interesting findings.
The mean for the Finesse group’s OPS against came in at an increase of .081, while the Power group had a much higher .118, both larger than the total groups’ .074 OPS increase. A few important details can be extracted from these numbers. First of all, the Power group is clearly more affected by the third time through the order penalty. This is likely due to their inevitable decrease in velocity by the end of the game affecting them much more than the other group. Because these pitchers rely so heavily on their pure stuff, they struggle to battle through later at-bats once it fades. They do not possess the ability to consistently get free strikes by painting on the corner so late in the game, when batters get used to their one-dimensional attacks. The Finesse group, on the other hand, does possess this ability, and a well-rounded arsenal often comes with this “pitchability.” This allows them to fare better late in games as they have more ways to get a batter out. The downside, however, is that they don’t have the plus fastball that can often bail a pitcher out when he gets fatigued and starts losing command of the zone. Here is a plot of the two groups laid on top of each other. To find a particular player, match the ID number to the horizontal axis.
One anecdote from the graph is that the Finesse group had two clear outliers, which skewed its data up a bit, showing that, as a whole, they were better than the mean portrays (Hellickson and Odorizzi).
Perhaps the most intriguing takeaway from my analysis is that the players that fall into neither of these categories are the far more superior group. This group is made up of multiple types of pitchers so I believe there are many reasons for this. The first player type that is in this group is the guy who possesses average velocity and average command. These pitchers have the best of both groups and feature many different styles. They can command the zone relatively well, or at least enough to not fully rely on big fastballs but still have average velocity. This can be a deadly combo late in games due to the fact that the batter can’t as easily lock in on what the pitcher is looking to do with his offerings. The other two groups in this sample would logically even out as they are the high command, high velocity group (basically the best pitchers) and the low command, low velocity group (the worst pitchers) but you must remember both the MLB and this data are skewed samples. MLB teams are “usually” (discussion for a different day) trying to win games and therefore would not roster most of these low command/low velocity guys because, well they’re terrible. On top of this, the data I used includes only players who made it to the third time of the order 20 times over the last two years, and with the current, short-leash environment, the odds that any of these pitchers make it that far in the game is getting increasingly lower by the year. This causes the high/high group to outnumber the poor group by quite a bit. These players are some of the best, most well-rounded pitchers in the game so they are perfectly suited to flip over the lineup card three times. They are what you look for in a pitcher and feature the lowest drop-off during the third time through the order.
The overall conclusions that can be reached from this analysis is the obvious: You want pitchers with good control and velocity on your team. But there’s more than just that. The prototypical velocity with limited control arm that seems to be getting more popular is by far the worst option to go deep in games. This popularity could be part of the reason why the third time though the order penalty is becoming more important and starters are going much shorter. If a team wishes to offset this idea and target a strategy that goes against the flow of the league, this data shows a versatile skill set is the way to go. While still seeing a drop-off in production, these players are significantly better than their extreme counterparts in this regard. A staff built around these guys would not only relieve pressure from the bullpen, but could also shift a team to a 12-man staff, creating more offensive versatility and options for a manager.