April 21, 2021

The Three-Batter Rule – There is a Better Way

April 21, 2020 by · 3 Comments 

When baseball resumes, the three-batter minimum rule for pitchers will be implemented. Under this rule, relief pitchers would be required to face at least three batters or to complete the half-inning before they could be taken out. The rule appears to be an attempt to address the related problems of pace of play and length of games. The notion is that the parade of relief pitchers in the late innings slows games down and lengthens them, causing fans to lose interest. Reactions to the rule have been less than enthusiastic.

Managers have had a mixed reaction. Joe Maddon, Joe Girardi, and Ron Gardenhire expressed displeasure, mainly on the grounds that it would detract from the strategic element of the game. “Strategy is sacred,” according to Maddon, and should not be interfered with. Other managers’ reactions ranged from uncertainty about its implications (e.g., Dave Martinez, Derek Shelton) to doubt that it would affect them much (e.g., Alex Cora).

Analysts thus far have agreed that it probably won’t matter a great deal. (E.g., Ben Clemens, https://blogs.fangraphs.com/the-three-batter-minimum-barely-matters/, December 18, 2019). Among other points, they have noted that the role of LOOGY’s and other one-out relief specialists has been declining and doesn’t figure into an appreciable number of appearances, that it will shave little time off the length of games, and that it won’t materially affect strategy and game results.

Prompted by the impending rule, I examined data on number of pitchers per game and time of games per nine innings, and have found that there is a strong and interesting relationship between the two. Here is what the data show since World War II – what I think is fair to consider the modern era of baseball:

The data are from Baseball-Reference.com.
Note that I have converted the minutes in Time/9I to a percentage of 60 minutes.


Both the number of pitchers used and the time of games have increased steadily over this period. At the outset of the era, about two pitchers were used on average per game, and the average game length was two hours or a little over. The averages climbed in synch to about two and three-quarter pitchers and hours in the late 1980’s. In 1989, an interesting divergence occurs. The number of pitchers per game jumped and increased more steeply. The time of games continued to rise, but not as sharply. This coincides with the advent of what I have called the LaRussa Orthodoxy – the replacement of the fireman model with the use of multiple relievers, match-up focused, and closer-driven. (Mark Pelesh, https://tht.fangraphs.com/ending-the-larussa-orthodoxy-part-1/, August 28, 2019; https://tht.fangraphs.com/ending-the-larussa-orthodoxy-part-2/, September 4, 2019). By the late 20-teens, the number of pitchers per game was over four and heading up, and the time of games was three-plus hours.

In light of these data and trends, it is unsurprising that Commissioner Manfred would look for something on pitcher usage to arrest the upward climb in the length of games.  While length of games and pace of play are not quite the same thing, the delays attendant to pitching changes – visits to the mound, entry of relievers, warmup pitches – affect both. Reducing those could move the game back toward when the games lasted less than three hours and must have had a crisper feel. But is the three-batter rule the right solution?

I submit that the rule is a blunt and misdirected way to try to get at the problem. As Ben Clemens stated in his post, the rule is “inelegant” and will likely confer a narrow benefit – “30 seconds and some tiny sliver of runs per game,” according to Clemens. I propose that Major League Baseball instead look to a rule used in another sport to address the issue. In soccer, teams are restricted to a limited number of substitutions per game – usually three with possibly one additional sub if games go into extra time. MLB could adapt this rule and restrict the number of relief pitchers used per game. There could be a limit for a normal nine inning game with additional relievers allowed for extra innings. Exceptions to the limits could be permitted for (genuine) injuries. Provided that teams stayed within the limit on number of relief pitchers, they could pitch for as little or as long as the manager wished.

Instead of a kludgy rule restricting the number of batters faced, a rule addressing the number of pitchers allowed per game is more likely to reduce length and increase pace of play. The data I have assembled support drawing upon the soccer substitution rule and suggest what the baseball limits should be. The data show that holding the number of pitchers to three or fewer per game strongly correlates to games under three hours. Accordingly, the soccer-inspired rule for baseball would allow for three pitchers per game – a starting pitcher and two relievers. If the game goes into extra innings, a team would get an additional relief pitcher for each three-inning increment. Again, injuries would not count against the limits. And for a starting pitcher who gets shelled and is taken out in the early innings (perhaps defined as before the game becomes official) a replacement for the starter could be permitted and not count against the three-pitcher limit. If a two-reliever limit per nine innings is too abrupt a change, MLB could consider a three-reliever limit for a period of time to allow for a transition. While the data don’t indicate that this would significantly reduce game length, it would arrest the upward trend in time of games per nine innings and eliminate the games when long lists of relievers are used.

What about the objections of Joe Maddon and others that a rule restricting pitcher usage would detract from the strategy managers can employ and thereby lessen interest? That objection, I believe, should not apply to the rule I am proposing. We can look to the sport that inspires the proposed rule for the reasons why. In soccer, when a manager utilizes his allowed substitutions and who he substitutes are key strategic decisions. They are made based on the state of the game and the team’s needs as the game evolves. Each one is made knowing that the substitutions are scarce assets and have to be deployed for maximum effect. In baseball, with a limit on the number of relief pitchers allowed, a manager would be faced with similar strategic decisions. He might still bring in a pitcher just to face one batter if the game circumstances warranted and the risk/reward balance made it worthwhile. With a one-run lead in the seventh or eighth inning, the bases loaded, and Bryce Harper up, for example, an opposing manager might well bring in his lefty relief ace even though it would burn one of his allowed pitchers. The need for strategic judgment would be enhanced compared to the frequent push-button matchup managing that often now occurs. Almost every decision to use a relief pitcher would become a source of discussion and debate for commentators, fans, and pundits.

I acknowledge that this proposal is probably not a cure-all for the problem of length of games and pace of play. Other factors are likely contributors, especially hitters who habitually step out of the batter’s box and repeat their set-up routine after every pitch – a phenomenon also of recent vintage (take a look at replays of games as recently as the 1970’s and ‘80’s). But a rule limiting the number of relief pitchers allowed would likely help a great deal and make baseball even more intriguing.


3 Responses to “The Three-Batter Rule – There is a Better Way”
  1. Mike B. says:

    Great article – very well-written. Here’s my only problem… I completely disagree. I have been in favor of the three-batter rule for years.

    In my view, the issue is not the number of pitchers used per game. After all, if a team used one pitcher per inning that would mean a team used nine pitchers in a game. That’s a ton of pitchers, but if each of them started an inning the game wouldn’t slow down at all and pace of play would be completely unaffected.

    The issue is pitching changes during a half-inning. That is what the rule is trying to limit – the manager’s slow walk out to the mound, the waiting at the mound for the reliever to jog in, the discussion when the new pitcher arrives, the eight warm-up pitches, etc. And if it’s a lefty-lefty match-up, then the whole sequence repeats if the next batter is a righty. That is what slows the game down.

    I simply don’t agree that this rule won’t help speed up the game.

    It’s not a cure-all, but it would help. I would also make these changes:

    • Get rid of instant replay. Nothing slows the game down more – and most of the time the umps are right.
    • Only three warm-up pitches when a pitcher comes out to inning, and when a reliever comes in.
    • Strict limits on a batter stepping out of the box.
    • Call the strike zone in the rule book! A pitch at the letters is a strike!

    Thanks for posting this. It’s great to talk some baseball, even if we cannot watch any!

  2. Mike Lynch says:

    I’m a proponent of fewer warm-up pitches as well. As someone who pitched in 20+, 30+, and 40+ leagues I hated warm-up pitches between innings and why a pitcher needs seven per inning is beyond me. Anyone who’s able to go nine innings–rare these days, I know–ends up with 63 extra pitches on their arm because of warm-ups. Even a hurler who gets through four innings will have 35 extra throws before he starts the fifth.

  3. Jerry Murphy says:

    Excellent idea and very typical of your knack for preparation … meaningful data. I like your idea and will expand on it later. However, allow me to briefly reflect on one of the most important aspects of the national past time … the importance of all players being able to play offense and defense … attack and defend. The need to go both ways exists in soccer (world football), hockey and basketball. As a Boomer, I’m a traditionalist and remain opposed to the designated hitter. Yes, some positions/players are in the line-up solely for their defensive position/abilities (Nolan Ryan) but in the National League, they still have an offensive role that influences the game’s strategy. Unfortunately, in recent history fewer pitchers can lay down a good bunt. This is a function of greater specializations which is exacerbated in relief pitchers who play virtually little to no offensive role. Seeing your data and graph reminded me of the days of Elroy Face, the great Pirate reliever in the late 50’s through the 60’s during the two pitcher era. No one talked about the length of the game back then.

    My fundamental problem with three hitter or half inning requirement is that it can and will change the outcome of the game only for the sake of timing. A reliever could be getting hammered in a close game that turns into a blow-out due to the three hitter requirement e.g. enters the game with the bases loaded and no outs. Not good for baseball.

    I like limiting the number of changes but would expand it to limiting the total number of “substitutions” or line up changes; apply it to pinch hitting, position substitutions and relief pitchers. Like soccer, the flow of the game will determine which position, how many and who you use which will add a major strategic element to the game. It may speed the pace of the game by timing changes and sharing a limited number of changes.

    Better use and promotion/visibility of the timing between pitches would add home fan engagement like the 24 second clock in basketball. Pitch timing is intended to help speed the game, so improve it. Getting input from pitchers to determine the ideal time (the second that collectively raises their anxiety!!!) could identify the most appropriate length of the pause.

    Be well and be safe.

Speak Your Mind

Tell us what you're thinking...
and oh, if you want a pic to show with your comment, go get a gravatar!