clock menu more-arrow no yes mobile

Filed under:

Shortened Schedule and Competitive Balance

Commissioner Rob Manfred has again expressed willingness to change the rules of Major League Baseball. A shortened schedule would have many cascading effects, but what would it do to competitive balance and the role of random chance?

Evan Habeeb-USA TODAY Sports

Rob Manfred lit fire to the Internet again when he expressed willingness to make changes to the MLB rulebook. In a conversation with ESPN.com’s Darren Rovell, Commissioner Manfred said that he would be open to discussing shortening the regular season back to 154 games. Many have already discussed potential impacts of a slightly shortened season. Earlier this week Chris Teeter, who is in favor of such a change, wrote about the potential benefits. Teeter wrote that a shorter season could lead to fewer cold weather games, reduced injury risk, and a more balanced and logical schedule. A shorter regular season also leaves more to random chance. With some slight modifications to previous code, a simulation can be run to compare competitive balance with a 162-game and with a 154-game schedule.

Set Up

For the particulars of the simulation, you can refer to the original article here. In summary, the simulation uses a normal distribution to generate a run scored and runs allowed value for a particular game given one team’s runs scored and runs allowed talent levels. This means the wins of one team are unaffected by the wins of another. Previously this simulation was used to determine how often the most talented team in baseball had the best record in the majors at varying season lengths. The simulation is not the most in depth by any stretch, but for this sort of thing it works great (and is reasonably fast).

This time, instead of looking at how often the most talented team finishes with the best record I will be looking at the balance of a particular division. For each simulated season I will keep track of the division winner. As such, the simulation will have one division of 5 teams.

Team

Runs Scored Talent Level (RS/G)

Runs Allowed Talent Level (RA/G)

Differential

Team Very Good

4.4

3.6

0.8

Team Pretty Good

4.2

3.8

0.4

Team Average

4.0

4.0

0.0

Team Pretty Bad

3.8

4.2

-0.4

Team Very Bad

3.6

4.4

-0.8

Each time we run a season we will keep track of who tallied the most wins and award them the division title. We’ll play 100,000 seasons with a 162-game schedule, 100,000 seasons with a 154-game schedule.

Results

As I wrote in October, much of the regular season is left to chance, and it would take a lot more than eight games to swing the pendulum much one way or the other. First we will look at how often each team won the division title under the standard 162-game schedule. This result alone is fairly interesting in and of itself. We find Team Very Good to walk away with a division title 68.0% of the time. Team Pretty Good captured the automatic playoff berth 23.9% of the time and Team Average won the division 6.5% of the time. The full results are summarized in this table:

Team Very Good

Team Pretty Good

Team Average

Team Pretty Bad

Team Very Bad

68.04%

23.89%

6.52%

1.35%

0.19%

Reducing the season length allows random variation to play a larger role. We would expect in a shorter season Team Very Good would not win quite as easily with all the less teams seeing varying rises in their playoff chances. The simulation confirms our suspicion, though Team Very Good’s playoff odds only decreased by .7 percentage points. That share of titles is then distributed amongst the rest of the division with Team Pretty Good and Team Average capturing the lion’s share ticking up .2 percentage points and .29 percentage points each.

Team Very Good

Team Pretty Good

Team Average

Team Pretty Bad

Team Very Bad

67.34%

24.09%

6.81%

1.51 %

0.25%

For fun we can also look at much shorter season lengths, like the 140-game schedule of the early 20th century.

Team Very Good

Team Pretty Good

Team Average

Team Pretty Bad

Team Very Bad

65.87%

24.42%

7.61%

1.79 %

0.32%

What if the season was cut in half and MLB instituted an 81-game schedule similar in length to the NBA season.

Team Very Good

Team Pretty Good

Team Average

Team Pretty Bad

Team Very Bad

58.51%

26.17%

10.49%

3.81%

1.03%

Now we start to see significant changes, with Team "Very Good" dropping nearly 10 percentage points from the 162-game schedule. Finally what if the MLB decided to play once a week and instituted a 16-game NFL-like regular season.

Team Very Good

Team Pretty Good

Team Average

Team Pretty Bad

Team Very Bad

44.27%

26.07%

15.25%

9.17%

5.24%

Things start to get crazy here, with even Team "Very Bad" having a five percent shot at the division title!

Conclusions

Commissioner Manfred has made it clear a shorter schedule is not something MLB’s New York offices are currently prioritizing. As Commissioner Manfred continues to express openness to change we should continue to empirically explore whether or not these changes will accomplish what they set out to do. Further, we should consider what unintended consequences these changes might have. In the case of a shorter season and competitive balance, the change is small and cannot be characterized as good or bad. Ultimately the question of shortening the season will be decided by economic forces, not fractions of expected playoff percentages. Either way, it appears an 8-game reduction would have minimal effects on competitive balance, so we may proceed to consider the other pros and cons of returning to a 154-game schedule.

Simulation was built by me in Python and is based off of run total data from Retrosheet.org

Daniel is a junior at Colby College and contributor to Beyond the Box Score and will be presenting at the upcoming SABR Conference. You can follow him on twitter @dtrain_meyer.