README.md

In Tylermarshall21/AstrosCheating: Analyzing the Houston Astros Cheating Scandal

AstrosCheating

The goal of AstrosCheating is to give insight into the Houston Astros 2017 cheating scandal by providing data and sample analysis regarding the situtation.

Installation

You can install the released version of AstrosCheating from CRAN with:

install.packages("AstrosCheating")

And the development version from GitHub with:

# install.packages("devtools")
devtools::install_github("Tylermarshall21/AstrosCheating")

Example Analysis

This is example analysis that can be done with this package.

library(AstrosCheating)

The first thing I would like to look into further is to see when the cheating began for the Astros. By looking at the total number of bangs by game, it should be clear when the Astros really started cheating.

#Find the number of bangs by game
date_bangs <- AstrosCheating %>%
  group_by(game_date) %>%
  filter(has_bangs == "y") %>%
  mutate(bangs = length(game_date)) %>%
  arrange(bangs)

#Plot the number of bangs by game
ggplot(data = date_bangs, aes(x = game_date, y = bangs)) +
  geom_point() +
  geom_smooth(se =  FALSE) + 
  theme_minimal() +
  labs(title = "Number of Trash Can Bangs by Game", x = "Date of Game",
       y = "Number of Bangs")

From this, we can see that the Astros did not start employing this cheating strategy until the middle of the season. There were very few bangs on the trash can during April’s home games, but it looks like the cheating began at the end of May. On May 28th, 2017, the Astros had 28 bangs on the trash can which seemed to be the real start of their cheating strategy. There is still some variation in the number of bangs per game, but it becomes clear that the Astros were cheating for the rest of the season. However, it is interesting that the Astros seemed to quit cheating in the final two home games of the regular season. The Astros had already clinched the division at this point, so that could explain why they did not feel the need to cheat, but it is definitely interesting that they stopped banging on trash cans in those last two home games.

The next thing I would like to look into further is to see which hitters were most involved in the cheating strategy. The Astros did not do this all the time, and hitters were involved to a varying degree, so I’d like to know more about who cheated most often.

#Find total number of bangs for each player
player_bangs <- AstrosCheating %>%
  group_by(batter) %>%
  filter(has_bangs == "y") %>%
  mutate(bangs = length(batter)) %>%
  select(batter, bangs) %>%
  distinct() %>%
  arrange(desc(bangs)) %>%
  rename(`Batter Name` = batter, `Pitches with Bangs` = bangs)
#Create table to display results
knitr::kable(player_bangs) %>%
  kable_styling(latex_options = c("striped", "HOLD_position"))

Batter Name Pitches with Bangs Marwin Gonzalez 147 George Springer 139 Carlos Beltran 138 Alex Bregman 133 Yuli Gurriel 120 Carlos Correa 97 Jake Marisnick 83 Evan Gattis 71 Brian McCann 45 Josh Reddick 28 Tyler White 28 Jose Altuve 24 Norichika Aoki 16 Derek Fisher 16 J.D. Davis 14 Juan Centeno 13 Max Stassi 13 Cameron Maybin 13 AJ Reed 4

From this table, we see that Marwin Gonzalez, George Springer, Carlos Beltran, Alex Bregman, and Yuli Gurriel were most frequently the beneficiaries of trash can bangs with all of them having at least 120 pitches where they received a trash can bang before the pitch. While this shows the total number of pitches where a bang occurred, it is a little misleading because hitters that see more pitches would be more likely to have more bangs. I would also like to look at a percentage of pitches where the hitters received a bang to see if the results are similar.

#Find the percentage of bangs for each player
percentage_bangs <- AstrosCheating %>%
  group_by(batter) %>%
  mutate(total_pitches = length(batter)) %>%
  filter(has_bangs == "y") %>%
  mutate(total_bangs = length(batter),
         percent_bangs = round((total_bangs/total_pitches)*100,3)) %>%
  select(batter, total_bangs, total_pitches, percent_bangs) %>%
  arrange(desc(percent_bangs)) %>%
  distinct() %>%
  rename(`Batter Name` = batter, `Total Bangs` = total_bangs, 
         `Total Pitches` = total_pitches, 
         `Percentage of Pitches with Bangs` = percent_bangs)
#Create table to display results
knitr::kable(percentage_bangs) %>%
  kable_styling(latex_options = c("striped", "HOLD_position"))

Batter Name Total Bangs Total Pitches Percentage of Pitches with Bangs J.D. Davis 14 49 28.571 Tyler White 28 106 26.415 Max Stassi 13 52 25.000 Cameron Maybin 13 56 23.214 Jake Marisnick 83 364 22.802 Juan Centeno 13 65 20.000 Marwin Gonzalez 147 776 18.943 Carlos Beltran 138 762 18.110 Yuli Gurriel 120 670 17.910 Evan Gattis 71 427 16.628 Alex Bregman 133 800 16.625 Carlos Correa 97 594 16.330 George Springer 139 933 14.898 AJ Reed 4 27 14.815 Brian McCann 45 507 8.876 Derek Fisher 16 211 7.583 Norichika Aoki 16 261 6.130 Josh Reddick 28 725 3.862 Jose Altuve 24 866 2.771

When looking at percentage of pitches where the hitter received a bang, there are some different names at the top of the list. J.D Davis, Tyler White, and Max Stassi had the highest percentage of bangs when they were hitting, but were not everyday players which is why they weren’t at the top of the total bangs table created above. Also, none of these three players were on the opening day roster for the Astros which could explain why their percentages were higher than the typical starters. Based on the analysis above, it seems the Astros did not start cheating until May 28th, 2020 which could be part of the reason that the starters have lower percentages compared to some of the backups that were not on the roster to start the season. Another interesting observation from this table is that Tony Kemp was the only player to not receive any bangs which suggests he was not a part of the cheating scandal. Josh Reddick and Jose Altuve who started almost every game for the Astros very rarely received bangs either which indicates they were not very involved in the scandal either.

The next part of the scandal that I would like to explore further is when the banging on trash can occurred during the game. I am interested to see if there are any patterns in the innings in which the Astros did this.

#Number of bangs by inning
inning_bangs <- AstrosCheating %>%
  group_by(inning) %>%
  filter(has_bangs == "y") %>%
  summarize(n = n())
#Plot Number of bangs by Inning
ggplot(data = inning_bangs, aes(x = inning, y = n)) +
  geom_line() +
  theme_minimal() +
  labs(title = "Number of Trash Can Bangs by Inning", x = "Inning",
       y = "Number of Bangs")

From the plot above, we see that the number of bangs are lowest in the tenth and eleven innings. This makes sense given that the game typically ends after nine innings unless the score is tied. Since the Astros were only cheating at home, they also would not bat in the bottom of the ninth unless the score was tied or they were losing, so this explains why there were fewer bangs in the ninth inning compared to the previous eight. The number of bangs is lowering in the first inning which I am a little curious about. It is not exactly clear why there are fewer bangs in the first inning compared to the rest of the game, so I would like to know the reason for this.

Finally, I would like to explore what pitches the Astros banged on the trash can for. Banging on the trash can had to be associated with a specific pitch type to help the hitter, so I will explore what pitches they used banging on the trash can to identify.

pitch_bangs <- AstrosCheating %>%
  group_by(pitch_category) %>%
  mutate(total_pitches_pitch = length(pitch_category)) %>%
  filter(has_bangs == "y") %>%
  mutate(total_bangs_pitch = length(pitch_category),
         percent_bangs_pitch = round((total_bangs_pitch/total_pitches_pitch)*100,3)) %>%
  select(pitch_category, total_bangs_pitch, total_pitches_pitch, percent_bangs_pitch) %>%
  distinct() %>%
  rename(`Pitch Category` = pitch_category, `Total Bangs by Category` = total_bangs_pitch,
         `Total Pitches by Category` = total_pitches_pitch,
         `Percentage of Pitches with Bangs` = percent_bangs_pitch)

#Create table to display results
knitr::kable(pitch_bangs) %>%
  kable_styling(latex_options = c("striped", "HOLD_position"))

Pitch Category Total Bangs by Category Total Pitches by Category Percentage of Pitches with Bangs BR 702 2518 27.879 FB 233 4959 4.699 CH 207 772 26.813

From the table above, it is clear that the Astros typically used the trash can to indicate that an offspeed pitch was coming. Breaking balls and changeups are both offspeed pitches and the Astros banged on the trash can for 27.9% and 26.8% of those pitches, respectively. The Astros only banged on the trash can for 4.7% of fastballs, so it seems pretty clear that the Astros were using the trash can to indicate when an offspeed pitch was coming. This means that in the situations where they were cheating, a bang typically indicated an offspeed pitch was coming while no bang would indicate a fastball was coming. While this evidence clearly shows the Astros were cheating during the 2017 season, this analysis also shows that they weren’t always doing it. It seems they were only doing it some of the time, but it is not clear how they decided when to do it. Further analysis into their methods could potentially give a better idea of when they were cheating.

Tylermarshall21/AstrosCheating documentation built on Dec. 31, 2020, 5:34 p.m.