Coryat Scores

The joy of Jeopardy! judgement.
Published

October 17, 2022

Modified

October 26, 2023

Introduction

Coryat Scores

The Coryat score is a way of measuring one’s performance when playing along with Jeopardy! at home. It is named after musician, philosopher of physics, and two-day Jeopardy! champion Karl Coryat.

A player’s Coryat score is the total value of clues answered correctly, minus that of those answered incorrectly, counting correctly-answered Daily Doubles according to their board position and ignoring Final Jeopardy! and any incorrectly-answered Daily Doubles.

Thus the Coryat score is a measure of one’s knowledge of the trivia material used on the show, ignoring other strategic elements like wagering.

J! Scorer

J! Scorer is a convenient way to record games and determine one’s Coryat score. The site was created by two-time TV game show contestant Steve McClellan and has a public GitHub repository.

J! Scorer users can download a JSON file of their games.

Data

It took a little effort to reverse-engineer the format of the files produced by J! Scorer. (I did this before finding the above-mentioned GitHub repo.) I used jsonlite::read_json plus a little trial and error. Then it’s just a matter of straightforward data transformations using dplyr.

# read the JSON file, convert to data frame, and unnest some columns
read_json(list.files()[grep(".jscor", list.files())], simplifyVector = TRUE) |>
  as.data.frame() |>
  unnest(cols = c(games_attributes.sixths_attributes,
                  games_attributes.final_attributes),
         names_repair = "universal") |>
  # keep only regular difficulty games
  filter(!games_attributes.play_type %in% c("toc", "masters")) |>
  # keep only the columns we need, add a round indicator, pivot to one row per
  # clue instead of per category, and calculate each clue's contribution to the
  # Coryat score
  select(2, 10:14) |>
  mutate(round = rep(c(rep(1, 6), rep(2, 6)), n()/12)) |>
  pivot_longer(cols = 2:6, names_to = "clue", names_prefix = "result") |>
  rename(result = value) |>
  mutate(result = recode(result, `1` = -1, `3` = 1, `7` = 1, .default = 0),
         score = round * as.numeric(clue) * 200 * result) |>
  # group by game, add up the total score, sort by game order, and throw away
  # the time stamps
  group_by(games_attributes.date_played) |>
  summarize(sum(score)) |>
  arrange(1) |>
  mutate(game = row_number()) |>
  select(-1) |>
  rename(score = `sum(score)`) ->
  # store the results
  coryat_scores

Table

coryat_scores |>
  kable(caption = "My Jeopardy! Coryat Scores") |>
  scroll_box(height = "5in")
My Jeopardy! Coryat Scores
score game
17800 1
10200 2
16000 3
12400 4
16400 5
16800 6
18200 7
16000 8
16000 9
18400 10
18600 11
22600 12
25000 13
17600 14
19600 15
24000 16
19600 17
18600 18
20200 19
17800 20
26800 21
21000 22
20800 23
19600 24
17800 25
18200 26
16600 27
25400 28
4000 29
27800 30
18600 31
27000 32
24200 33
21600 34
22600 35
21400 36
18800 37
11800 38
23200 39
28000 40
24400 41
14800 42
17000 43
16200 44
27800 45
18000 46
18200 47
28600 48
22600 49
13600 50
18000 51
16200 52
18400 53
17200 54
24200 55
19400 56
20400 57
21000 58
16800 59
19800 60
24800 61
19600 62
20400 63
24400 64
18600 65
17200 66
23400 67
26600 68
24800 69
21600 70
24600 71
21600 72
17600 73
12000 74
21600 75
17800 76
23600 77
21800 78
22400 79
18800 80
28400 81
16800 82
21400 83
5600 84
15000 85
19400 86
14200 87
16400 88
23200 89
20400 90
22600 91
18200 92
14000 93
17000 94
19000 95
18400 96
18400 97
26200 98
26400 99
22800 100
33200 101
19800 102
26400 103
18600 104
13600 105
22200 106
24800 107
27400 108
13400 109
14000 110
13600 111
15200 112
12000 113
20800 114
26200 115
16800 116
20400 117
28800 118
32400 119
10400 120
20000 121
12400 122
11400 123
20600 124
19800 125
18200 126
20200 127
20000 128
15800 129
16400 130
17000 131
22400 132
15600 133
20200 134
20600 135
18600 136
19400 137
22800 138
13400 139
18800 140
15200 141
20000 142
14000 143
25200 144
25800 145
20000 146
28600 147
16200 148
16400 149
18000 150
20200 151
18600 152
19800 153
13000 154
10200 155
15600 156
14400 157
25200 158
26800 159
14800 160
21400 161
16800 162
19200 163
17600 164
20800 165
25000 166
27400 167
20800 168
20000 169
26800 170
22800 171
33800 172
16600 173
20800 174
20800 175
27800 176
19200 177
16200 178

Average Score

An average score of around $25,000 is considered appropriate for prospective contestants.

# calculate my mean score
coryat_scores |> pull(score) |> mean() |> round()
[1] 19787

Clearly, I have some studying to do before I consider trying to compete on the show.

Visualization

We can create a histogram showing the distribution of my scores and a line chart showing the evolution of my scores over time using ggplot2.

Code 
# main plot
(coryat_scores |>
  ggplot() +
  # aesthetic mapping
  aes(x = game, y = score) +
  # visual elements representing the data
  geom_line(colour = "#b4b4b4") +
  geom_smooth(se = FALSE, colour = "black") +
  geom_hline(yintercept = 25000, linetype = "dashed") +
  geom_point(colour = "#dc2828") +
  # scales
  scale_y_continuous(labels = scales::label_dollar(), 
                     n.breaks = 6, 
                     limits = c(0, 30000)) +
  scale_x_continuous(expand = c(0, 0), breaks = NULL) +
  # labels
  labs(title = "My Jeopardy! Coryat Scores",
       subtitle = "Trend and distribution", 
       x = "", 
       y = "Score",
       caption = "Dashed black line indicates target average of $25,000.") +
  # theming
  theme_bw()) |>
  # add the marginal histogram
  ggMarginal(type = "histogram", margins = "y", fill = "#b4b4b4")

Code 
coryat_scores |>
  arrange(score) |>
  ggplot(aes(x = score, y = cumsum(score)/sum(score))) +
  geom_line(size = 2, colour = "#dc2828") +
  scale_x_continuous(labels = scales::label_dollar(), 
                     n.breaks = 6, 
                     expand = c(0, 0)) +
  scale_y_continuous(labels = scales::label_percent(), expand = c(0, 0)) +
  labs(title = "My Jeopardy! Coryat Scores",
       subtitle = "Cumulative distribution", 
       x = "Score", 
       y = "") +
  theme_bw()

References