The many games of Jeopardy

January 14, 2020. Double Jeopardy Daily Double:
What is Chad?

Jeopardy is incredibly instructive. It has come up a number of times on the blog. The best summary is the June 2021 NPR Planet Money podcast.

First, Jeopardy is not a trivia game. Jeopardy is a television show. It’s a game within a game within a game within the game of life. Life is a game in the sense that there are rules (hard like physics, soft like psychology), consequences, and randomness.

Penn Jillette notes the gaminess. ‘Winning’ a game like Jeopardy or The Apprentice, wrote Penn, meant being interesting not being champion. His victory is selling tickets to the Penn & Teller theater show, not to make the best beans.

Like all games, the game maker can change the rules. Jeopardy is a pretty clean operation, but games like Survivor change the rules all the time. Governments change rules like “interest rates” or “bailout”. Regulators change rules too, maybe that’s their chief job?

Until 2003, Jeopardy contestants could only win five times. Then the rules changed and Ken Jennings showed up to win seventy-four games. Jennings has a very particular set of skills. Jeopardy was one of the few American shows Jennings watched growing up in South Korea and Singapore. But it wasn’t the clue collection, it was the cadence.

Jeopardy really isn’t a trivia game so much as (per the NPR episode) “a really crappy computer game.” Just as Jeopardy is a game within a game, there are three games within the game of Jeopardy too.

Before addressing the three games, it’s helpful to remind ourselves about the skill and luck spectrum. The “success equation” is to disentangle what games within games are skill based and which are luck based. Skill based games, wrote Michael Mauboussin, are games where some set of actions consistently returns some set of results. Another way to find skill based games is to ask: can you lose on purpose? Most games are a mix.

The outcome of this arrangement, says Mauboussin, is the “paradox of skill”. As each game within a game optimizes, the impact of luck grows.
Kawhi Shot

Jeopardy’s Three Games.

Jeopardy is a series of sequential games. If a contestant wins one, they get to try at the next, then the third, until the process resets. The first game is the buzzer.

Buzzer. “A crappy computer game”. There are a lot of ways to practice the buzzer, the most common among fans is the toilet paper roll holder.

During IBM’s Watson appearance some fans thought that the computer was too fast, though the engineers note the accommodations. Humans, wrote David Gondek, have the ability to anticipate when the buzzer window will open.

Like a basketball free-throw is preceded by a foul, winning the buzzer is precedes a chance to answer.

Trivia. Jeopardy screens contestants via a written test, then (a randomly selected) audition. Everyone that’s on Jeopardy is already good at trivia and getting much better is difficult. Not only that, Jeopardy as a sampling issue: it’s only full of people who really want to be there and do well.

The potential contestants (by now) know that not all trivia is important. The United States (capitals, geography) and its history (presidents) are important. Literature and Science are important. Professional wrestling and heavy metal are unimportant. Jeopardy’s trivia topics are wide but not deep, except for a few areas.

This used to be a data problem. Besides watching each night, how might a potential contestant figure out what mattered? Since 2004 there is a J-Archive, a fan curated collection of all things Jeopardy. Thanks to the internet, this part of the game within a game has shimmied from the luck end of the spectrum to the skill portion.

Though luck still matters. James Holzhauer only knew Sadie Lou was a nickname for Sarah Lawrence College because he and his wife considered ‘Sarah’ as a baby name.

In the season before James Holzhauer, the top ten Daily Double bets ranged from seven to fifteen thousand dollars. Holzhauer raised that to eleven to twenty-five. Holzhauer nearly doubled the Daily Double bets.

It’s not just the betting. Daily Doubles are tactical. If a player finds one early in the round, they won’t have enough to bet to make it ‘a true Daily Double’. Bet too little of a big stack and it lets the other players ‘stay in the game’. Bet too much of a big stack too late and a lead will evaporate.

It was gambling that let Holzhauer to reframe the Daily Double. Think of a coin flip. What is the price and payout someone would play? A dollar to win a dollar is kinda boring.

Or not? At our house we play this dice game called Left, Center, Right. At the start, each player has three chips and they roll three dice, one for each chip. If the dice say left or right, a player must pass a chip left or right. Center and the player adds a chip to the pot. A player can also roll a dot and they keep that chip. It’s a fun game, especially for the younger-non-gaming-gamers. But there’s no skill. The “successes equation” of LCR is: have fun.

The bet a dollar to win a dollar coin toss is meh. What about bet a dollar to win two? That is interesting. That’s how Holzhauer saw Daily Doubles. It cost a dollar to win a dollar but rather than the payout being skewed the odds were.

Jeopardy seems simple. It’s fun to feel smart, that’s part of the big game too, and play along at home. Like poker’s appeal maybe there’s more to it. What is, a richness of thinking?

Timestamps in the episode:
A jarring style, First 20 years, Trivia’s paradox of skill, Finding Daily Doubles, J-Archives, Betting Daily Doubles.

The Paradox of Skills Cascades

To rephrase the aphorism then: If I can’t spot the fool at the table then the fool at the table is me — if all that matters is this game.

In the chapter on The Poker Bubble, Nate Silvers writes that the paradox of skill can have a cascading effect. “The subtraction of fish from the table can have a cascading effect on the players. The one who was formally the next to worst player is now the sucker and will be losing money at an even faster rate than before and he may bust out too, making the remaining players’s tasks more challenging.”

We can revisit the post on Jeopardy James to consider this idea more fully. Broadly, success in Jeopardy is the outcome of four inputs said Ken Jennings:

  1. Trivia knowledge.
  2. Buzzer skills.
  3. Game board strategy.
  4. Luck.

However, these are not obvious at all times. Jennings, for example, says that once his winning streak approached ‘absurd’ levels the Jeopardy producers allowed new contestants more time on the buzzer. Then it was James Holzhauer who demonstrated the third component, looking for Daily Doubles.

Through each step: trivia, buzzer, and board, the weaker players filter out and the competition grows, forcing out further weaker players.

Here’s where our Jeopardy analogy breaks down. The goal of Jeopardy isn’t winning, the goal is entertainment. Much like The Only Honest Sport, the objective isn’t brilliance or bafflement but for the people at home to feel in the game.

Sometimes the game isn’t about winning, sometimes there’s a game within the game, and to win one game one only must stand out in another. Different games have different finish lines too. If someone says ‘best’ or ‘winner’ it’s only in that context.

Tracking Tom Update: We guessed that there’s more that can ‘go wrong’ than right for Tom Brady this season and so far that guess looks just okay. It might be wrong in the end, as Brady is +108 of passing yard pace, but it still feels like good reasoning for taking the under. At the right price, of course.

Jeopardy: Paradox of Skill

Michael Mauboussin introduced the Paradox of Skill as a condition where as skills improve luck becomes more important to determine who succeeds. For example, combine many very talented basketball players and you get something like Kawhi’s shot.

The simplest way to think about the spectrum of luck to skill is to ask, ‘can I lose on purpose?’ If so, it’s more skill and the chance for the paradox.

So what about Jeopardy? Ken Jennings spoke with Nate Silver (of 538) about what his original run was like, what the GOAT challenge was like, and some of the logistics of playing.

Jennings explained that there were four aspects of Jeopardy that made him become the GOAT.

  1. Trivia knowledge.
  2. Buzzer skills.
  3. Game board strategy.
  4. Luck.

Let’s go through those and see where the Paradox of Skill exists and where it doesn’t. Given Mauboussin’s theory, we’d expect it to be present less often given Jennings long run of success.

Trivia knowledge. Ken Jennings is smart in Jeopardy question terms but so is everyone else in Jeopardy question terms. Everyone on the show has self-selected to be there and has passed a quiz to demonstrate their acumen.

“Most of the players know most of the answers most of the time,” Jennings told Silver.

Buzzer skills. The biggest advantage is winning because it means you’ve experienced the game. Unless there is a break to check an answer, Jeopardy is filmed in nearly real-time. Each game is followed by a break to change clothes and onboard new contestants before filming immediately begins again.

Like a basketball player getting into a rhythm, Jennings synced his buzzer cadence with Alex Trebek’s voice. When the buzz-in window opened, Jennings was first in line. (This is also the reason he attributes to Watson’s success).

Game board strategy. It was James Holzhauer who took strategy to its ultimate mathematical end. Jeopardy James figured that if he had the most experience buzzing in, but his competition probably also knew the answer, his best chance was ‘shock and awe’. That meant going for large clues and finding Daily Doubles.

Which are not quite randomly hidden. Jennings said, “The Daily Doubles are not placed randomly. A human physically looks at the gameboard, reads through some clues, and sees what type of clue might be Daily Double friendly. They try to scatter them but pseudorandomness is not actually randomness.” This same thing occurred with Spotify, Ben Cohen told Barry Ritholtz (22:25).

Luck. Ken admits he got lucky. There were probably dozens of categories he could have gotten bounced on. But they didn’t come up or didn’t come up at a bad time.

Ok, so is Jeopardy a game of mostly luck or mostly skill?

Trivia knowledge is a wash with similar competition. Buzzing is a hard skill (once Jennings really got going new contestants were offered longer training sessions to find the rhythm). Strategy is a wash too, now at least. Luck, is, well, just luck.

When Jennings was asked if Jeopardy is a sport, he pauses. It doesn’t really seem like a sport, but it’s the coverage by sport writers that feels the most accurate to him.

Like Jeopardy? We’ve written more about James Holzhauer and Watson. There’s also my ebooks.

Edit, the first version of this post had goofed up formatting. Have fixed.