Analyze “Jeopardy Pavlovs” with Python | by Katy Hagerty | Sep, 2023

Using Python and data analytics to test the strength of Jeopardy Pavlovs

Jeopardy is one of America’s most beloved game shows. Each show, contestants are answered a trivia question and must respond with the correct question. Each game consists of three rounds with 13 categories and 61 questions. Categories include anything and everything from science to popular culture and wordplay to foreign languages.

Over the years, fans have discovered some keywords are automatic giveaways to the correct response and dubbed these pairings “Jeopardy Pavlovs.” For instance, if the clue contains “bachelor president,” the correct response is often “Who’s James Buchanan?”

I recently found an Anki deck of Jeopardy Pavlovs and wrote some Python scripts to test them.

All the code for this project is available on GitHub or Deepnote.

First, I needed a database of past Jeopardy clues. I used a dataset from Kaggle containing clues from shows airing between 1984 and 2019.

In keeping with the Jeopardy format, the answer column contains the clues, and the question column contains the correct response. For simplicity, the question column does not phrase the correct response as a question.