Models & metrics
BUTTREY Ratings
A single-number strength rating for every Jeopardy player who has appeared on the televised show. BUTTREY uses a Bradley-Terry framework: it observes who beat whom across the entire history of the show and back-solves the rating that best explains the outcomes. A rating of zero is league average; positive ratings are above average, negative below.
Ratings are derived from each player's Coryat performance — the score they would have had if Daily Doubles and Final Jeopardy didn't exist — relative to the Coryat performance of their opponents, weighted across every game they played. Ratings update daily as new games air, and feed directly into the BRING IT Forecaster.
BRING IT Forecaster
The win-probability model behind every chart on Board Control. Given any three Jeopardy players (current or historical), it returns each one's pre-game probability of winning the match. The forecaster takes the players' BUTTREY Ratings, derives expected score differential, and converts that to win probability — the same model shape professional sports analytics teams use to forecast games before they're played.
The same engine powers the in-game win probability charts on the homepage: at every clue, the model recomputes each player's chance of winning given the current score, the clues remaining, and the strength of the field.
Coryat
A long-standing Jeopardy fan-stat: the score a contestant would have had if they couldn't risk anything. Coryat counts every correct answer at the clue's full dollar value, ignores Daily Double wagers (counts the clue at face value), and excludes Final Jeopardy entirely. It strips out wagering luck and asks the simpler question: how many clues did this player actually know?
Board Control uses Coryat as the input to BUTTREY Ratings because it's a more stable measure of a player's quality than final score, which is heavily influenced by wager strategy and the location of the Daily Doubles.
Win Probability
The model's estimated probability that a given player will win the game, computed clue-by-clue from each player's current score, the number of clues remaining, the BUTTREY ratings of all three players, and the locations of any unfound Daily Doubles. Charts on the homepage plot win probability across every clue of every game since 2002 — every spike, every collapse, every Final Jeopardy reversal traceable to its exact moment. Win probabilities sum to 100% across the three players.
Get Rate
The probability that a player will respond correctly to a clue of a given dollar value, computed by tracking historical correct rate. Used by the Daily Double Optimizer: a player with a 90% get rate at $1000 should wager more aggressively than one with 60%, all else equal. Get rates also feed the in-game forward simulation that produces win probability.
Bradley-Terry
A statistical framework for rating players in pairwise competitive games. Originally developed for chess (now generalized as Elo and its variants), Bradley-Terry takes a series of head-to-head outcomes and finds the set of player ratings that best explains the win/loss pattern. BUTTREY Ratings are a Bradley-Terry implementation adapted to Jeopardy's three-player format, with Coryat differentials standing in for binary win/loss.
↑ Back to top
The Jeopardy Excitement Index
Jeopardy Excitement Index (JEI)
A single 0–10 score per game measuring how dramatic the game was. Higher means more drama. The score is computed as a weighted combination of ten excitement drivers — components like lead changes, late-game swings, near-miss Final Jeopardies, and clutch wagering. Each driver is normalized against the entire Jeopardy corpus so percentile thresholds stay stable across eras.
Every game from January 2002 forward is scored. Tier names ("All-Time Classic", "Thrilling", "Competitive", "One-Sided") are derived from population percentiles. Browse the full ranking on the Excitement Index page.
Excitement Drivers
The ten components that combine into the Jeopardy Excitement Index. Each is normalized to a 0–10 contribution, with weights re-fit against an LLM-judged drama benchmark on 2,420 modern games. The ten drivers, in descending weight order:
Round Tempo
The pace of scoring through the Double Jeopardy round. Faster scoring correlates with more clues being correctly answered and a tightly contested game; slow scoring usually signals stalled play. Highest weight in the index.
Final Stakes
Total dollars on the board going into Final Jeopardy. Bigger pots produce bigger possible swings on the last clue, and tournament finals sit at the high end. Second-highest weight in the index.
DD Wagering
The amount of win probability that moved on each of the three Daily Doubles. Bold wagers in close games score high; tiny defensive bets score low.
FJ Cover Tightness
How close the second-place player was to forcing the leader into a perfect cover bet on Final Jeopardy. If the leader could lose by getting Final Jeopardy wrong, the game scores high on this dimension.
Hot Start
Pace of scoring across the first quarter of the J-round. A fast opening means the game came alive immediately; a slow start often previews a sleepy episode.
Buzzer Dominance
A measure of how concentrated buzzer wins were across the three contestants. Specifically, the share of clues won by the most dominant buzzer in the game.
Stakes Context
A multiplier for high-stakes formats — Tournament of Champions, Masters, finals, Invitationals, and similar. Regular gameplay scores neutral on this dimension; tournament games get a context boost.
Comeback Depth
The biggest deficit the eventual winner had to climb out of at any point in the game. A wire-to-wire victory scores zero here; a deep comeback scores high.
FJ Swing
The amount of win probability that moved on the Final Jeopardy clue itself — the size of the FJ-induced reversal, regardless of which player benefited.
Run-of-Correct
The longest consecutive correct-response streak by any single player in the game. Captures dominant single-player runs.
Excitement Tiers
Categorical labels assigned to each game based on its Jeopardy Excitement Index percentile within the entire corpus:
- All-Time Classic — top 2% of all games (score ≥ 7.29)
- Thrilling — top 10% (score ≥ 6.49)
- Competitive — top 25% (score ≥ 6.05)
- One-Sided — bottom 50% (score < 5.60)
Games scoring between 5.60 and 6.05 sit in the unlabeled middle band. Thresholds are recomputed when the index version changes; current thresholds are from JEI v14.
↑ Back to top
Game mechanics
Daily Double
The three "Daily Double" clues hidden on the Jeopardy board: one in the Jeopardy round, two in Double Jeopardy. The contestant who selects the clue wagers any amount from a small minimum up to either their current score or the round's maximum value, whichever is greater. Daily Doubles are the single most consequential decisions in the game — a well-timed all-in can flip a runaway, and a panicked underbet can blow a lead. See the Daily Double Optimizer for the bet recommendation on every Daily Double in show history.
True Daily Double
A wager equal to the contestant's entire current score — risking everything on a single clue. The phrase comes from the original Jeopardy where wagers were capped at the player's score. Today contestants can wager more than their score in some situations, but the term still means betting everything they have.
Final Jeopardy
The single clue at the end of every Jeopardy game. All three players write their wagers in advance (any amount up to their current score), then see the clue and write their responses. Whoever has the highest final score wins. Final Jeopardy is the source of most last-second drama on the show, and feeds two excitement drivers: FJ Swing and FJ Cover Tightness.
Optimal Wager
For Daily Doubles, the wager that maximizes the player's expected win probability, given the score, clues remaining, and player get rates. Computed by simulating both outcomes (right and wrong) across every plausible wager and choosing the size that produces the highest weighted-average win probability. Sometimes the optimal wager is a true Daily Double; sometimes it's a small defensive bet; sometimes it's split-the-difference.
Lock Game
A game where the leader has more than twice the second-place player's score going into Final Jeopardy. Mathematically the leader cannot lose: even betting the maximum and getting Final Jeopardy wrong still leaves them ahead of any conceivable second-place finish. Lock games kill all drama and score very low on the Excitement Index.
Runaway
A near-synonym for lock game — a leader so far ahead that the result is essentially decided before Final Jeopardy. A runaway can technically be broken if the second-place player bets everything and gets Final Jeopardy right while the leader wagers $0 and gets it wrong, but this is vanishingly rare.
Cover Bet
The wager amount the leader needs on Final Jeopardy to "cover" the second-place player — typically (2 × second-place score) − leader's score + $1. A cover bet guarantees the leader wins as long as either they get Final Jeopardy right or the second-place player gets it wrong. The closer this minimum cover is to the leader's full score, the higher the game's FJ Cover Tightness.
↑ Back to top
Tournament types
Tournament of Champions (TOC)
The most prestigious recurring tournament in Jeopardy, inviting the top regular-play winners from one or two seasons. The format historically runs two weeks of quarter-finals and semi-finals leading to a 2-game cumulative final. The TOC champion takes home the largest non-Masters prize on the show.
Jeopardy Masters
A modern invitational launched in 2023 featuring an elite handful of recent champions (typically six players) competing in a points-style round-robin format spread across multiple matches per night. Higher-stakes than the TOC and with a different bracket structure.
Jeopardy Invitational Tournament (JIT)
A bracket-style tournament inviting selected fan-favorite returning champions. Format and stakes vary by edition.
Second Chance Competition (SCC)
A tournament inviting players who appeared on the show but didn't qualify for the Tournament of Champions. The SCC winner historically earns a TOC berth in addition to the prize money.
Champions Wildcard (CWC)
A tournament structured to give one-game-winning regulars a path back to the show. A variant of the SCC theme — selecting players from the broader champions pool.
College Championship
An annual tournament featuring undergraduate students. The College Championship has been running since 1989. The winner traditionally earns a Tournament of Champions invitation.
Teachers Tournament
A tournament featuring K–12 educators, run periodically since 2011. The winner traditionally earns a Tournament of Champions invitation.
↑ Back to top