Jeopardy James
As far as I can recall, the most Jeopardy has cracked into mainstream pop-culture is through some pretty famous SNL sketches. Other than that, Jeopardy has been a long-standing and consistent presence in American homes for decades, happily staying in its quirky and nerdy lane of content. So, I’m not exaggerating at all when I say that over the past few months Jeopardy has been flipped on its head and launched into a new tier of relevance. Jeopardy has beaten out Judge Judy for the number one slot in syndication ratings in recent weeks and stormed into popular culture through news sources, magazines, Twitter, and more. All of this is a result of the success and playstyle of one man, Jeopardy James a.k.a James Holzhauer.
By now, I think just about everyone has heard of Jeopardy James (which is pretty wild if you think about it) but I also think most folks do not know just how dominant his performances were and how interesting his playstyle was. Unfortunately, since I moved out to California and starting working I have not been able to watch Jeopardy as frequently and was not able to watch much of James’ run in real-time (his run just ended on Monday, June 3rd). For that reason, I wanted to dig in and summarize what I missed and hopefully iron out how James plays Jeopardy and what happened in his last game.
Who is Jeopardy James?
James Holzhauer is a 34-year-old professional sports gambler from Illinois. He was a math prodigy growing up that loved sports, statistics, and online poker. He graduated in 2005 with a B.S. in Math from the University of Illinois - Urbana Champagne spending most of his time playing low stakes poker with friends. James won his first big pool of money from the 2006 World Baseball Classic by exploiting what he saw as skewed odds (based on the round-robin format of the tournament and just how variable the outcome of baseball games can be) and betting on all teams except the US and the Dominican Republic to win. He used his winnings from this bet to move to Las Vegas and bankroll his sports gambling career while appearing on game shows like The Chase, 500 Questions, and Jeopardy in his spare time.
How did James play Jeopardy?
Beyond this point, I’m just going to assume the reader knows how Jeopardy works so if you do not, read the Jeopardy Wikipedia page and or watch an episode or two on Netflix (each episode is ~20 minutes long).
So let’s start with the most obvious but most important part of James’ game: he has an unbelievably deep trivia knowledge. In other words, he knows the answers (actually questions). Jeopardy is first and foremost a trivia game. A contestant’s betting strategy can only be successful if they are correct when they buzz in, and James was correct an unfathomable amount of the time during his run on Jeopardy. His response accuracy during his winning streak was a whopping 97%. Ken Jennings (the all-time leader in total winnings and original Jeopardy rock star) had an accuracy rate of 92% at the same point in his Jeopardy run.
Still, just knowing the answers is not enough. Jeopardy does not just let anyone on the show. Ken Jennings once said, “almost all of the contestants know almost all of the answers almost all of the time.” This brings us to the next key piece of James’ success: how he buzzed in. On Jeopardy, it’s not enough to quickly read the prompt and recall the answer, you need to do that and then buzz in faster than the other contestants. Jeopardy’s buzzer system is famously finicky, applying dead-time penalties if you buzz in too early while buzzing in late obviously means someone else gets to answer. Claire McNear at the Ringer explains the system well here:
The moment [Trebek] finishes [reading the clue], a dedicated Jeopardy! staffer sitting at the judges’ table just offstage manually activates a switch that illuminates blue lights alongside the outer edges of the Jeopardy! board. The moment the “enable light” switches on, the three onstage contestants are permitted to ring in, but if they press their buzzers (“signaling devices” in official Jeopardy! parlance) even a fraction of a beat too early, they will be locked out of the system for a quarter-second, which is generally enough time for a competitor to swoop in instead. It’s a mechanism that’s hidden from viewers—you can’t see the blue lights in the telecast.
Per accounts of staffers on the show, James was keen to get as much information as possible on the buzzer system at the start of his run and put in tons of practice based on a guide put together by former Jeopardy champion Fritz Holznagel called Secrets of the Buzzer. So, long story short, James was notably fast and efficient at buzzing in and remarkably accurate with his responses when he did. Those two factors alone would likely be enough for James to accumulate quite a bit of Jeopardy winnings but, of course, there’s more.
The final and most talked about piece of James’ game is the combination of his clue selection and betting strategy. James’ game centers around the Daily Doubles - clues where the contestant can bet up to double the amount of winnings they have accumulated so far. In the first round there is one Daily Double on the board and in the second round (Double Jeopardy) there are two Daily Doubles on the board. With that in mind, here is James’ overall strategy:
Start with the highest value clues on the board
Accumulate as much money as you can while hunting for the Daily Doubles with these 4 main scenarios for each clue:
a. Buzz in fast and get the question right (happens very often if you’re James)
b. Get beat on the buzzer and either pick the clue up if the other contestant gets it wrong or wait for the next clue and try again (happens now and then if you’re James)
c. Buzz in fast and get the question wrong (happens only 3% of the time if you’re James)
d. Don’t try to buzz in (happens rarely if you’re James)
If you find a Daily Double, you hopefully have accumulated enough winnings to make a wager such that, if you get the Daily Double right, it will put you far and away into the lead but, if you get it wrong, you can still feasibly come back to win (based on the other players’ winnings at that point, how many clues are left, and the value of the clues remaining)
a. Note: James’ average Daily Double wager was $9,000 while the average Daily Double wager for all of last season of Jeopardy was $1,664
And that’s it. 29 of James’ 32 wins were runaway games, which means by the time Final Jeopardy came around, he’d already won. His Final Jeopardy accuracy was also 97% and he was always in the lead going into Final Jeopardy, so by following general Jeopardy betting theory in the Final Jeopardy rounds of the non-runaway games, he still won.
James’ Jeopardy performance by the numbers
Jeopardy has some official summary stats here and a fan site digs into more detail here while benchmarking against Ken Jennings. Here are some stats I think are interesting, with Ken’s stats at 33 games in parenthesis:
Total winnings: $2,462, 216 (Ken - $1,100,460)
Average final score: $75,364 (Ken - $33,347)
% of the time first to buzz in: 58% (Ken - 59%)
Final Jeopardy % correct: 97% (Ken - 67%)
Average Final Jeopardy wager: $27,891 (Ken - $7,292)
Average Margin of Victory: $62,269 (Ken - $25,144)
James’ final bet
In his 33rd game, airing on June 3rd, James lost to Emma Boettcher, a 27-year-old librarian from Chicago. Emma was a fierce opponent and ultimately beat out James fair and square, using his playstyle of aggressive Daily Double betting against him.
As laid out above, James’ strategy is straightforward, but not fool-proof. It depends on James landing the Daily Double clues and, on Monday, Emma landed both Daily Doubles in Double Jeopardy. So, for the first time in his entire run, James was not in the lead going into Final Jeopardy. The board was as follows: Emma - $26,600, James - $23,400, and Jay - $11,000. What followed was an uncharacteristically small bet by James of $1,399, but that was the correct bet. Here’s why:
To start, here is James’ primary anchoring assumption for his Final Jeopardy wager: Emma is going to bet $20,201. Emma must cover the amount that James would have if he were to double his money, so she bets $20,201 which would put her at $46,801 to James’ $46,800 if both of them answer correctly and he doubles his score — just enough for her to win. Why is this assumption important? This means that no matter what James bets, he needs Emma to get the question wrong to win.
With this in mind, here are the three possible bet ranges for James (excluding bets that would result in a tie):
Case 1: James bets $0 - $1,399
If Jay doubles up to $22,000 and James gets it wrong he will still beat Jay (James: $23,400 - $1,399 = $22,001)
If Emma gets the question wrong with her bet of $20,201, he will beat her (Emma: $26,600 - $20,201 = $6,399, James: $23,400 - $1,399 = $22,001)
Approximate win probabilities:
Emma: 55% (chance Emma is correct, assuming she is an average Jeopardy player)
James: 45% (chance Emma is incorrect)
Jay: 0%
Case 2: James bets $1,401 - $17,000
As established, Emma wins if she is correct, regardless of what James bets
If James and Emma are both incorrect and Jay is correct, however, then Jay wins (assuming he bets $11,000, which he should if he wants any shot at winning), eating into James’ win probability
Approximate win probabilities:
Emma: 55% (chance Emma is correct)
James: 44% (97% chance James is correct * 45% chance Emma is incorrect)
Jay: 1% (chance Jay is correct and both James and Emma are incorrect)
Case 3: James bets $17,002 - $23,400
Again, Emma wins if she is correct, regardless of what James bets
Now, however, Emma can win even if she is incorrect, so long as both James and Jay are incorrect as well
Approximate win probabilities:
Emma: ~55.5% (55% chance she is correct + ~0.5% chance all three of them are incorrect - again assuming both Emma and Jay are average players and are right 55% of the time in Final Jeopardy)
James: 43.5% (97% chance he is correct * 45% chance Emma is incorrect minus the ~0.5% chance all three of them are incorrect)
Jay: 1% (chance Jay is correct and both James and Emma are incorrect)
Unsurprisingly, James chose Case 1 here and bet the perfect amount for his position to maximize his shot at winning and ensuring he held off Jay. While the estimates above put James’ odds at a little worse than a coin toss, it is safe to assume based on Emma’s performance that she is a far above average Jeopardy player and was shooting way higher than 55% on Final Jeopardy, meaning James’ chance was probably much lower (but the math stays the same). He did the best he could in the situation. But he lost. Indeed, Emma bet $20,201 and was correct (as was James, for what it’s worth) and she sealed her victory.
Ultimately, James finished just about $60,000 shy of Ken Jennings’ all-time winnings record (less than one of his average winnings per game away from the record). What a shame. He also won just about $2.5 million, so don’t feel too bad. Regardless, in terms of winnings per game and excitement to watch, James is probably the Jeopardy GOAT, and we should all take a minute to appreciate his 32 game tear.